ISFDB:Community Portal/Archive/Archive11

< ISFDB:Community Portal‎ | Archive
Revision as of 00:37, 4 June 2015 by Nihonjoe (talk | contribs) (+template)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

This is an archive page for the Community Portal. Please do not edit the contents. To start a new discussion, please click here.
This archive includes discussions from

Archive Quick Links
Archives of old discussions from the Community Portal.

1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 · 9 · 10 · 11 · 12 · 13 · 14 · 15 · 16 · 17 · 18 · 19 · 20 · 21 · 22 · 23 · 24 · 25 · 26 · 27 · 28 · 29 · 30 · 31 · 32 · 33 · 34 · 35 · 36 · 37 · 38 · 39 · 40 · 41 · 42 · 43 · 44 · 45 · 46 · 47 · 48 · 49 · 50 · 51 · 52 · 53 · 54 · 55

Page Protection Alert

Monday, March 17, 2008 - ISFDB's wiki is getting attacked by spammer robots. Al is working on a long term fix but in the mean time the wiki pages that the spammer is hitting are protected from editing by the general public. It's expected that Al's solution will be installed within a week. If you find yourself locked out from editing a page then please drop a note here, or on ISFDB talk:Community Portal, and we will either unprotect the page, or will do the edit, for you.

Note to moderators - When reviewing Special:Recentchanges if you see a page getting spammed then

  1. Protect the page (using the Protect tab). You don't need to give a reason but do need to check the confirm box.
  2. View the page's History, click on the spammer's account
  3. If there are items in history before the spam then
    1. click "user Contributions" (in th left menu bar). You should see a [rollback] link.
    2. Click on rollback - if the page had already existed before the rollback then it'll work and the page is now protected from spams.
  4. If the spammer is the only item listed in the history then it created a new page.
    1. Click Delete and confirm - you don't need to give a reason.
    2. Click Edit and create a new page that has just {{protected}} on it. You don't need to sign this with ~~~~.
    3. Protect the page.
    4. Note that it's better to delete the page rather than blanking out the spam as that also removes the record of the spam from Special:Recentchanges making for fewer red links to the spammer accounts.
  5. Consider protecting the talk page, or article page, so that both are protected as I've found that this particular spammer robot will try the other if one page is protected and then will give up. Each robot account only seems to spam one page meaning that once the page is blocked that account is also blocked.

The reason we don't block the spammer accounts is because the ISFDB server is behind a Squid web-caching server. ISFDB's wiki sees all editors as coming from a limited range of 10.192.168.X addresses and the version of MediaWiki we are using will also add an IP block when you block a user. Thus even blocking a single spammer account causes random non-moderator editors to get blocked from editing. Thus we have instead been protecting the wiki pages the spammer targets. It's not easy to upgrade this particular wiki as it's fairly out of date and the ISFDB Python code is accessing some MW database tables directly to validate accounts, see if you are a moderator, and to handle the comment/notes from titles and publications to the wiki. The new ISFDB server will have the current MediaWiki code meaning we can install modern add-ons to deal with spammers. Marc Kupper (talk) 17:02, 18 Mar 2008 (CDT)

With the upgrade to MediaWiki 1.12.0rc1 I've unprotected all of the ISFDB pages other than the protected template itself and a few deprecated pages that should not be edited. MediaWiki 1.12.0rc1's improvements include a simple captcha in order to add a link to an external site and a captcha in order to create a new account. Hopefully those two will keep the spambots at bay. Marc Kupper (talk) 19:21, 15 April 2008 (UTC)

Intended Order of Work

Work is progressing on feature changes to the ISFDB, but just to make sure we're all on the same page, I'm posting the general priorities of what I intend to work on. In general, I want to make sure (in the following order) that all data can be edited (it currently isn't), that we support all required bibliographic features needed to make this a serious bibliographic site, that we make the editing of data as easy and error free as possible, and that data is displayed in an informative yet easy-to-read manner. As such, here are the priorities that I am assigning to each area of work:

  1. Fix any showstopping bugs.
  2. Finish port of ISFDB to MySQL/Wiki. This includes:
    1. Removing last vestiges of static html pages. After doing the author directory last month, the only real static area left to look at is publisher information.
    2. Finish editing tools. This would include awards and publishers.
  3. Add features considered necessary for solid bibliographies. These include (but are not limited to):
    1. Translator support
    2. Direct linkage of serials and reviews to titles.
    3. Possible additions to the publication table, such as editions, printings, or non-author roles.
    4. Other requested bibliographic features
  4. Error detection support.
  5. Changes/Enhancements to the display code.
  6. Internationalization.
  7. Dissembler overhaul.
  8. Minor Bug Fixing.

That's the order I consider important to minimize rework and to address current problem areas. Feel free to post here any order changes you would like considered, or any major feature areas not included above. Alvonruff 12:20, 24 Jan 2008 (CST)

Looks good: not necessarily my preferred order, but close enough. If you can spare some time to tell us what some of the intentions are in more detail it would be good - e.g. what will the dynamic publisher pages give us? Will editing Publishers mean we're going to have to start canonicising certain publishers? What about imprints? Translator support sounds useful, but is it at title or publication level? Can we have "Narrator" as well? It should be similar to Translator but even more likely to be pub-level. Direct links of reviews looks good - recent work means that we are matching far more than before, and we don't want those undone as somebody "corrects" the review - but sometimes a review states the exact edition, could there be an optional link to a specific pub as well, if that's what the review is actually about? What are the display improvements - e.g. are you dealing with the ever longer, messier pages of results, or doing something about variant displays? What does "Internationalization" actually entail? BLongley 14:30, 24 Jan 2008 (CST)
One of my current problems is that everyone has a particular area that they are interested in, and I just don't have the time to defocus across large numbers of projects simultaneously. So rather than try to answer all the questions you've posed (not that I have the answers anyway), I'd like to focus on areas one at a time. I'm currently working on awards editing (which will also entail cleaning up some of the columns that are no longer needed), and that's pretty noncontroversial (except what year should be attached to a particular award). So I think the next area to gather input on would be potential changes to the publication structure. Some areas of interest are:
  • Roles - This would encompass ideas like Narrator, Illustrator, Editor, etc. I personally think that a translator creates a new work (title) that can be published in multiple formats (publications), but that's something that should be debated.
  • Fields - Do we have all the necessary fields? For instance, most libraries (which obviously track physical objects) record a publication's dimensions. Do we care about that? People have mentioned tracking editions or printings as separate fields. Are there other fields we should support that we currently do not?
  • Publishers - We currently do not have the ability to edit publishers (like we do with authors), merge them, have supporting fields like web sites, dates of existence, or related bibliographies. We also don't link imprints to publishers like we do with pseudonyms (or even variant titles). This is an area that needs to be designed before writing code, and we should get some consensus on how it should work. There are similarities between Publishers and Authors that we can borrow from.
We can fork off a separate wiki page to discuss this particular topic if we wish. Alvonruff 14:58, 24 Jan 2008 (CST)
I'd suggest you provide a little more detail about your intentions JUST for the items at the top of the list then: "I'm going to do this unless anyone objects" is a nice stimulant. I'm not going to argue about awards, but if the next item is Roles then we can all argue^discuss that while you deal with the first problem. Or Fields. Or Publishers. Work on one, let people discuss the next, and if it gets too complicated break it up and move it down the list. This is what I'm using tomorrow afternoon for MY 2008 workload plans: not sure if it will work, but we can try. (Project management? Program planning? I've heard of them... and know people that will sell you such services... it's usually better to just let something get DONE though.) BLongley 17:43, 24 Jan 2008 (CST)
If no one objects, I intend to create a sub-page for the roles discussion, another for the editions/publications/printings discussion, another for the editable publisher info discussion, and a fourth and final one for discussion of revised/related/based on info and how we cna best capture it. I will link the new sub-pages here. -DES Talk 15:26, 25 Jan 2008 (CST)
Back in December, when it looked like Al may have only a couple of weeks of ISFDB time, I thought that the "low hanging fruit" approach was the best way to get the most bang for the buck. However, now that it looks like Al may be able to spend more time on the project (what's the emoticon for "fingers crossed"?), a more structured approach that will help avoid defocussing sounds like a better deal overall.
As far as creating sub-pages goes, we may want to use the currently existing ISFDB Feature List namespace or, for more extensive discussions, the "Project" namespace. Structurally, it sounds like we may want to do a few different things:
  • Identify any additional data elements that we want added, e.g. printing numbers, recommended ages for juvenile titles, translators/narrators/etc.
  • Identify any additional relationships between existing data elements, e.g. the proposed "relationship" functionality between variant titles and the "based on" relationship for fixups. We may want to look into the standard data elements used by libraries (i.e. the MARC21 standard) and see if any of them are applicable in our case
  • Identify any additional codes that we want to add to our standard tables, e.g. do we want to add anything to "ss"/"nt"/"nv"? Similarly, what are the codes defined for the previously mentioned "roles" by the MARC21 standard?
  • Possibly create new tables of allowed codes for certain fields, e.g. the binding field, and make our data entry forms use drop down lists instead of free text
  • Analyze which database fields are currently used for more than one purpose and create new fields for the abused data elements when necessary, e.g. "Catalog ID/ISBN" and "storylen/jvn"
  • Identify which fields need to be changed to multiples, e.g. add the ability for multiple people to mark publications as Physically Verified
  • Analyze whether any ISFDB fields should be moved to another database record, e.g. do translators belong in the Publication record or in the Title record?
  • Identify any additional functionality that we may be missing and whether we need to add more fields/records to support it or whether we can do within the existing framework, e.g. we may be able to add support for Author level reviews without changing the database structure.
As far as show stopping bugs go, I am not sure we have any at this time. We seem to have a data loss problem associated with pseudonymous reviews, but it's fairly limited and can be identified/fixed after the fact. The only major way for our editors to accidentally cause logical data loss is to change Contents level data in a publication, which then changes that Title in all other Publications -- not the desired outcome in well over 80% of all cases. I have also seen evidence that Author Merge has been used to merge pseudonyms with their respective canonical names, but I can't tell how widespread the problem is until I run some searches. Other than that, our data integrity is generally improving as the ISFDB:Data Entropy demonstrates. Ahasuerus 06:15, 26 Jan 2008 (CST)
Pseudonymous reviews is a problem, but not a major one as ONE of the review entries is still completely valid, I think. It might be worth a script to check for some pseudonymous reviewers we haven't noticed yet. But I can point people at "Fred Patten" Author and "REVIEW" Ttype in Advanced search if they want to start cleaning some. BLongley 18:25, 28 Jan 2008 (CST)
The ISFDB Feature List lacks prioritisation data, so isn't quite up to our needs. If Al wants "Roles, Fields, Publishers" in that order, I've no problem with that. Roles could be comparatively simple to record (add "Translator", "Designer", etc, columns to pubs), more complex to display. Fields likewise: it's easy to add printing number, far harder to USE it. Publishers - well, that's a minefield indeed, but one well worth tackling I think. BLongley 18:25, 28 Jan 2008 (CST)
That's right, one way to handle roles would be to simply add individual columns for each additional role ("translator", "narrator", "designer", etc). However, that could lead to a lot of additional (usually empty) columns in each record and we would have to ask Al to make programming changes every time we added a new role.
Another, hopefully more flexible, approach would be to do what most libraries do and use a table of allowed codes a la the MARC21 list of "relator" codes. If we chose the latter approach, then we would have 0-to-many "role-person" relationships for each record. Our editors could then choose from a drop down list of "jacket designer", "narrator", etc and moderators could add additional items to the "list of allowed roles" (after a Wiki discussion) the way we can add new Verification sources now without having to touch the software. Whether this is something that Al could implement easily -- well, that's a different question :) Ahasuerus 20:39, 28 Jan 2008 (CST)
One of my chief personal desires as far as ISFDB goes would be to be able to run my own copy for development/testing. Last year I set it up on a win-2003 server but am missing a copy of MediaWiki that matches the one used on ISFDB. Where can I get a copy of MW 1.4.5 plus is a database backup available that contains the MW tables that are normally not included in the backup.gz? I'm thinking of rebuilding the development machine on Linux but which version is TAMU using? While I had MySQL and some of the Python code running on win-2003 had had to butcher the build scripts to do it. I'd rather run ISFDB "off the shelf" so that changes can be sent back to Al or others interested in this project for integration. Related to this is should I use PHP 4.4.6 and MySQL 5.0.27? When I installed I took the current versions that were available. Marc Kupper (talk) 03:20, 3 Mar 2008 (CST)
The full ISFDB backup is about 1.5 Gb in size and compresses down to almost 600Gb. I download it weekly, import its contents into MySQL, then delete the history of all Wiki pages as well as all XML formatted submissions and all user specific information. I then export the results to a 100Mb backup file, zip it up and post the resulting 23Mb file for public consumption. If Al doesn't mind, I could create a custom version that would include, e.g., Wiki page histories, but I would need a place to upload it to. Ahasuerus 20:56, 3 Mar 2008 (CST)
It's more important to get the MediaWiki 1.4.5 code. Once that's located we could then work on a version of the backup that's substantially stripped of wiki page histories but does include all of the tables. The only reason I can think of for getting a full backup is to see if a script can be figured out that safely deletes all of the spammer accounts and the spam. For pages that did not have content it's easy - just delete all of the records so that even an undelete is not available. Marc Kupper (talk) 23:28, 3 Mar 2008 (CST)
At a tool called "SubVersion" or SVN is described, with which old versions of MedisWiki software can be downloaded. At is a link to download a copy of release 1.4. i am not quite clear on whether one can get exactly 1.4.5 (which was a bugfix release accordign to relase notes) or only a version that also includes later bugfixes, but it should probably be close enough for the purposes desired. I hope this is helpful. -DES Talk 09:58, 4 Mar 2008 (CST)
There is also a tag for 1.4.5. The 'branch' 1.4 should have the latest version of 1.4. At the bottom of this page there is an example how to download REL1_6_2. --Roglo 12:22, 4 Mar 2008 (CST)
I think the MediaWiki drop I made last year is on an old laptop. If I find it I'll post it on the download page. I don't have a problem with releasing a full backup, but we should at least remove the email strings to conform with the Privacy Policy, and it would be a good idea to zero out the password hashes. Alvonruff 13:09, 4 Mar 2008 (CST)

Project namespace

(unindent) The "feature" namespace seems to be devoted mostly to detailed specs on specific features already on the requested feature list -- what we need here is more general, IMO. I think that one or more pages in the project namespace might be a good way to handle this. I have no problem with that order either, except that roles and fields may interact a bit, perhaps they should be discussed on the same page? -DES Talk 18:42, 28 Jan 2008 (CST)

Sounds reasonable -- go right ahead! :) We can always rearrange things later and set up a bunch of redirects if we decide to change the page layout. Ahasuerus 20:41, 28 Jan 2008 (CST)
Very well, i am about to create ISFDB:Proposed Design Changes. -DES Talk 11:00, 29 Jan 2008 (CST)
David, you mentioned the Project namespace a couple of times. While I understand that ISFDB is the same as a Project, Project:Community Portal is this page for example, I'm not aware that there is anything special about this namespace other than it's common to MediaWiki servers and thus templates, etc. can be copied to another MW server without needing to edit it. Am I missing something or why did you bring up this namespace? Marc Kupper (talk) 02:59, 3 Mar 2008 (CST)
Nothing special about it -- it is the conventional place to put pages about "projects" such as discussions about how the ISFDB ought to be changed/improved, how we should better handle various issues, etc. I was just saying that we needed to discuss these issues and one or more pages devoted to such a discuss would IMO be a good idea -- the page that resulted is ISFDB:Proposed Design Changes, but it could have been called soemthing else or designed differently. such pages had to be put in some namespace, under some name, and it seemed to me that the project/ISFDB namespace was appropriate, because such a page was about the design of the ISFDB. -DES Talk 09:11, 3 Mar 2008 (CST)

March 1, 2008 backup posted

By popular demand, the 2008-03-01 backup has been posted one day ahead of schedule. Ahasuerus 00:16, 2 Mar 2008 (CST)

Namespace Status

Note: I am posting the following message on Al's behalf. Ahasuerus 13:42, 2 Mar 2008 (CST)

My ability to post in the wiki is pretty limited due to a separate Internet issue, so I'll try starting a new, shorter thread.

I have updated the namespace numbers of all affected pages via direct SQL statements. I don't trust the php scripts since our version of MediaWiki is very old (yet another problem), and even if I did we don't have shell access to the system, and those scripts require being run from the command line. So the pages are in the correct namespaces now, which should prevent any dataloss problems.

There is still the issue of the Wiki links, which are not all working properly. The old versions of the articles were not in a real name space, so a magazine like Analog had an article title of "Magazine:Analog". Now that it has a real namespace, the Wiki software prepends the namespace to the article title, so access to that page is of the form <namespace>:<article_title>, or in Analog's case: "Magazine:Magazine:Analog". In order to fix this, the pages need to be moved. So far, I have moved all Publisher:Publisher:Title articles to Publisher:Title. I have moved magazines starting with 'A' from Magazine:Magazine:Title to Magazine:Title.

The move process will leave a redirect page. Eventually we will remove those as well. For now the publisher.cgi bibliography looks for TITLES of the form Publisher:Title - it doesn't know about namespaces, only titles. Publisher:Publisher:Ace has a namespace of 'Publisher' and a title of 'Publisher:Ace'; when it is moved to the correct title, it will have a namespace of 'Publisher' and a title of 'Ace'. Since the publisher bibliography is looking for the TITLE of 'Publisher:Ace' it will no longer find it - but it will find the redirect. I would change the online tools, but unfortunately MY PASSWORD DOESN'T WORK TODAY!

Can anything else go wrong at exactly the same time? —The preceding unsigned comment was added by Alvonruff (talkcontribs) .

Al had e-mailed me with something similar this morning but then I saw that he was posting here. I'm not sure what he's working on at the moment. I've cleaned up the Easton Press / James Gunn articles I added yesterday and am working on the DAW list pages at the moment. Marc Kupper (talk) 13:58, 2 Mar 2008 (CST)
I modified the cur_namespace field in the mw_cur table via a web MySQL interface for all pages that started with Author:, Bio, Fanzine:, Magazine:, Publication, Publisher:, and Series:. When that was completed I still didn't have WEBdav access to TAMU, so I started moving some pages in the Magazine and Publisher namespaces via MediaWiki interfaces. Once that became immensely tiresome, I rechecked my password status at TAMU, and it was finally up. I then made some custom python scripts to remove the pseudo namespace strings from the affected titles. I then modified the ISFDB apps so that they know about namespaces. I did not track down any temporary wiki pages that may have been made in the last couple of days.
So anything that had a pseudo-namespace this morning, is now located in the proper namespace and has had its pseudo-namespace prefix removed. I don't vouch for any other pages in the system at this time. Alvonruff 15:52, 2 Mar 2008 (CST)
Thank you very much. I apologize if in any way my moves and posts on the namespace issue contributed to this being done in a way that caused more work than was needed -- I have tried to help in the cleanup. I think all talk pages for pages that were in pseudo namespaces have now been moved to the proper names, and all references/wiki-links to them have been corrected. It looks aas if Marc Kupper has deleted a lot of the redirect pages created during various moves, adn has done a lot of moves to get things where they properly belong. Thanks also to him and anyone else who helped. -DES Talk 17:10, 2 Mar 2008 (CST)
When it rains, it pours :) Thankfully, the impact was primarily on the Wiki side and regular data editing/lookup were not affected. But now we have real namespaces, so the headache was worth it :) Ahasuerus 17:46, 2 Mar 2008 (CST)
One thing I ran into with the DAW list is when a page is modified/saved that Wikimedia expands the templates and builds an internal table of links. As I used templates in the DAW list the "what links here?" things were showing links from pages that had no links. I ended up needing to do a nonsense edit to add a character to affect a "modify/save" that caused the internal links table to get rebuilt for each page. The SFBC pages probably should get moved to the Publisher namespace and we will run into this as it also uses templates for the table of links from page to page. Marc Kupper (talk) 17:54, 2 Mar 2008 (CST)
Yes, that is a standard and well-known effect of editing a template -- a dummy edit on pages that transclude the termplate may be needed to correct the 'what links here" list.
If you like i will move the SFBC pags and their templates to Publisher:SFBC. -DES Talk 18:33, 2 Mar 2008 (CST)

Fantasy Book Magazine, 1980s incarnation

This was, according to locus, a "Semi-pro zine", but at least two well known SF authors (Alan Dean Foster and Esther Friesner) had work published in this magazine. I have created a wiki page: Magazine:Fantasy Book (1980s) -- the parenthetical is because of the existence of Magazine:Fantasy Book which was published from 1947-1951 and appears to be a completely separate magazine.

That's right, the two magazines were not related except for the fact that they were both semi-prozines. I am sure that the folks behind the 1980s incarnation were familiar with William Crawford's classic/collectible/etc magazine and re-used the title as an homage. Ahasuerus 14:15, 8 Mar 2008 (CST)

Some of the info from the two authors above indicate that specific shortfiction titles were published in specific magazine issues, but full contents and other bibliographic info for those issues is not available at the moment -- should woefully incomplete entries be completed, or should we wait until more complete info is available? —The preceding unsigned comment was added by DESiegel60 (talkcontribs) .

I have approximately half of the issues (perhaps a bit more) in my collection, but I won't have consistent access to them until at least June. It's a borderline professionally looking magazine (think Locus 10-15 years ago), so I am sure that we will eventually want to enter and verify its contents. In the meantime, there is no harm in creating partial entries, but we could also use secondary sources like Miller/Contento as a stopgap measure. Ahasuerus 14:15, 8 Mar 2008 (CST)
I followed the link from Sources of Bibliographic Information to Contento, and they don't appear to have anything more than a list of issues, unless I didn't look in the right place.
By the way, the sources page lists the wstcity pages as a source -- does this mean it is acceptable to enter contents based on covers and contents listings from there? and do we have permission to link to cover images from that site? Is there a comprehensive list of the sites from which we do have such permission somewhere? -DES Talk 14:33, 8 Mar 2008 (CST)
We don't have permission to link to cover images on westcity, but wouldn't want it anyway. Look again at the images. They're all watermarked with the westcity logo. About Fantasy Book: I have most of the issues. Between the two of us, Ahasuerus and I should eventually have them indexed. I'll place it on my projects list. In the meantime you can use alternate sources (Contento, Ashley, NESFA indexes), but please note those sources in your submissions. MHHutchins 15:40, 8 Mar 2008 (CST)
Turns out I only have the 11 issues from March 1984 through September 1986. Hopefully Ahasuerus can fill in the gaps. :) MHHutchins 15:57, 8 Mar 2008 (CST)
Hopefully, it won't be another Thrust-like case! ;-\
As far as "a comprehensive list of the sites from which we do have such permission" goes, there is a list on the Help:Screen:NewPub page in the "Image URL" section. Also, the online version of the Miller/Contento Index doesn't have contents level data, but the CD version does and we have a few editors with access to the CD. Ahasuerus 18:13, 8 Mar 2008 (CST)
The CD edition usually comes out on an annual basis - probably around April. I generally don't enter data from it unless I can verify it from another reliable source (Day or MIT). There are occasional discrepancies and Contento/Miller quite often replaces the credited (and sometime mis-credited) author names with canonical names.--swfritter 18:47, 8 Mar 2008 (CST)

New Series cleanup project

While the servers were resting, I put together a couple of data cleanup scripts and you can see the results on the newly created ISFDB:Data Consistency/Series Numbering Issues page. Keep in mind that while some of the problems are obvious, e.g. a series may start with novel #3 or have 2 novels #4, other problems may not be immediately visible because our software may not display them correctly. For example, if "Series Number" is set to "0", it will not be displayed by the display logic. Also, some Series start with a higher number since "its" earlier books also belong to another series and we don't allow Titles to belong to more than one series. Other than that, have at it :) Ahasuerus 00:30, 11 Mar 2008 (CDT)

If a series number is set to zero, will the zero be displayed in the title edit screen? If not, where will it be displayed and where can it be fixed? (I looked at the title screens of all the volumes that displayed no number in the Amber series, and did not find one that displayed a zero when editing the title data. Does this mean that series is fixed?) -DES Talk 11:17, 11 Mar 2008 (CDT)
When a series starts at a number higher than 1 because it is a sub-series and multiple subseries are numbered continuously, should a note be put on the ISFDB:Data Consistency/Series Numbering Issues page? In the fixed column, or where? (See my note on the Esmay Suzia line for an example of what I thought was a good idea.) -DES Talk 11:20, 11 Mar 2008 (CDT)

We are back! (and so are the spambots)

The ISFDB has been up since about 23:30 server time on Sunday night and we have been under spambot attack ever since as you can see in "Recent changes". Please do not try to block the spam accounts -- due to a known bug in our (ancient) version of the Wiki software, this may block all users! We will have to revisit the issue of protecting certain pages, but the ultimate answer is to upgrade our Wiki software, which will give us the ability to distinguish between humans and bots. Now that the TAMU staff is (presumably) back, Al may be able to get a better sense of when the long promised upgrade may happen. Hopefully. Ahasuerus 23:26, 16 Mar 2008 (CDT)

We had a few dozen spam posts overnight, all reverted/deleted now. The last infestation lasted for a few weeks, on and off, we'll see how it pans out this time around. Ahasuerus 06:11, 17 Mar 2008 (CDT)
Is there anything that can be usefully posted, about what caused so long an outage, and what we can do to avoid it in future? -DES Talk 10:42, 17 Mar 2008 (CDT)
The Cliff's Notes version of the story is that the relevant TAMU's support staff were unavailable due to the spring break. We also ended up with a shiny new blog (, but you probably saw that announcement on rec.arts.sf.written. As far as avoiding this kind of problems going forward goes, well, the long awaited TAMU upgrade (if and when it happens) is supposed to give us a newer/beefier server as well as better support. Another option would be moving to another host. Ahasuerus 12:01, 17 Mar 2008 (CDT)
Thanks.-DES Talk 12:23, 17 Mar 2008 (CDT)

Move to new site

At some point we may want to discuss the pros and cons of moving, but perhaps this is not the time. -DES Talk 12:23, 17 Mar 2008 (CDT)

It's hard to tell what we would have done if we had our druthers, but the decision has been rather unexpectedly made for us. This morning TAMU notified Al that the "ISFDB upgrade/relocation project" is a no-go. Consequently, we will be moving to a commercially hosted site shortly. Al is already in the middle of juggling nameservers, bits, bytes, etc. Hopefully, the interruption of service as far as our users are concerned will be minimal... Ahasuerus 15:26, 17 Mar 2008 (CDT)
If there are ways in which we can help, please post. -DES Talk 15:29, 17 Mar 2008 (CDT)
There were a number of e-mail exchanges over on the "ISFDB Moderators" mailing list while the ISFDB was down and some ideas were mentioned, e.g. enabling offline submissions so that editors could continue working while the site is down, but nothing definite at the moment. Al is doing all the heavy lifting and I am helping with moving backup files around because Al's internet connection is not all that it could be. We will have 500Gb of disk space and a fair amount of bandwidth (5Tb/mo), so the issue of hosting our images is back on the agenda, but again, we can't do anything until we have finished moving. Now that the TAMU server is up, we are trying to drag the discussion back to the Wiki, but the ongoing spam attacks are not helping. BTW, thanks for all the spam-busting help! Ahasuerus 23:25, 17 Mar 2008 (CDT)
Any spam-busting I can do, of course, is gladly offered. I am inclined to think that upgrading the wiki should have first priority once the move is complete, because that will enable several spam-fighting tools. In particular semi-protection would deal with most of the spam, particularly if we ad a CAPTCHA for new accounts, which recent wiki versions support, i think. Then we could discuss things like image hosting, including legalities, and whether in the wiki or in a separate area. -DES Talk 01:51, 18 Mar 2008 (CDT)

Request for at least partial Mod status

As lots of pages are being protected again, to help deal with the spambots, i would like to request to have my wiki-privileges set to admin/moderator status. i think i have demonstrated the ability to handle wiki pages and help out on the wiki, and someone without this status simply cannot edit any protected pages.

I understand that you have normally linked this status to that of moderator on the DB itself, and it may be that people don't feel that I am ready to moderate submissiveness to the DB yet. If the wiki status can be granted without the DB stats, that will be fine for now for me. Or, if the two must be granted together, I will promise not to use the DB rights until others here think I am ready to do so. -DES Talk 16:55, 18 Mar 2008 (CDT)

A list of my wiki-edits can be found here. -DES Talk 16:57, 18 Mar 2008 (CDT)

That's fine with me - I'm not a bureaucrat and so can't turn the moderator flag on myself. Marc Kupper (talk) 16:59, 18 Mar 2008 (CDT)
I think David has demonstrated that his extensive experience as a Wikipedia administrator can be usefully applied here, so I support giving him administrator privileges within the Wiki. At this point there is no way to separate Wiki admin privileges from the ability to moderate submissions on the database side, but I am sure David won't use the latter until and unless he goes through the standard moderatorizing process. Ahasuerus 17:48, 18 Mar 2008 (CDT)
The Borg welcomes you David. Marc Kupper (talk) 19:28, 18 Mar 2008 (CDT)
I've always been an assimilationist. :) -DES Talk 19:31, 18 Mar 2008 (CDT)
Admin flag set. Marc has already protected a metric ton of pages, so things should get a little more stable. Signed: Ahasuerus of Borg ("Your books will be cataloged. Resistance is futile.") Ahasuerus 20:02, 18 Mar 2008 (CDT)
I've always liked "'Oh, bother', said the Borg. 'We've assimilated Winnie the Pooh'."
Or "'Your ass will be laminated!' said Dyslexic of Borg." Still, there's no real need for Borg jokes, welcome to the Collective. BLongley 18:46, 19 Mar 2008 (CDT)

2008-03-18 backup file posted

The latest backup file has been uploaded. Now that the spambots have been mostly thwarted, it looks like we are back to normal after what will undoubtedly become known as the Spring Break Fiasco of 2008 :) Ahasuerus 21:04, 18 Mar 2008 (CDT)

Nah, lousy title, it'll never catch on. It doesn't work outside the US. How about "The TAMU Tumult of Twenty-Oh-Eight"? "The ISFDB Downtime Massacre"? "When Harry Met 404"? "2008: A Specfict Oddity"? "The Incredibly Strange Creatures Who Stopped ISFDB-Editing and Became Mixed-Up Zombies"? BLongley 19:01, 19 Mar 2008 (CDT)

Reality gets stranger than specfict at times

I generally don't forward stuff but this one is amazing from a specfict perspective - YouTube W1czBcnX1Ww is a robotic "dog" that walks more like a goat. Marc Kupper (talk) 18:06, 19 Mar 2008 (CDT)

Walks like a Goat, sounds like a Bee! Does it sting like a Butterfly? ;-) BLongley 18:31, 19 Mar 2008 (CDT)

ISFDB:Privacy policy

One of the new pages introduced by MediaWiki is ISFDB:Privacy policy but I have no idea what to put on the page. I'm doing the thread here on ISFDB1 so that when we cut over to the new server we'll have the Privacy policy page set up. Marc Kupper (talk) 21:52, 20 Mar 2008 (CDT)

See Privacy Policy :-) Ahasuerus 22:06, 20 Mar 2008 (CDT)
Why don't we move the existing policy to the expected location under the latest version of MediaWiki, so that when we cut over, it'll automagically appear? Alvonruff 15:24, 21 Mar 2008 (CDT)
Done. -DES Talk 15:48, 21 Mar 2008 (CDT)

ezine policy

The dust may have settled on the ezine issue and the ISFDb Policy Page needs to be updated to reflect current practice: ezines are included if they are downloadable and have been assigned an ISSN. Although my personal preference is for a broader policy, applying the same standards to downloadable magazines as print magazines and even perhaps including website only zines, I intend to update the policy sheet only to reflect the current standards: downloadable with an ISSN.--swfritter 11:15, 21 Mar 2008 (CDT)

I thought that the policy was that any ezine that was downloadable, stable in its issue contents, and seemed likely to be reasonably persistent was includable. I took the ISSN as proof of issue stability, and strong evidence of persistence, but not as a sine qua non of ezine inclusion. Please note that I have added multiple issues of baen's Grantville Gazette. None of these have ISSNs. Some -- mostly the ones also issued in print form -- have ISBNs. Some have DOIs -- or more accurately they have strings in the form of DOIs, but which seem not to have been registered a valid DOIs. -DES Talk 11:30, 21 Mar 2008 (CDT)
Any discussion should probably be on the Policy Talk page or Rules and Standards. I find your definition acceptable but based upon the extensive discussions on this subject matter I only felt comfortable adding ezines with ISSN's. I would like to add Aeon magazine which is currently at issue 13, has made the transition from semi-pro (3 cents a word) to pro (6 cents a word), is regularly issued and available at Fictionwise, has well recognized authors but does not have an ISSN. With Grantville Gazette totally blurs the concepts of ebook anthology and ezine.--swfritter 12:41, 21 Mar 2008 (CDT)
Fair enough. I will copy this to the Policy Talk page, as it seems to me this is a matter of inclusion policy, rather than bibliographic standards -- the latter seem to me more focused on how we enter and display data that we have agreed we want in the DB. -DES Talk 13:05, 21 Mar 2008 (CDT)
BTW, it is usually better to use wiki links to pages in the wiki, rather than URL links -- they make "what links here" work properly. Not a big deal though. -DES Talk 13:05, 21 Mar 2008 (CDT)

2008-03-24 backup file uploaded

The 2008-03-24 backup file has been uploaded. The next one is unlikely to be created until 2008-04-07 or so due to an unusually busy travel schedule. Ahasuerus 23:37, 24 Mar 2008 (CDT)

Encyclopedia of Speculative Fiction

Is anyone here using Encyclopedia of Speculative Fiction?

I had just did a search for an author and found The bottom of the page says

No, I don't use it, although "" keeps turning up in searches where it shows a typo HERE is propagated to several other sites. It seems to be a purely Wiki version of what we're doing - i.e. it's not going to give me the chance to discover that "Barclay Shaw" is a price via SQL. I can do more via the database here, so I'm not tempted to move over. BLongley 18:29, 25 Mar 2008 (CDT)
At one time there were a few Web spiders that periodically crawled the ISFDB and made the data available on unrelated Web sites in various formats. "" seems to be the most prominent one at this point, although I occasionally run into what appear to be snapshots of various ISFDB pages (and variations thereof) as of 5-10 years ago elsewhere. Ahasuerus 23:36, 28 Mar 2008 (CDT)

Analog Magazine Page

Could a moderator unlock this for a short while please? Thx, rbh (Bob) 16:31, 28 Mar 2008 (CDT)

Hurry quick. Stay away nasty spambots.--swfritter 20:36, 28 Mar 2008 (CDT)
Done! Quick, close the door! rbh (Bob) 21:57, 28 Mar 2008 (CDT)
Re-protected.--swfritter 22:04, 28 Mar 2008 (CDT)

Developer Question

Hi, I've written a small program for managing my (html-based) ebooks that I'd like to incorporate your data into indirectly by querying over the web, but I've found no mention or plans of a web-based API. Is it considered acceptable to just 'scrape' the pages for the information I'm interested in? (I'm definitely making every effort to cache the results of all queries in order to use the minimum of both your bandwidth and server resources) Are there unpublished plans for a web API in your future? Relatedly, do you need any development help? My spare time is limited, but I'm willing to pitch in somewhat - I've got several years of python experience. --PaulJ 14:51, 29 Mar 2008 (CDT)

Greetings! We don't have a Web-based API at this time, but we post the backup file of the underlying MySQL database on the ISFDB Downloads page every week (give or take), which is probably the easiest way to get to the data programmatically. Would that work for you? Ahasuerus 14:56, 29 Mar 2008 (CDT)
Well, I'd rather scrape pages since that way I always get the 'freshest' data without having to keep doing the download/update process on my end. Also my needs are pretty simple - I'm just doing an isbn->series(name,number) lookup for a couple hundred books initially and then a trickle of them as I get new ones. Are there any plans for a Web-based API? I'd be more than happy to lend a hand with it if so, or to help with design and implementation if there's no plans but you think it's a good idea :) --PaulJ 20:16, 29 Mar 2008 (CDT)
That's a good question for our programmer, Al. I'll drop him a line, thanks! Ahasuerus 22:58, 29 Mar 2008 (CDT)
Hi Paul. Something I've thought about is the concept of watchlists. If someone was interested in a couple hundred books they would watchlist them and get notified as there were updates. For a long time Al was the only one that could do meaningful ISFDB development as the current server is based on a rather old version of MediaWiki. However, a new ISFDB server is coming on line where we have upgraded to the current MediaWiki. There's still a few bugs being worked out but we are getting to the point where people can set up test/development systems to play with ideas that could then get merged into the main source tree. That said - I suppose the first priority is to get some test systems set up. :-) Marc Kupper (talk) 02:44, 30 Mar 2008 (CDT)
Re: watchlists, how about RSS feeds? On arbitrary objects or collections of objects, of course. Then you could do a search and use the RSS link that the search creates for you to watch for changes in the results of that search. Or you could subscribe to book/author/series RSS feeds individually. --PaulJ 22:20, 3 Apr 2008 (CDT)
Hm, how is the MediaWiki related to installation of the ISFDB? If you need just the data on publications, you can live without Wiki (or use online Wiki). --Roglo 13:23, 30 Mar 2008 (CDT)
For the most part they are not related but the following parts of ISFDB use the MediaWiki MySQL tables; Login, author display (for link to Bibliographic and Biographic wiki links), publication display (for Bibliographic wiki link), Series display (for Bibliographic wiki link). Your login (or lack of login) also controls which options are shown to you on all pages. I'd need to test to see if the login state is stored in cookies or the MediaWiki database. It's not clear if/when the WebAPI is made public if anonymous data extraction will be allowed or if the code will require you to log in. Marc Kupper (talk) 03:51, 2 Apr 2008 (CDT)
Certainly it is useful to have MediaWiki and ISFDB together but we do not have backups of the Wiki, so it won't help much for browsing the data off-line. From my experience, the application is working with backup installed without MediaWiki (i.e. the code looking for wiki data doesn't cause errors) and initially I didn't realize that mw_user, mw_user_rights tables are from MediaWiki. I used to create mw_user_rights (the fields required by ISFDB code), insert some minimal data into these two tables and comment out the password check but finally I wrote a script to setup a moderator's account. I can paste the scripts into wiki or upload them to, if anybody is interested in them. One problem I have is that in, HTMLHOST defines both were the static files are (e.g. stylesheet) and were the server with wiki is, so that if I want the on-line wiki links, then the styles are downloaded, too. --Roglo 04:46, 2 Apr 2008 (CDT)
At present there is support for two Web API calls. The current XML submissions (such as NewPub, PubUpdate, etc) are support via a REST implementation that calls for embedding those submissions within an IsfdbSubmission tag. Additionally there is a REST implementation to extract publication data via ISBN. Both of these APIs are currently used by Dissembler. These have not been made public as yet, as we should have registration support to prevent abuse. Only registered users can currently edit the ISFDB, and without a registration key, it would be trivial for someone to spoof submissions using known moderator logins.
The registration support would be pretty rudimentary. It would require a new MySQL table that contains user IDs and registration keys. There would need to be an edit application that generates a key for a logged in user, or displays the current key for a registered user who has already generated one. The REST api would have to be updated to check for the key, and the IsfdbSubmission tag would then have a child tag for the registration key. Once that's done, then we can publicize the API (and I can make the REST API part of the source distribution). Alvonruff 08:45, 30 Mar 2008 (CDT)

Nominating Davecat for moderatorship

Ref: Moderator Qualifications#Becoming a moderator for the nomination process.

Nomination statement

I nominate Davecat (talkcontribs1) for moderatorship; he has accepted the nomination. Davecat has 2333 edits and counting. He has been doing some very detailed and comprehensive work on Analog including Project Gutenberg records associated with that magazine. He has actively participated in numerous discussions and gives his opinion freely. At the request of the currently traveling Ahasuerus and on my own behalf I gladly place this nomination.--swfritter 18:45, 31 Mar 2008 (CDT)


  1. Support, as nominator.--swfritter 18:45, 31 Mar 2008 (CDT)
  2. Support I generally don't have much to do with the magazines but I do monitor the general discussions and Dave appears to be ready to join Al's merry band of moderators. Although some caution would be prudent when he deals with areas unrelated to magazines.Kraang 21:04, 31 Mar 2008 (CDT)
  3. Support More moderators the merrier. Dave certainly has the knowledge to approve his own submissions, especially within his chosen specialization. And I feel he'll be able to spread out to other areas at his own pace. MHHutchins 21:53, 31 Mar 2008 (CDT)
  4. Support Magazine editors and moderators are in short supply. It feels like it has been over 6 months since the last moderator was nominated and sometimes the backlog backs up. With the amount of work involved in magazine entries, I often make partial entries (all entries from the TOC and illustrations, then reviews) and have had to wait up to 3 days to finish. Another moderator will help! Thx, rbh (Bob) 22:51, 31 Mar 2008 (CDT)
  5. Support I've approved a lot of Davecat's submissions and find the quality of his work to be uniformly excellent.--Rkihara 10:31, 1 Apr 2008 (CDT)
  6. Support Lots of good work over a number of months; good working knowledge of the magazine maze; good communication skills. Ahasuerus 13:52, 3 Apr 2008 (CDT)



  1. Neutral Leaning towards Support - we don't really overlap often, although as I like my Short Fiction we no doubt will work together again. I'll leave it to the Magazine Mods though, as the things he apparently does particularly well are not really my area. BLongley 18:16, 1 Apr 2008 (CDT)
  2. Neutral Leaning towards Support - As with Bill, we David and I don't overlap often. I do recall approving submissions at times and have also never had a reason to drop a note on his talk page. Marc Kupper (talk) 03:56, 2 Apr 2008 (CDT)


Nomination is successful as per the consensus above and is closed after 5 days. Administrator flag set on Davecat's account. Congratulations! Ahasuerus 20:32, 4 Apr 2008 (CDT)

Modified rapture ...
I'm honored (and, for Bill, also honoured). I expect to feel my way slowly on this. Thanks for the vote of confidence & comments. -- Dave (davecat) 12:14, 6 Apr 2008 (CDT)

Nominating Roglo for moderatorship

Ref: Moderator Qualifications#Becoming a moderator for the nomination process.

Nomination statement

At Ahasuerus' request Roglo (talkcontribs1) is nominated for moderatorship; he has accepted the nomination. Roglo has 1526 edits and counting. Based upon his contributions he brings a unique knowledge of foreign language works. At the request of the currently traveling Ahasuerus I place his nomination--swfritter 18:45, 31 Mar 2008 (CDT)


  1. Support, as surrogate nominator. Although we work in different spheres and I am not totally familiar with his work. But it would be a little unseemly for even a surrogate nominator to do otherwise than support. Based upon what I see on his talk page he is a thorough editor and a good communicator.--swfritter 18:45, 31 Mar 2008 (CDT)
  2. Support All the submissions I've approved have generally been without issue and his work on the Stanislaw Lem page put it in excellent shape. I believe Roglo has a good knowledge of how the database functions and will have no problem approving submissions, his and others.Kraang 21:37, 31 Mar 2008 (CDT)
  3. Support I've not had any major problems with Roglo's submissions. His progression from a neophyte to a consistently knowledgeable submitter was surprisingly fast. And just take a look at his user page! The man has plans, and we need to give him the access and means to fulfill them. MHHutchins 22:02, 31 Mar 2008 (CDT)
  4. Support Although I know little of his work with foreign language works, we certainly need specialized coverage as well as general coverage. Thx, rbh (Bob) 22:51, 31 Mar 2008 (CDT)
  5. Support I've had no problems with his submissions except when he's more learned in foreign languages than me (OK, that's pretty easy) but I appreciate his work and also like the look of some of his proposed projects (Phil Dick, Publication series, etc) and am happy to let him get on with them. BLongley 17:53, 1 Apr 2008 (CDT)
  6. Support as the nominator-in-absentia who worked closely with him on Lem's bibliography and some other projects. Roglo has a long history of programmer level contributions to open source projects, knows Python and has a local copy of the ISFDB database running, which has enabled him to come up to speed quickly. Ahasuerus 13:57, 3 Apr 2008 (CDT)



  1. Neutral I am abstaining, since Roglo works mostly in areas that I am unfamiliar with.--Rkihara 10:48, 1 Apr 2008 (CDT)
  2. Comments I'd like to see Roglo START communications more often, especially when his projects overlap with other people's. BLongley 17:53, 1 Apr 2008 (CDT)
  3. Neutral leaning towards Support - I have not had a chance to deal with many Roglo submissions but from what I see on the talk page he's well on his way to understanding the system well enough that he won't press the bright yellow "press here" button right away. :-) Marc Kupper (talk) 04:15, 2 Apr 2008 (CDT)


Nomination is successful as per the consensus above and is closed after 5 days. Administrator flag set on Roglo's account. Congratulations, Robert! Ahasuerus 20:34, 4 Apr 2008 (CDT)

Thank you all for your support and kind comments. --Roglo 11:58, 5 Apr 2008 (CDT)

2008-04-06 backup file uploaded

...and we should be backup to our more-or-less weekly schedule for a while. Ahasuerus 01:03, 7 Apr 2008 (CDT)

Combine downloadable ezines with print magazines?

Marc Kupper's suggestion. I will merge them in a couple of days if there are no objections. A related issue: Black Static, the horror sister to Interzone, is both a print and ezine publication. Nobody has entered any print editions and I intend to enter the ezine editions if nobody enters the print editions. The electronic editions get to America a lot faster and cheaper than the print editions.--swfritter 16:31, 9 Apr 2008 (CDT)

Sounds reasonable to me. There are quite a few SF magazines straddling the print/ezine fence (e.g. Omni), so using the Magazine namespace for all of them seems to be the easiest way to keep track of them. Ditto with print/ezine fanzines. Ahasuerus 16:45, 9 Apr 2008 (CDT)
Webzine volatility alert!!! Check out this webzine. They completely lost issue #1 and all backups.--swfritter 17:29, 9 Apr 2008 (CDT)

Data Consistency - mismatches updated

I have re-run my script against the 2008-04-06 backup and linked the updated mismatch pages on the Data Consistency project page. We are still getting new mismatches from partial data fixes and new submissions, but at a greatly reduced rate, so overall the situation has improved significantly in the last 6-8 months. Now if we could clean up R. L. Stevenson's biblio... Ahasuerus 03:28, 10 Apr 2008 (CDT)

List of publishers uploaded

I have uploaded Bill's list of publishers (in CSV format) to our file library at . I also added a copy of a nice Web based secondary bibliography of Lovecraft. It's a Word document written in French, but all English language Titles are in English, so it's quite usable even if you don't know French. The third file contains connection data for all publicly available Z39.50 servers that IndexData was aware of a year ago. If you are into writing your own Z39.50 clients, it's very useful. Of course, feel free to upload any other documents that you may find useful! (You have to become a member to upload files, but that's easy enough to do.) Ahasuerus 12:56, 10 Apr 2008 (CDT)

ISFDB editing disabled tonight

Just another note to point out that, as per Al's updates to the Main Page, the database will be read only for a couple of hours after 4pm tonight as we move to the new server. Ahasuerus 15:47, 12 Apr 2008 (CDT)

Well, here we are! The furniture is slightly used, but looks OK. The china is somewhat on the cheap side, but should do. Carry on! :) Ahasuerus 02:35, 13 April 2008 (UTC)
P.S. Please note that links from the new ISFDB pages to the Wiki are currently pointing to, so they will remain broken until has been re-pointed to the new location. You can access all Wiki pages directly at Ahasuerus 03:14, 13 April 2008 (UTC)
At this point the DNS changes have propagated to some places, but not everywhere - I still see the old TAMU site, but pings from to go to the correct host. We're getting indexed by Yahoo, MSN Search, and Google (they must launch the bots upon receipt of a DNS change) so the site is a little busy today, as we haven't been indexed by any bots for a while. Alvonruff 14:02, 13 April 2008 (UTC)

Van Vogt's Quest for the Future

To quote Mike Schilling's post on rec.arts.sf.written:

_Quest for the Future_ is a fixup novel, not a collection. It combines three unrelated stories, giving them all the same protagonist and adding linking material.

We currently list 4 publications for this Title and 2 of them are verified (by Kraang and Bill). The two verified pubs do mention that it's a fixup, so it sounds like it should be safe to change all 4 pubs to "Novels", but I'll wait a day or two in case there are any issues. Ahasuerus 18:37, 13 April 2008 (UTC)

Fine by me. I've moved the notes about the component stories to title level for now, along with Icshi's notes about where the Stories can be found within the book. I've also added coverart, as Icshi granted permission to use his freely. I haven't added the remaining five publications he knows of yet, to save time in conversion, and because he lists my copy as third NEL printing but there's only evidence in it to suggest second printing. (He's probably more right than me and NEL though.) BLongley 19:31, 13 April 2008 (UTC)
OK, I have changed the book's Title record and all Publication records from "Collection" to "Novel" and Removed the three individual stories, adjusting Notes accordingly. Should be all set, hopefully. Ahasuerus 01:04, 17 April 2008 (UTC)

DNS Problems with

FYI, when accessing pages, all links now point to, which is currently down. I suspect that it's a DNS propagation issue which may only affect some editors. Ahasuerus 22:21, 13 April 2008 (UTC)

It appears that I'm locked out of the main data base. I keep getting the same message "The requested URL could not be retrieved". Earlier today I had excellent access and than it suddenly shutdown. Any suggestions or should I just wait and it will fix itself.Kraang 02:11, 14 April 2008 (UTC)
What URL(s) are you using and when you ping and what address to you get? You should be getting Marc Kupper (talk) 07:58, 14 April 2008 (UTC)
I'm using and that gets me to the home page without the nice graphics, but from there everything I try to do gets me this
The requested URL could not be retrieved


While trying to retrieve the URL: 

The following error was encountered: 

Access Denied. 
Access control configuration prevents your request from being allowed at this time. Please contact your service provider if you feel this is incorrect. 

Your cache administrator is 


Generated Mon, 14 Apr 2008 10:58:57 GMT by (squid/2.6.STABLE6)
Does this make any sense to anyone and what can I do?Kraang 11:07, 14 April 2008 (UTC)
It looks like your ISP's DNS server, which runs the application that routes your browser's requests to the actual computer running the ISFDB database, is still configured to send you to the old TAMU server, which generates the error message that you see. If you are using Windows, click on Start, then on "Run", then type "cmd" at the "Open" prompt. This will bring up a big box with a DOS-style command line. Type "ping" and check the numeric strings that the system displays after a couple of seconds. If you don't see, then you can't really do much ISFDB-wise until the DNS changes are propagated to your DNS server :( Ahasuerus 13:02, 14 April 2008 (UTC)
I'm getting which is TAMU's server. Hopefully it sorts itself out soon. Thanks :-)Kraang 13:17, 14 April 2008 (UTC)
> I'm using and that gets me to the home page without the nice graphics, but from there everything I try to do gets me this
There are some server configuration issues with the have of the server wired in too many places. When you are viewing the ISFDB images should come from but instead are referencing which causes problems for you as you can't see the new
Run "ipconfig /all" - what are the results? Specifically, I'm interested in what your DNS server(s) are set to.
Take a look at your etc/hosts file. For Windows this would be in %windir%\system32\drivers\etc\hosts and for Unix it's /etc/hosts. Are there any lines for ISFDB? If so, you can change them to use or better yet would be to remove or comment the lines out (with #).
If you really want to do a hack then edit the hosts file and add
Exit/restart your browser and you should see the new site. Marc Kupper (talk) 18:54, 15 April 2008 (UTC)

2008-04-14 backup file uploaded

This is our first post-move backup file. It should contain the same data as the old ones, but a number of things have changed, e.g. we have a bunch of new Wiki tables, so please post any issues that you may run into. Ahasuerus 15:48, 17 April 2008 (UTC)

I can confirm the old Entropy scripts still work, at least. I haven't had time to check the new tables yet though. BLongley 22:47, 17 April 2008 (UTC)
Well, the idea was to delete the new Wiki tables as part of the cleanup process, at least until we know what they do. Whether I succeeded is a different question :) Ahasuerus 00:07, 18 April 2008 (UTC)
When will be uploaded? I'd like to get my ISFDB server to match the production machine. Marc Kupper (talk) 07:47, 19 April 2008 (UTC)
The Python scripts that drive the application are maintained by Al, who uploads (hopefully) stable versions when he gets a chance, so you'll have to bug him :) Ahasuerus 17:45, 19 April 2008 (UTC)
Are the Sourceforge versions out of date then? BLongley 20:23, 19 April 2008 (UTC)
I don't see any versions. --Roglo 20:34, 19 April 2008 (UTC)
Oops - it looks like Al never had time to check in the code. I've e-mailed him. Marc Kupper (talk) 20:45, 19 April 2008 (UTC)
It would be good idea to set as the project's website - is empty. --Roglo 21:02, 19 April 2008 (UTC)
Done - I might as well upload the old code - is there a site that would serve as a good model for how to structure it? Marc Kupper (talk) 01:51, 20 April 2008 (UTC)
For the Downloads section something like this, I think. We have one package 'isfdb2', and release would be the date of the code. I would rename the file: isfdb2-2008-01-27.tar.gz . As the release notes: Info which backup is required (I'm not sure about the exact date; perhaps 'recent'?); pointer to our download page for the backup and installation info, pointer to our wiki for more info. --Roglo 09:44, 20 April 2008 (UTC)
Al has now uploaded the latest version to the ISFDB server, so we are back to normal :) Ahasuerus 16:45, 20 April 2008 (UTC)
Time to update :) But are we going to use more (e.g. upload patches there)? Upload the code? --Roglo 17:42, 20 April 2008 (UTC)
Roglo, Why don't you just add yourself to the project? :-) Yes, while Al has not been using I believe that once the code is there that he'll use it. I've never used which is why I was asking about best practices in uploading the code. For a long time the code was available as a tarball and with Al being the only developer that was fine. There was a desire by others to get involved in coding but building a fully functional development/test system turned out to be quite a challenge. Marc Kupper (talk) 07:12, 21 April 2008 (UTC)

Alternate bibliographic sources

Are listings at considered reliable enough to base publication entries on? Perhaps if confirmed by OCLC/Worldcat data? I am seeing titles listed there that we don't have here (for relatively obscure authors).

And speaking of OCLC/Worldcat would it ba a good idea to add that as a source that a publication can be verified against? or is there a good reason not to do this? -DES Talk 20:53, 21 April 2008 (UTC) is more or less and data with the author and story bios written by hand. As a bibliographic resource (of the works an author contributed to) it should be fine. Dave Wands has also put a lot of attention on series and those are usually accurate.
In general - anything can be used as a source. I frequently use Amazon, Abebook seller listings, etc. but when I do so I always add a citation noting the sources for each field. For example, the basic publication data may be from Amazon with the page number, cover price, and cover artist from Abebook seller listings. If you do this often enough you then you'll be able to spot which sellers just regurgitate Amazon garbage (not even bothering to get their listing to match the physical publication they are selling at all) and which sellers genuinely love books and take care in their seller listings. Marc Kupper (talk) 23:13, 21 April 2008 (UTC)
No site is going to be perfect - only today I found myself wondering why we and Wikipedia have mostly "Hanns Heinz Ewers" and Fantastic Fiction has "Hans Heinz Ewers". Neither is a Canonical source IMO, worth elevating to "Verified against" status. But a STABLE and confirmed as mostly accurate (to the 98%+ level) and widely-enough-scoped source would warrant another Verification source, IMNSHO. OCLC is looking quite useful, but for some fields only - they estimate dates at times, and Publisher info is weak or truncated - e.g. "London: Cape" for "Jonathan Cape" or "London: Panther" for the Panther imprint at a time when they were 20 miles north in St. Albans. :-/ OCLC is good for finding ISBNs (except for series of similarly named titles) that can be used elsewhere though, and foreign variants of titles. However, most good sources I find are very narrow in scope (at best for a specialist SF imprint or publisher and more often for a particular author) and so I mention those on Publisher or Author pages as POSSIBLY worth a look. We can't always find people as GENERALLY fanatic about SF books as us, but it's often not hard to find someone else fanatic about an Author, a Genre, a Publisher, an Imprint, a Series... BLongley 00:29, 22 April 2008 (UTC)
I deliberately AVOID colouring general entries with sources though. If I'm trying to date/price a specific book I have, yes, I'll go into much detail about why I chose that date, that price, that form of Publisher/Imprint name, etc. When it's a case of "ISFDB has a title, author and year ONLY" then I'll often add a publication (preferably the first) to get things started and point people in the right direction. Don't worry, I'm not just making them up, I also tend to use at least two sources - which is why I've left the submission queue a bit messy recently as some good fannish sources need a bit of checking before I'd approve my own edits based on their data. But I don't want to say, for instance, "all the missing van Vogt pubs I entered came from Icshi" and have people accept them based on Icshi's reputation (or mine, if I have one) - I want people to see an unverified pub and go CHECK things. But I'm not intending to send people on Wild-Goose chases, so if I have entered such, you know it's for good reason. BLongley 00:29, 22 April 2008 (UTC)
Oh, and before anyone gives up on Amazon entirely - for books that were around before Amazon even existed, they have some useful data still. The TRUE imprint or publisher might be there - but in the title rather than in the publisher field. Some serial numbers are there too. And often a format - there's many "unknown" format pubs with "hardcover" or "paperback" in the TITLE! Amazon is often a good start for finding things to search on - just don't submit their data unchecked. That's Dissembler's job. ;-) BLongley 00:38, 22 April 2008 (UTC)
Oh, and another thought - sometimes Amazon are good at placing an UPPER limit on page numbers. I'm specifically researching British paperbacks of the 1960s and 1970s at the moment. So I've seen a lot of books Amazon obviously don't really know about with a page count of "192" for instance - where this IS in fact the number of pages inside the book, they just don't know how many were used for adverts after the main text. I'm not happy to use Amazon page-counts in general (I've seen them several HUNDRED pages out) but occasionally I think that "default page-count" for a paperback from a particular publisher for a particular time-period is something worth recording. At least it should tell you that it's going to be "192, less a few pages". (Same with 168 and other suspiciously round numbers.) BLongley 00:54, 22 April 2008 (UTC)
384 seems to be 'default page count' for many books on Amazon ;) --Roglo 06:53, 22 April 2008 (UTC)
Apparently, publishers send their "forthcoming books" data to Amazon before they know how many numbered pages the book will contain, so they just use the maximum number of pages that they had contracted with the printing company for. As Bill said, this number can serve as a useful upper limit which is about 95%+ accurate, but the actual page count will be almost always different. Ahasuerus 00:10, 24 April 2008 (UTC)
As far as OCLC goes, it can be an extremely useful source of information about otherwise hard to find books, e.g. SF published by Christian publishers, especially if you use the FirstSearch interface (note that the URL has recently changed -- see Sources of Bibliographic Information for the new URL), which lets you run combined searches by publisher/year/subject/etc. Of course, OCLC has its issues, including the ones that Bill mentioned, but overall their data is pretty good. Ahasuerus 00:23, 24 April 2008 (UTC)
Two general comments about using secondary sources. First, it's important to make sure that your cut-and-paste worked correctly. I have approved and corrected a number of submissions where the Author was entered in the Title field, the Number of Pages value was missing a digit, there was extraneous punctuation at the end of the title, etc etc. One submission had three typos, all of them seemingly due to cut-and-paste errors.
Second, please make sure to record where the data comes from since it makes other editors' life much easier when two records disagree and we need to determine their source to figure out what's going on. And if you have a typo in your submission, the approving moderator will thank you for making it easy to go to the source and determine what it says :) Ahasuerus 00:23, 24 April 2008 (UTC)

2008-04-21 backup uploaded

...and now I am trying to figure out why the Wiki slows down to a crawl from time to time. Ahasuerus 05:30, 22 April 2008 (UTC)

It looks like our server is experiencing memory problems and the bad RAM module will be replaced on Friday during a brief downtime. Hopefully, it will address our performance problems! Ahasuerus 00:48, 25 April 2008 (UTC)

Strange Series Numbers

Just a heads up that one of our editors, Alibrarian (who, unfortunately, is yet to find this Wiki), apparently has a strange problem with cut-and-paste. Sometimes his submission will have a valid Series name, but then it will add a bogus 4 digit Series Number. I am not sure what may be causing these problems (a cut-and-paste from a text editor that uses Unicode?), but the solution is to remove the bogus number. Ahasuerus 00:32, 24 April 2008 (UTC)

I think it's on the wish list but having the wiki new message top banner show up in the ISFDB when you are logged in would be nice. Even if you use the wiki it can mean finding out about a message hours earlier. Dana Carson 07:34, 24 April 2008 (UTC)

2 William Blakes?

I wonder if we may be conflating 2 different William Blakes, one a famous poet and the other one a contemporary artist. Would anybody happen to know anything about the cover art that the latter Blake is presumably responsible for? Ahasuerus 02:15, 25 April 2008 (UTC)

The cover here is fragment of The Ghost of a Flea (c.1819)Tate. The Great and Secret Show by Clive Barker (cover on Amazon US) looks like fragment of Satan Arousing the Rebel Angels (1808), Bridgeman Art Library. The History of Hell (1995) - cover on Amazon + 'Search Inside' - has back cover credits 'from The Book of Urizen' if I read correctly (note that this one has also an example of 'Backcover art'). So I don't quite believe in contemporary William Blake. On the other hand, the Tate link above states that John Varley was a watercolourist, landscape designer and astrologer whom Blake met towards the end of his life. :-) --Roglo 12:33, 25 April 2008 (UTC)
Yes, William Blake was a famous poet and ARTIST. Even Wikipedia knows that! ;-) More confusing might be the 1974 poem "Infant Sorrow". It's not transposed middle digits, it's a default from the publication entry, I think. Apparently this poem is the counter poem of "Infant Joy" in the "Songs of Innocence". BLongley 18:14, 25 April 2008 (UTC)
It didn't occur to me that modern publishers might be using Blake's 200 year old paintings (or fragments thereof) as cover art, but it makes sense in retrospect! Ahasuerus 19:01, 25 April 2008 (UTC)

2008-04-27 backup file uploaded

FYI, the script that backs up the database now runs at 2am Central (US) time, which results in an effective 15 minute outage every night. Ahasuerus 22:09, 27 April 2008 (UTC)

That means I don't get my "ISFDB not responding" reminder to go to bed any more... :-( BLongley 23:28, 27 April 2008 (UTC)

"Light Novels"

I was reviewing Dissembler's last run and followed a few links to Publisher: Seven Seas. A cursory review of their 2005-2006 editions suggested that they specialize in manga and should be deleted since we don't generally list manga. However, according to their Wikipedia article they have also been translating and publishing Japanese "light novels".

Apparently, "light novels" are real novels, mostly illustrated YA versions of manga novels, so we presumably do want to list them as per our Rules of Acquisition, but the tricky part is distinguishing "light novels" from their parent manga titles. Wikipedia has a list of light novels, which may be particularly helpful since all titles that have been licensed in English are displayed in bold. If you enter these "light novels", please make sure to add a note explaining that they are real novels and not manga. Thanks! Ahasuerus 23:13, 27 April 2008 (UTC)

I like watching a Dissembler run, it's normally inspirational in some ways and distracts me from whatever rut I've got myself into. Rachel Roberts might have been left a bit incomplete and disorganized otherwise. Or David Colbert might still be a novelist. I normally leave it to Al to do the rejections, but I'll jump in and go approve/merge/serialize as appropriate. Watching what Al does to the rest is interesting too, and what he lets Dissembler do on the NEXT run as well - Dissembler seems to like some erotica that Al doesn't, for instance. I never knew we had a Kinky AI Bot... BLongley 23:53, 27 April 2008 (UTC)
Keep in mind that certain terms like "adult fantasy" don't always mean what an SF-centric bibliographer may take them to mean... Ahasuerus 00:02, 28 April 2008 (UTC)
Dissembler doesn't seem to have reached "SF-centric bibliographer" yet, but he's showing his adolescent tendencies! Now, if only he'd stop showing us the picture-books he likes... ( I mean the "Ooh, Pretty Flowers!" kind, not the Boris Vallejo books with the androgynous bums on the cover.) BLongley 00:30, 28 April 2008 (UTC)
Here's a Boris cover[1] you'll never seen on a book shelf in North America.Kraang 00:57, 28 April 2008 (UTC)
I suppose you may find it on a shelf, but then it would be in the "adult" section of the bookstore or wrapped in some kind of protective cover the way Forbidden Planet used to do it. The French definitely have a higher threshold for pictorial nudity in public places.
BTW, some eight-ten years ago I saw another type of cover that you won't find on a mass market SF book in NorAm. It was for some kind of Russian YA space opera and the cover showed an 11-12 year old boy casually holding a very big blaster-cum-machine gun. My first reaction was "Gee, no 12 year old could possibly lift that monster, much less fire it with a modicum of accuracy, unless he is a cyborg!" But verisimilitude aside, it was an interesting illustration of what's considered "child-appropriate" in different countries. Ahasuerus 02:03, 28 April 2008 (UTC)

Pseudonyms with non-Variant Titles

I have finally caught up on the sleep that I missed during the ISFDB migration (thanks for introducing me to the concept, Marc, it's great!) and wrote a new data cleanup script. You can find the results at ISFDB:Data Consistency/Pseudonyms With Titles, which is also linked from the Projects page. Basically, any Author record which is flagged as a "pseudonym" should have only variant titles associated with it, yet we have 700+ Pseudonym records that have non-variant Titles. Some of them are pretty easy to fix since they have just a couple of titles that need to be turned into variant titles or merged with pre-existing titles. Others can require quite a bit of work. And then there is "uncredited"... Ahasuerus 03:54, 28 April 2008 (UTC)

P.S. I skipped all REVIEWs since they are likely to change in the foreseeable future, but listed all pseudonymous Interviews. Ahasuerus 04:04, 28 April 2008 (UTC)
I would strongly suggest not fixing up many of the artists until we have the capacity to reverse pseudonym assignments and have verified their credits. A particular case is Emsh - as I make my way through the 50's and 60's stuff I am constantly having to change the wrongly attributed Ed Emshwiller art to the actual credit Emsh, the name with which he was overwhelmingly credited. If it's decided that the artist's real name should be used, making the assumption that he was credited with it at least a few times, that would be fine with me but that is not the current policy. As long as all credits for an individual end up in the same place it is not a big deal to me.--swfritter 17:02, 28 April 2008 (UTC)
Oh yes, artists are still in a state of flux, so we can't always tell which way the eventual pseudonym association will go. Sorry, I should have mentioned this important caveat in the original post! Ahasuerus 18:42, 28 April 2008 (UTC)
Artwork at least is easier to fix the name on, as it's rarely merged and you can update singletons easily. The Add/Remove for fiction contents is a pain especially when the page and content length is stated and the date isn't the default for the pub... I may make a feature request for that, it should be easy to put a button beside each entry to clone it to a new content section and mark the current entry as needing deletion. Actually, it's probably easy to do that as a Firefox extension, maybe plain javascript: but it would be nicer to have a list of known variants to replace a content entry with and that needs Al. BLongley 19:04, 28 April 2008 (UTC)
Anyway, I digress - can I at least ask people to check whether the reason for the lack of variants is because people feel the variation is the wrong way round, and if so pipe up and say so? It's easier to fix BEFORE clean-up. BLongley 19:04, 28 April 2008 (UTC)

Moving these discussions?

I have started to note problems found during the Pseudonym Project on the project's talk page as they seem to be rather cluttering this page. I am inclined to move these discussions -- those on specific problems, anyway -- to that page. Does anyone else have a view on this matter? -DES Talk 21:18, 1 May 2008 (UTC)

Seems like a good plan - particularly as new additions here are now not appearing at the bottom of the page, which is where most people look for new additions. (I know, past ISFDB practices are creaking as we try and use Wiki things too.) So long as the interested parties know where to look (via prompting on personal pages if necessary) we should be OK - it's getting very difficult to spot important messages via "Recent Changes" alone. We will have to learn to use Watchlists and suchlike eventually. BLongley 22:53, 1 May 2008 (UTC)
Done. These discussions have been moved to ISFDB talk:Data Consistency/Pseudonyms With Titles. Pleae watch that page or check it in some way onm a regualr basis.

Publisher:Project Gutenberg

I have reworked the Publisher:Project Gutenberg page, moving all threaded discussion to its talk page, and framing what remains as suggestions/guidelines, most of which I believe have reasonable consensus support. If you have a chance, please take a look at this page and indicate if you disagree with any of the suggested practices or contribute to any of the ongoing discussions on the talk page. -DES Talk 20:24, 28 April 2008 (UTC)

Linking templates

Pursuant to a discussion on User talk:Davecat, I have organized the templates we have for inserting links into wiki pages into Category:Linking templates. Have a look. I have created Template:Publisher to faciliate links to publisher pages. All of this is documented in Help:Linking templates and on the talk pages of the individual tempaltes. I hope this is helpful. -DES Talk 14:14, 29 April 2008 (UTC)

I have come to realize that some editors know very little about wiki-markup and how to ue the wiki features. Much of this is documented in the page Help:Editing and the other pages in Category:Wiki Help. Where should a link to Help:Editing be placed so that more people can see and make use of it if they wish to, without getting in the way? Is there any part of Help:Editing or any of the pages in Category:Wiki Help that is unclear or needs improvement? I'll be glad to do what I can to make things better. -DES Talk 15:46, 29 April 2008 (UTC)

I have also created Template:Series, for use in placing a link to an ISFDB series display on a wiki page. -DES Talk 22:48, 29 April 2008 (UTC)

Something new? Chapterbooks in authors bibliography

Just approved a submission for a new chapterbook for Patricia Wrightson and it has its own title in her bibliography[2], is this something new?Kraang 03:40, 30 April 2008 (UTC)

It's a leftover from an earlier (mid-2006) experiment when we allowed Titles to be Chapterbooks. After playing with it for a bit, we realized that the same Shortfiction Title may appear in a collection/anthology as well as in a standalone chapterbook, so there is no way to distinguish "Chapterbook Titles" from "Shortfiction Titles". We changed out policy to limit Chapterbooks to Publications and modified the software accordingly, but there is still a back door in the application that allows users to create Chapterbook titles. For now, I search for them every few weeks and change the type to Shortfiction, but hopefully a software fix (requested on Sourceforge a while back) will be available soon. Ahasuerus 04:04, 30 April 2008 (UTC)

Dates of living authors

Wikipedia has now strongly suggested that birth dates for living people be shown only as years, not as exact dates, because an exact birthdate can significantly facilitate identity theft. Should we adopt a similar policy? We are a much less visible site than wikipedia, of course. -DES Talk 11:41, 30 April 2008 (UTC)

As far as I know we have had one author ask us to remove her birth date since she had been a victim of identity theft and didn't like the experience much.
It would be easy to change the display logic not to show the exact date (unless you are a registered user and click on "Edit", I guess), but it would also affect other parts of the system, e.g. the main page, which shows the names of the writers/artists who were born/died on this day. Hm, let me leave a message on Al's Talk page... Ahasuerus 15:39, 30 April 2008 (UTC)
Authors are IMO not likely to ask for this until after they have had such experiences, when it is to some extent a case of barn door and departed horse. I was actually thinking of a script to remove all such info -- change the month and day to 00-00 -- on all authors for whom a) no death date is recorded, and b) the birth date is less than, say, 105 years ago. This would be followed by a request to editors not to enter such info in future with greater precision than the year. However, a display change would prevent such info being returned by Google and similar searches, which is probably sufficient for the purpose -- anyone seriously looking for such info can find it if we could, whether we include it or not. -DES Talk 16:06, 30 April 2008 (UTC)
Well, if we decide to remove this information from the database, then it will be likely easier to modify the filing logic to convert dates to "YYYY-00-00" years rather than rely on editors and moderators to prevent the data from being submitted. However, masking this data at display time is a gentler and easier-to-undo solution, so if we agree that we want this functionality added (or removed, as the case may be), I think we will want to consider the masking solution first. Ahasuerus 03:25, 1 May 2008 (UTC)
I presume that if we keep it and mask it it would reappear automatically when a death date is entered. So moderators would still have to be careful of Author Data edits. BLongley 18:30, 1 May 2008 (UTC)
Well, I'm generally against the wholesale slaughter of public data, and old data has a way of sticking around cached elsewhere (for instance Google still has a link to, even though we moved off there umpteen years ago). I'm also concerned that if we whack the data now, we won't be able to easily reconstruct it later. I'd support suppressing the display of the month and day for living authors, but otherwise make no changes (other than getting rid of the born on/died on section of the front page). That means that the data would be available in the downloads, and would also be visible while editing. It theoretically wouldn't show up in any search engines, would make obtaining specific dates more difficult than they are now, and allows us to display the data after they no longer care. It also presumes that the set of people interested in identity theft doesn't intersect with the set of people interested in editing SF bibliographies. On the other hand I wonder if this is much to do about nothing - the birthdates are present in a plethora of locations, which is how we obtained most of them in the first place. Alvonruff 17:57, 2 May 2008 (UTC)
Actually, my presumption would be that people interested in committing identity theft want to be able to collect lots of info in an automated manner, and won't mostly bother with things that require one-at-a-time collection over a period of time or a complex non-automatable set of searches. Anyone interested in a specific individual target can find the info whether the ISFDB includes it or not, but perhaps there is no reason to be a repository of a large number of easily stripped birth-dates. This logic suggests that the "who was born on this day" feature can be left unchanged, since it rarely has more than half a dozen or so entries, and only the display of month and day should be suppressed for living authors (with "living" being defined as something like "no valid death date entered, and birth date less than X years ago" with X sat at something like 105 or 110). Whether this is worth doing at all i can't say. Wikipedia's size and prominence as a search target means that there are pressures on it which don't apply, or at least are weaker, on us. But if even one SF author was a victim of ID theft due to the ISFDB, I suspect we would feel rotten about it. I'm sure this is legally public info, so there is no legal jeopardy, just a question of what is the right thing to do. There is such a thing as ID theft paranoia, and maybe this suggestion is an example. -DES Talk 19:13, 2 May 2008 (UTC)
Can a citation be dug up for "Wikipedia has now strongly suggested?" Is this from Jimbo Wales, the stewardship committee, or oversight committee? I'm just wondering who "Wikipedia" is. One of the concerns I've had about the birth date is I'll sometimes fix an author's date in ISFDB from Wikipedia and there's no easy way to cite the source unless I want to add an author note. I've also done the reverse is using ISFDB as a source for Wikipedia.
Anyway - This is an interesting challenge. I don't want to toss the data but agree that it's personal information about a living person. One thought is to add a boolean flag, bShowDOB, that could be enabled if an author discloses the full DOB on his or her own web site. bShowDOB would be off by default and editors would only turn it on if the DOB is cited on the author's web site. When bShowDOB is off we would only show the year in displays and the author edit would have YYYY-XX-XX with the "XX-XX" indicating the system does have a full DOB but it's not disclosed. An editor would be free to enter a DOB and if it's different than the existing DOB the moderator will see it.
When bShowDB is enabled by an edit we should have a yellow warning box on the moderator screen that would also have the author's web site info (both old and new if they are different) available as clickable links so that the moderator can verify that the author has chosen to dislose their DOB. Marc Kupper (talk) 09:12, 3 May 2008 (UTC)
As to the citation, see Wikipedia:Wikipedia:Biographies of living persons#Privacy of personal information, an offical policy page, which says, in part: "Wikipedia includes dates of birth for some well-known living persons where the dates have been widely published, but editors should exercise caution with less notable people. With identity theft on the rise, people increasingly regard their dates of birth as private. When in doubt about the notability of the subject, or if the subject complains about the publication of his or her date of birth, err on the side of caution and simply list the year of birth.". In practice this is being taken as warrent for the removal of precise birth dates from articles about the vast majority of living persons.
As to your proposal, I think it is more mechanism and more maintanance effort than the subject warrents, but it would do the job. Assuming we want to do this at all, and that we want it done in software (which would be required to mask dates but retain them) I think the simpler rule of "no death date, brith date within living memory, don't display" would save a lot of work. But your idea would be a bit more flexable. -DES Talk 11:04, 3 May 2008 (UTC)

Magazine Projects Page

I intend to create a project page for magazines so that anyone editing magazines will have a central location to document their plans and view the plans of others. This follows a recent event where two editors were working on the same mag and made concurrent submissions that had a predictably unpleasant result. It would also be a good place to document issues that are specific to magazines. I do not want to further bloat Help with issues that are rarely of concern when editing book data.--swfritter 21:04, 1 May 2008 (UTC)

Sounds like a good idea, but a brief pointer to that page in the help pages might be useful. -DES Talk 21:21, 1 May 2008 (UTC)
As I understand Bob's posting, the collision was in wiki editing, not DB submissions. That's not to say that your magazine project page isn't a good idea; we're sure to run up against collisions there sometimes. Dave (davecat) 15:13, 2 May 2008 (UTC)
Wiki-editing collisions can actually be recoverd from fairly easily, once you know how. See Help:Edit conflict. -DES Talk 15:36, 2 May 2008 (UTC)
That is where the collision was but that is not why I think a magazine project page would be useful. The current help pages allow choices on some data entries. I would like to be able to discuss preferences among those of us who do a lot of magazine editing only to have someone else change an entry to another acceptable alternative. It would allow us to spend more time entering new data and less time revising it. We could also address other issues that are of little interest to non magazine editors as well and list magazines/date ranges we are working on to get better coverage. I think there are a lot of positives. Thx, rbh (Bob) 22:52, 2 May 2008 (UTC)
That sounds appropriate and useful to me, just wanted to be sure that you know how to deal with a wiki edit conflict if one should occur again. -DES Talk 01:03, 3 May 2008 (UTC)

Update needed to Privacy Policy

ISFDB:Privacy policy#Policy on release of data derived from page logs currently says: "As ISFDB staff do not have access to the page logs, all policies relevant to the release of page log data are controlled by Texas A&M University."

Obviously, this is out of date and should be changed. I don't know who does now have access, and exactly what the policy should now say -- clearly log info is not distributed widely, but I'm not sure exactly what, if anything it is being or might legitimately be used for. -DES Talk 15:53, 2 May 2008 (UTC)

I've added some text to cover the post-TAMU situation. Alvonruff 22:53, 3 May 2008 (UTC)

Publisher Editing

I've added a first pass at publisher editing. This allows adding/modifying a Wikipedia article link, adding/modifying web site links associated with the publisher, and adding notes that will be displayed on the publisher bibliography page. More dramatically, you can also edit the publisher's canonical name - once done this applies to all publications that are linked to that particular publisher. As such, the editing feature only shows up in the navbar for moderators.

Note that you still can't merge any publishers, so if someone accidentally changes a publisher to something undesirable (like changing Ace to Tor), it is reversible. Alvonruff 22:56, 3 May 2008 (UTC)

Thanks, that sounds helpful. Do I correctly understand that if someone changes, say, "Baen Books" to "Baen" we will have two (or more) different publisher records with identical names? Do we plan that a publisher merging function will be added in due course?
It seems to me that this facility means that we ought to restart discussion on cannonical names or forms of names for publishers. But it would help if we knew what else is in the works near term. in particular, are there plans to add an "imprint" field separate from publisher, or is the publisher record intended to really be a publisher/imprint record, whenever imprints are significant enough to distinguish. -DES Talk 00:09, 4 May 2008 (UTC)
At the moment when you change an alternate name to an existing name they do not merge, you just end up with two identical names.Kraang 00:53, 4 May 2008 (UTC)
I discovered that after a few test changes. I'm sure Al will be able to merge publishers eventually (or give us moderators the tools to do so). Changing publishers' names wholesale is an awesome power that must be used judiciously in order not to create damage that could potentially take hundreds of hours' work to undo. MHHutchins 01:48, 4 May 2008 (UTC)
I'm in complete agreement with Michael, merging can be helpful but if done without an in-depth check of the publishers it would be quite a chore to undo. In my cleanup of publishers names I've already found two different publishers with the same name(small number of books) who are separated by a few decades and I'm only in the "A's".Kraang 02:34, 4 May 2008 (UTC)
I absolutely agree, an injudicious publisher merge has the potential to be even more trouble than an unwise author merge. But like the author merge, it will be a very helpful tool when used properly. -DES Talk 03:41, 4 May 2008 (UTC)
I can place some restrictions on publisher merges. For instance, perhaps you can only merge publishers with identical names. That would require 2 deliberate steps to merge: first altering a publisher to a new name, then merging the result. Alvonruff 17:32, 4 May 2008 (UTC)
Well, we could police ourselves for a bit first - leave Publisher changes for other Mods to approve, even if we could approve them ourselves. That way we'd at least occasionally see what people were doing and which direction they're pulling in: e.g. I'd have no objection approving "Futura Orbit" and "Futura/Orbit" and "Orbit/Futura" merges and at least I could see who it is that feels that we need to record more than "Orbit" on every publication (the imprint is quite informative enough for me and ownership of Orbit can be recorded elsewhere, IMNSHO). Or which people are trying to do Imprint/Publisher and which are going for Publisher/Imprint and messing up my plans either way. BLongley 18:55, 4 May 2008 (UTC)
I'd like some standards discussions restarted - e.g. we can reduce the numbers of publishers a lot if we agree on regularization of "and Company", "and Co", "and Co.", "& Co." and "& Co" for instance. I've fixed a lot of those just by merciless updates of all the unverified pubs under the variations with the fewest pubs, but when there's scores of one and dozens of the other then I've just redirected the Wiki entries to the leading one for now. This in no way constitutes an endorsement of that version though. But we don't seem to be talking about the way forward enough, and at times I think we're already working AGAINST each other, which has to be stopped ASAP. BLongley 18:55, 4 May 2008 (UTC)
I fully agree. Standards are needed here. once we agree on them, publsiher editing will alow us to implement them far more easily than we could have before. I also think that Al's suggestion of forcing a publisher merge to be a two step process might have merit, but I'm not sure. -DES Talk 22:20, 4 May 2008 (UTC)
I also agree, the statement in the help "The publisher is not a key entity in the ISFDB, so you are free to choose an imprint ("Ace Books"), a division ("Berkley") or the parent corporation ("Penguin Group USA") as you wish." means that I for one, have not been trying to ensure that the etries I edit are consistent. I like the idea of a consistent set of publisher names. Thx, rbh (Bob) 02:05, 5 May 2008 (UTC)
OK, the help could do with updating to make it clear we're TRYING to regularize a bit, and/or make the Publisher/Imprint data useful. I lean towards the Imprint being most important and the Publisher is going to be messy nowadays anyway - "A is now an imprint of B which is a division of C who are part of the D conglomerate..." How much we want to record on each pub and how much can be moved to the Publisher pages is still unclear (that will depend on Al's design changes) but we can generally work toward what we each want for a particular "Publisher". The BIG "Publishers" are going to be harder to separate into useful data than it will be to regularize several versions of smaller publishers that really are the same. But just keep trying and when you encounter a verified pub with a Publisher name you aren't keen on, start talking! BLongley 22:12, 5 May 2008 (UTC)
I just edited the help to start heading in this direction, see Template:PublicationFields:Publisher and consider if my changes look reasonable. -DES Talk 16:41, 8 May 2008 (UTC)

(Undent) I always like to start by abstracting the real-world rules and see how that impacts record relationships, database layout, and display rules. So here's what I think I know:

  • Publisher names are generally recorded in a variety of databases with less care and regularity than author names. This suggests that we need to pick publisher canonical names, just as we do with authors. Publications which clearly vary textually from the canonical name should be documented in the notes section of the publication. I would suggest making a single wiki page that lists the agreed canonical publisher names, and shows their known variant spellings.
  • Publishers are generally legally recognized companies and corporations, and as such have legally recognized names. This suggests that if we are undecided as to the correct canonical name, there is a resource we can turn to.
  • Publishers may, over time, change their company and associated legally recognized names. I personally don't think we should have variant publishers. We can capture company history in the wiki, and leave the publishers as separate entities.
  • Publishers may be created, or sold to other publishers as an imprint. This suggests that we need imprint support. Imprints of a publisher should be displayed on the parent publisher bibliography as clickable links. We need to decide if books published under an imprint should be displayed with books published under the parent company. My leaning is: no.
  • An imprint may or may not have a legally recognized name, depending upon the business structure under which it was created. This means we'll have to decide on canonical names for imprints.
  • Imprints can have a history of multiple owners. This means that an imprint can have multiple parent publishers. The geneology of the imprint should be recorded on its wiki page - it may be too complicated to describe lineage in an SQL database.
  • Imprints can change their names. As with the publisher case, I think these should be handled as separate entities, and documented on the wiki pages. I'm not in support of variant imprints.

That's my first pass at the problem. Alvonruff 15:46, 8 May 2008 (UTC)

I tend to agree with the above, given a few questions and caveats.
  • When you say "we need imprint support", would this mean an "Imprint" field on publication records to be used going forward? If so, the question of how to fill this for existing records arises, but I can see several possible answers. I am in favor of this if it is achievable in a reasonable way and timeframe.
  • This would mean that pubs would record an imprint and a publisher. If a user wants to know what the further corporate structure was then s/he would need to consult the relevant wiki page (or other sources) based on the publisher name, the imprint name, and the date of publication. For most modern cases this seems reasonable to me. However...
  • You say you are opposed to having "variant publishers". But when dealing with relatively early works (as I have been doing in some cases through the original editions of works transcribed by Project Gutenberg) one gets publisher names no longer in use, where the name itself is a significant clue to dating. For example "Charles Scribner's Sons" or "Henry Holt and Co". I would be somewhat reluctant to list these works under the modern Scribners or Holt, but perhaps that is exactly what should be done, as i see more than 25 Publisher names are in our DB at the moment of which "Scribner" is a sub-string, and several of these list publications covering most of the 20th C and before or after.
  • Building on this, and in the interest of searchability, what do you think of the following idea. Rather than relegate the exact publisher name to the notes field, include two fields: "Listed Publisher" and "Canonical Publisher". "Listed" would be exactly what is in the publication or other source. Similarly we could have "Listed Imprint" and "Canonical Imprint" but I think that is less important, as it seems to me that imprints don't change their names as often as publishers do, although the do get sold and moved around the corporate structure.
I hope the above comments are helpful. -DES Talk 16:35, 8 May 2008 (UTC)
I'll just comment on the imprint question for now: my thought was to leave the publication fields untouched. There would instead be an imprint table containing parent/child relationships. You can quickly find if a publisher is an imprint by searching the child field. You can easily find all the imprints of a publisher by searching on the parent field. When printing a publisher's bibliography, there would be an Imprints section which would contain the names of all publishers which are imprints of there current publisher. Likewise there could be an Imprint Of section that would list the parent publishers the imprint has belonged to.
Starting with the publication then, the user could click on the publisher link and go to the publisher bibliography, which let's say it's an imprint like Forge. The bibliography will display the parent publisher, and the user may click on that and go to the parent publisher, say Tor. That bibliography would display all it's known imprints, including perhaps Tor Fantasy, and so on. So in this case, the publication record only records the publisher. Alvonruff 16:58, 8 May 2008 (UTC)
Interesting design, I can see it would be less work to implement than what I had suggested. This design implies that if there is an imprint the imprint, and the imprint alone, should always be entered in the "publisher" field, because from the imprint you can find the publisher (at least given the date), but from the publisher alone there is no way to find the imprint, if the publisher has more than one imprint.
I see some problems with this in practice:
  • If a new imprint is found, how is it linked with the proper publisher if the person who enters the data enters only the imprint (as the above logic suggests would be normal)
  • If it is unclear or unknown when an imprint changed ownership, the data in the individual publication records (which will not record both publisher and imprint) will not be a data point to help resolve the issue. How will such issues ever be resolved? Only through outside research sources?
  • If it is unclear or unknown when an imprint changed ownership, who will a user be able to follow the chain from publication to imprint to publisher? presumably there will be multiple parent publisher links to follow, how does the user know which one to pick?
  • If the date of the publication is unclear, and the imprint has had multiple parent publishers, the imprint/publisher data will not be available to help pin the date down, because (following the above logic) the publication record will record only the imprint, not the publisher. Also, in such a case, how will the user ever determine the correct publisher from among the various parent publishers of the stated imprint?
  • When an editor is entering a new publication that lists both an imprint and a publisher, what if anything does s/he do with the publisher data?
The above points make me think that it would be useful if publication records somehow separately encoded both imprint and publisher, if that is feasible. If we once implement the kind of design described, I think that becomes less feasible in future than at present. -DES Talk 17:26, 8 May 2008 (UTC)

Error on adding web page to author data

Tried to add some web pages to the John Dalmas author data. The first one seemed to add without an error. When I went to add a second it seemed fine until I approved the edit. Then I get an error. The entry I'm approving is 969268 and the error is

 --> -->
<type 'exceptions.TypeError'>	Python 2.5: /usr/bin/python
Mon May 5 01:35:58 2008

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.
 /var/www/cgi-bin/mod/aa_update.cgi in ()
  193                         while record:
  194                                 if from_value == '':
  195                                         from_value += record[0][2]
  196                                 else:
  197                                         from_value += ","+record[0][2]
from_value = '', record = ((5233L, 1443L, None, ''),)

<type 'exceptions.TypeError'>: cannot concatenate 'str' and 'NoneType' objects
This is an issue related to a change I made to the webpages table over the weekend. Will fix this evening. Alvonruff 13:26, 5 May 2008 (UTC)
Fixed. Alvonruff 23:06, 5 May 2008 (UTC)
Thanks Dana Carson 19:43, 8 May 2008 (UTC)

Nominating DESiegel60 for moderatorship

Ref: Moderator Qualifications#Becoming a moderator for the nomination process.

Nomination statement

At the request of Ahasuerus, DESiegel60 (talkcontribs1) is nominated for moderatorship; he has accepted the nomination. DESiegel60 has 1305 edits and counting. At the request of the currently traveling/sick/tired Ahasuerus I place his nomination--Alvonruff 17:14, 6 May 2008 (UTC)


  • Support as the nominator. David has an extensive technical and Wiki administration background, which he has put to good use here, creating various and sundry templates and improving our Help pages. He has tackled numerous (often rather complex) areas of the application and has actively participated in policy discussions. Over the last few months he has acquired a working understanding of the underlying database structures, which I believe makes him ready for moderatorship. Ahasuerus 22:44, 6 May 2008 (UTC)
  • Support Most of David's submissions have been without issue and any that I communicated with him about were quickly resolved. Learning how the database works is about communication and interaction with others and David has been a very active participant.Kraang 00:53, 7 May 2008 (UTC)
  • Support Very solid rate of approvable submissions. Dana Carson 02:35, 7 May 2008 (UTC)
  • Support Very active in both submissions & discussion, good understanding of what we're trying to do. Work seems to be of high quality. Dave (davecat) 13:56, 8 May 2008 (UTC)
  • Support While I don't see his work, I often see his comments and he has provided me a few helpful hints as well. rbh (Bob) 21:48, 8 May 2008 (UTC)
  • Support David has shown a lot of initiative and the willingness to put his ideas into action. Thanks for the Project Gutenberg standards.--swfritter 22:30, 8 May 2008 (UTC)
  • Support Those submissions of David's that I've handled have consistently been good ones, and he's very communicative as well as being a master of the wiki. MHHutchins 23:45, 8 May 2008 (UTC)



  • Neutral I'm abstaining as David works in areas that I am mostly unfamiliar with.--Rkihara 02:15, 8 May 2008 (UTC)


Nomination passes, moderator flag previously set to facilitate David's work on the Wiki side. Congratulations! Ahasuerus 01:23, 12 May 2008 (UTC)

Response Thank you to all who supported me. I am honored. I will do my best to enhance the project. And a particularly large thank you to Al, without whose work there would be no project for me to be part of. -DES Talk 15:07, 12 May 2008 (UTC)

Congratulations - and now get back to editing. :-) I had not had time recently to look at the community portal and so missed the election. Marc Kupper (talk) 07:12, 13 May 2008 (UTC)

Nominating Hall3730for moderatorship

Ref: Moderator Qualifications#Becoming a moderator for the nomination process.

Nomination statement

At the request of Ahasuerus, Hall3730 (talkcontribs1) is nominated for moderatorship; he has accepted the nomination. Hall3730 has 3424 edits and counting (good enough for 13th place in total contributions). At the request of the currently traveling/sick/tired Ahasuerus I place his nomination--Alvonruff 17:14, 6 May 2008 (UTC)


  • Support as the nominator. Bob has been doggedly working on magazines for many months and has required almost no help with his submissions lately. He also participates in policy discussions on the Wiki side. I believe that Bob will be able to function independently in the magazine arena, although he will probably want to be a little more careful on the book side, which has its own gotchas. Ahasuerus 22:23, 6 May 2008 (UTC)
  • Support Not my area of interest but I do follow the talk pages and Bob seems up to the task of self moderation.Kraang 00:24, 7 May 2008 (UTC)
  • Support Not my area of expertise but almost all the ones I've seen were obviously correct and I approved them. Dana Carson 02:33, 7 May 2008 (UTC)
  • Support Bob has a lot of experience, and I've found no fault with his submissions.--Rkihara 01:39, 8 May 2008 (UTC)
  • Support There was a time I thought I was going to have to do the magazines all on my own. Bob will make another fine addition.--swfritter 22:36, 8 May 2008 (UTC)
  • Support Those submissions of Bob's that I've handled have been mostly in the magazine side of the db, so I'm sure he'll be able to approve those submissions. It shouldn't take long before he's broadened his area of knowledge into other areas. MHHutchins 23:49, 8 May 2008 (UTC)



  • Neutral leaning toward support I haven't seen enough of Bob's work, but what I have has been good. Dave (davecat) 14:05, 8 May 2008 (UTC)


Nomination passes, moderator flag set. Congratulations, Bob, approve away! :) Ahasuerus 01:20, 12 May 2008 (UTC)

2008-05-06 backup posted

Can anyone think of a reason not to delete the old versions of all Wiki pages? They are getting close to 2Gb, which makes it ever more time consuming to generate the public version of the backup file that I post weekly. Ahasuerus 04:02, 7 May 2008 (UTC)

It can be sometimes helpful to go into the page history and see when or by who something was written, or how a page came to have the shape it does, and in an ideal world I'd say retain it all. But I doubt that there is anything really vital lurking in the old versions. Is it possible to retain old versions that are not older than X, where X is say a couple of months? or is it an all or nothing process? I have found that in doing the archiving of the verification requests old versions were helpful to sort things out in a few cases. But again, not vital. -DES Talk 04:35, 7 May 2008 (UTC)
I am afraid I am not sure. Let me leave a message on Al's Talk page and see what he discovered when he looked into this issue a few weeks ago... Ahasuerus 05:01, 7 May 2008 (UTC)
Well, it's quite easy to delete all but the most recent page (there's an existing PHP script to do that). That reduced the size down radically on isfdb2, but also got rid of too much history. I'd rather save the last N versions of a particular page, rather than deleting everything older than X, as there may be some pages where we wind up with only a single page, due to it not being updated for some time. Since every update saves a completely new version of the page, there are some large pages which have been updated numerous times (like the DAW and some magazine pages), causing them to use quite a bit of storage. As such I had the following in mind:
  • Pareto the current wiki pages. Create a tool that determines the amount of space use by each wiki page and all its old revisions.
  • Create a general tool for pruning a specific wiki page.
  • Attack the big resource hogs, and not worry to much about the little guys.
Alvonruff 11:53, 7 May 2008 (UTC)
That seems to make good sense to me, if it isn't too much work. It also seems to me that we could probably safely drop any old revison marked as minor, if it is easy to sort those out. -DES Talk 14:29, 7 May 2008 (UTC)
Anything that makes the backup file smaller is fine by me! :) Ahasuerus 15:52, 7 May 2008 (UTC)
I don't see a reason to include the edit history in the public backups. I could see making private backups of the entire db but see the public backups as providing a snapshot of where things are at now. If someone has a need for a backup that contains historical data, either on the wiki or db side, then I could see constructing a backup for that project. Marc Kupper (talk) 07:19, 13 May 2008 (UTC)

Creating a standard "request for deep linking" letter/e-mail

As per this ongoing discussion about "deep linking" to other sites' images:

In the past, it was usually individual editors who would send ad hoc requests to various Web site administrators and get their permission to link to their images. However, going forward, it would be probably easier to have a standard "deep linking request" form letter, which we could then translate into French and any other language that we may need. As an added benefit, we can make double sure that the form letter makes it very clear what kind of linking we intend to do, which may help avoid misunderstandings down the road.

Any volunteers? :) Ahasuerus 05:01, 7 May 2008 (UTC)

I'll give it a whirl. -DES Talk 14:30, 7 May 2008 (UTC)
How about this for a first draft:
Dear <web site operator>
I am a volunteer with the Internet Speculative Fiction Data Base (ISFDB) which is located at The ISFDB displays information about the titles, authors, and publications of works of Science Fiction, Fantasy, Horror, and related works of speculative fiction. When we display information about a particular publication (an edition or printing of a particular work) we prefer to display an image of the cover of the work, so that the public can see exactly which edition is being described. We do this by including a URL of a publicly posted image in the database. When the publication page is displayed, using the image found at that URL as an image source for the page display. An example is for an edition of The Lord of the Rings. We also link to images of authors in the same way. However, we only use URLs from sites that have agreed to permit us to do so.
Your site <name of site>, located at <base URL>, has images that would enhance the ISFDB. We would greatly appreciate it if you would permit us to link to those images. The ISFDB is a completely non-profit, volunteer-run organization. We think that it does provide a valuable public service, and does help promote speculative fiction in general. Please let us know if we may display images from your site on our pages by linking to your site. Please copy your response to <official email address> so that the ISFDB as an organization can preserve a record of it.
Thank you
<signature of volunteer>

Can we offer any sort of credit or link-back to sites granting permission? Can we have a page that says:
"Images provided courtesy of
  • <site 1>
  • <site 2>
  • <site 3>
or something of the sort. Or perhaps the pub display page could actually provide a credit line "Image courtesy of" and the domain name from the URL, or a description based on a table lookup of the domain? That would take coding, I'm sure, but a credit line might well help induce image providers to agree, and if we can offer it, it should be mentioned in any standard request letter.
This is just a draft, I hope it is helpful. Fell free to propose minor or drastic changes, or criticize it in any way at all. -DES Talk 14:52, 7 May 2008 (UTC)
It seems pretty clear and well stated, now all we need is it translated into French, German, Spanish.....Kraang 00:35, 8 May 2008 (UTC)
I have no created ISFDB:Image linking permissions. Any translations or related text should be placed there, and any discussion of what should go in our standard letter should be done at ISFDB talk:Image linking permissions. To support this, I created Template:Image Host Sites, which is now used on that page and in Template:PublicationFields:ImageURL, and through that template, on several help screen pages. -DES Talk 21:44, 9 May 2008 (UTC)
Can you remove the French van Vogt site for the moment please? I think you misread Author:A._E._van_Vogt, I only acquired permission for THREE of the four sites. Mind you, the fourth would be a great test of Circeus' French version! BLongley 21:49, 11 May 2008 (UTC)
Done, info moved to the template's talk page until/unless permissison is obtained. -DES Talk 15:23, 12 May 2008 (UTC)
It occurs to me that having a version of this letter in Japanese/Klingon won't do us much good unless we have somebody who can read Japanese/Klingon and interpret responses. Should we have a list of editors who can help with foreign language responses and/or willing to send the letter out on the ISFDB's behalf? Ahasuerus 22:29, 9 May 2008 (UTC)
Should we unverify The Klingon Dictionary before your example lands us with MORE work, Ahasuerus? Although if pushed, I might be able to track the creator down, he was a fellow alt.callahans poster for a while. BLongley 21:58, 11 May 2008 (UTC)
True enough. such a list might well be a good idea, but then a list of editors with the ability to read and understand various languages might also help in dealing with publications in those languages. Should there be a separate list for the purpose of permission requests? if there should, perhaps ISFDB:Image linking permissions would be a suitable place for such a list. If not, perhaps ISFDB:List of editors with skills in non-English languages or some such name? -DES Talk 23:06, 9 May 2008 (UTC)
I think we could use both. There seem to be no volunteers for "<official email address>" so it may be that we have to have several semi-official email addresses - which reminds me, if there IS an official repository of permissions then I have three emails to forward. But I'm not volunteering for Speaker-to-Japanese or Speaker-to-Icelanders or Speaker-to-Americans in general: none of them can be guaranteed to speak the Queen's English intelligibly. ;-) But if people want to list their multi-lingual skills it would help me interpret the replies I get to some of my other attempts at Alien Contact. BLongley 22:24, 11 May 2008 (UTC)

I'll try to work a translation soon. Prod me in a few days if I haven't. Circeus 00:30, 10 May 2008 (UTC)

Comment: I think "The ISFDB displays information about the titles, authors, and publications of works [...]" should be "The ISFDB is a bibliographic database of works [...]". Circeus 20:21, 11 May 2008 (UTC) Posted my changes on the relevantr page.
Here goes my French version. I adjusted it to take into account that such emails are not usually adressed to anybody specific in French:
Je suis bénévole à l'Internet Speculative Fiction Data Base (ISFDB, L'ISFDB est une base de données bibliographiques sur les oeuvres de science-fiction, de fantasy, d'horreur et autres genres apparentés. Lorsque nous affichons les informations relevant d'une édition ou réimpression précise, nous affichons lorsque possible une image de la couverture, afin de rendre plus explicite l'édition concernée. Cela se fait en incluant dans la base de données une URL vers une image accessible publiquement sur Internet. Cette image est alors affichée sur la page de l'Édition, avec un lien HTML vers son emplacement originel, ainsi la page est celle d'une édition du Seigneur des Anneaux avec son image. Le même système est employé pour inclure des portraits d'auteurs. Toutefois, nous ne lions qu'à des images de sites qui nous en donnent la permission.
Votre site (<base URL>), a de nombreuses images qui seraient d'une grande utilité à l'ISFDB, et nous apprécierions grandement la possibilité de les employer. L'ISFDB est une organisation à but non-lucratif entièrement gérée par des bénévoles. Nous croyons qu'il s'agit d'un service utile qui appuie la promotion de la litérature spéculative. Nous apprécierions de savoir s'il nous est possible d'envoyer des liens vers les images de votre site. Une réponse est <official email address> serait particulièrement utile afin que l'ISFDB en conserve une copie.
If the website has an explicit name (e.g. as opposed to being a company website), the outlined part can be instead worded as "Le site <name of website>", otherwise, it's just assumed to be some company's official website. Circeus 20:41, 11 May 2008 (UTC)

The Dying Earth

I would like to change The Dying Earth and all its publications from a Novel to a Collection. I don't think there can be much argument here, this isn't even a fix up, or a collection with linking text. This a group of stories, many published separately at one time or another, that retain separate titles, that in some cases don't even have common characters. What they have is a shared world, and a few cross-references. There is even a title-level note that says "Note: Slightly connected series of stories" and then lists them. The problem is that there are three publications of TDE that are verified as novels, plus three verified publications of omnibus titles that include TDE. Multiple verifiers are involved. What is the best way to proceed here? -DES Talk 15:36, 7 May 2008 (UTC) I have notified all the verifiers involved of this posting. -DES Talk 15:54, 7 May 2008 (UTC)

I'm not bothered either way really, I've not read it. My copy is definitely described as a novel but there are six chapters each with a sub-title correlating with the titles listed at title level. "Ulan Dhor Ends a Dream" is just "Ulan Dhor" though, if someone's going to change my copy: if so let me know and I'll add page numbers. BLongley 17:57, 7 May 2008 (UTC)
I think if you were to read it (which if you like early Vance at all i would reccomend, it is quite good IMO) you would agree. Note such lists of separate publications as "Mazirian the Magician", "Liane the Wayfarer" (aka "The Loom of Darkness" ), "Guyal of Sfere", and to a lesser extent "Turjan of Miir". Note that among these several later Vance collections (such as Eight Fantasms and Magics, The Jack Vance Treasury, andGreen Magic ) reprint only individual stories from TDE. Well, perhaps I'm going on longer than this warrants. -DES Talk 19:56, 7 May 2008 (UTC)
I agree on both counts. It's a really good example of Vance's early work and the stories are all basically standalones, although some characters and events are cross-referenced. Some sources call it a novel and others call it a collection of linked stories, but for our purposes it's easier to handle as a collection. Ahasuerus 20:05, 7 May 2008 (UTC)
OK, count me as generally "FOR" then, based on separate publications of constituent titles elsewhere. The work that someone put into the title record makes me think that someone has disagreed in the past - there are people here that seem to insist that if a book SAYS it's a Novel, then it should be counted as one: even if someone else checks chapters against other Shortfiction publications and finds them all substantially the same. I've given up on Times without Number, where we seem to have got the worst possible combination of solutions: and I've given up on many A. E. van Vogt titles too. I can always remove my verification if I don't like the end result. BLongley 21:28, 7 May 2008 (UTC)
As to reading it - I might well do so soon, it's one of the smaller books I'm being recommended at the moment, and if it really is a collection I should be able to read most of it during PC start-up times at work with very little interruption. (10-15 minutes where everybody KNOWS that you're not slacking off, boot-times really ARE that bad! Apparently it's currently more acceptable to pay someone £20 to sit there waiting for it than 5p for the electricity saved by switching off each day.) BLongley 21:28, 7 May 2008 (UTC)
10-15 minute boots, arrgh. It would probably be cheaper yet to pay someone minimum wage to come through and just switch on everyone's machine 20 mins before the usual start of business, but that would make too much sense. If you do read TDE, I'd be interested in what you think of it. -DES Talk 21:51, 7 May 2008 (UTC)
Just keep in mind that some stories in The Dying Earth (especially the last one, if memory serves) are rather pulpy. Nothing wrong with that in my book, but different readers have different "pulp tolerance thresholds". Ahasuerus 22:10, 7 May 2008 (UTC)
Changing it from "novel" to "collection" is fine by me, since I've never read the book and my verification was fairly early after I arrived here. I went with how it was originally described and never examined the book any further.Kraang 00:28, 8 May 2008 (UTC)
Very well, there seems general approval. I have started to make the changes. -DES Talk 16:59, 8 May 2008 (UTC)
Changes done. -DES Talk
Well, except that i have not added the individual stories to the contents lists of editions verifed by other editors -- should I do so? -DES Talk 22:13, 10 May 2008 (UTC)
I did mine. Pagination added. Also Variant created, contents added, title record updated, reverified one copy, not sure if I still have the other. Don't we normally put a Collection's contents into the Omnibuses as well though? I know I've skipped that step in the past, and if that's acceptable now please do say. Otherwise I'll drag out Tales of the Dying Earth tomorrow and add more pagination to that. (I haven't had it leave the house just yet.) But I have a bath run now and can actually attempt reading the book at last! BLongley 22:27, 10 May 2008 (UTC)
I've run across a few books that had named chapters and appeared to be collections. With the last one (I can't recall the title) I ended up making it a novel and adding a note explaining the named titles and that some people viewed it as a collection. My reasoning for keeping it as a novel in ISFDB was that none of the stories had ever been published separately meaning it was unlikely someone would ever search for the individual stories by title though if they used Google then the note I wrote should show up. Marc Kupper (talk) 07:30, 13 May 2008 (UTC)

Collections in Omnibuses

This section broken out from the discussion of The Dying Earth which spawned it.

I was under the impression that for an omnibus that included both one or more novels and a collection we typically did not list the individual shortfiction works in the collection, particularly not if the collection had previously been published separately and has that publication listed with contents. But I am not sure if we have ever really codified this policy one way or the other. (if an omnibus is a collection of collections I suspect we would just treat it as a different, larger, collection). -DES Talk 22:42, 10 May 2008 (UTC)

I was thinking about this the other day. I see two possible problems with using Collection Titles -- as opposed to using their constituent stories -- in Omnibuses. First, if you are looking at an Omnibus page and it contains a Collection, you need to click 2 times to get to the Collection's contents, once on the Collection's Title link and once on a Collection Publication link. Worse, once you click on the Collection Title link, you are then faced with anywhere from 0 to dozens of Publication records. And, of course, each Collection Publication may have somewhat different contents -- essays, forewords, afterwords, etc -- so how do you know what exactly the Omnibus contains? Second, if you are looking at the Title page of a story published in a Collection that later appeared in an Omnibus, you will not see that Omnibus publication, which defeats the purpose of listing all Publications for that story. It seems safer by far to list the constituent stories and avoid these problems. Ahasuerus 23:12, 10 May 2008 (UTC)
I see. A problen is that there is no concept of hierarchy to the contents of a publication, including an omnibus. If one could list both the collection title and its contents in such a way as to make clear what belongs to the collection, it would be much better. There are other cases where this would be helepful, too. For example it is not uncommon for a collection to contain groups of stories with a common group title, as well as individual titles. "X: Three tales" consistign of "A", "B", and "C". without a "group title" element, one is tempted to reder these as "X:A", "X:B", and "X:C". Sometimes this can be done with a series, as in Distant Friends, but this doesnt always work well. In any case, we have often entered omnibus pubs that contain collections without the short fiction -- Indeed I rather thought it a rule that no title in an omnibus could be of type short fiction. For example, see Mountain Magic which contains the collection Old Nathan or 3xT which contains two collections and a novel, but lists no shortfiction (and if it did, how would the user know which short works were part of which collection?) or Annals of the Time Patrol which contians two collections, and many in simialr state. If we decide that the individual works of short fiction should be enterd, i will do that. Help:Screen:NewPub is not very helpful on this point at present. -DES Talk 00:05, 11 May 2008 (UTC)
I think we have to enter the short-fiction. :-( More work, I know. Reasons: 1) We'll never be sure WHICH version of the collection is included, so which variant titles etc are in it. Or even contents: some collections are different in different countries due to having to split them up into separate volumes, or because one or two stories aren't free to be published in that country. (I have MANY books that say "For copyright reasons, this book is not available in the USA" - although Canada normally gets permitted.) 2) People use the pagination for Length-determination at times - personally, I'm not that bothered, but you can't get it from Magazine versions. E.g Story X may be entered as starting on page 22, another story starts on page 25, so it's 3 pages long and therefore Short Story? Not if it's "Continued on pages 125-148" - which those lazy Magazine editors never mention. ;-) It's rare (but not unknown) that a Book publication splits contents that way. ("Destinies" comes to mind though, as a Magazine in paperback format that we've had to deal with as "Anthology" - not sure if it ever did such magazine content-splitting though.) I think Ace Doubles of a pair of Collections have been the worst to deal with so far - workarounds such as Roman Numerals for one side and Arabic for the other have been suggested, but as usual never made official. BLongley 22:58, 11 May 2008 (UTC)
I see your point, and perhaps it is the best way, but it seems to me that at the least the collection titles ought to be entered in addition to the shortfiction titles. If full page numbers are entered, this will make it at least fairly clear what stories belong to what collection title. But note, lots of omnibuses are not entered in this way, even ones that are verified, and if this is to be the standard, the help should be changed. I am going to raise this on the rules and standards page, I think. -DES Talk 15:33, 12 May 2008 (UTC)
I'd favor entering both the collection & its contents. (If you did enter something on rules&standards I don't see it, so I'm sticking my oar in here.) A problem with it, which sometimes shows up in other places (book review columns, say) is that we have no way of forcing two titles with the same page number to display in a particular order. But that's relatively minor.
The ideal solution would be a whole slew of programming changes, with some database changes involved, which would always expand collections in searches & displays. I kind of shudder at the changes needed, though. -- Dave (davecat) 16:47, 12 May 2008 (UTC)
work called, that entry is comming soon to a wiki-page near you. -DES Talk 20:40, 12 May 2008 (UTC)
Oh, for a real horror of "continued stories" in books see the Merovingian Nights series of anthologies, which routinely had multiple stories continued multiple times in one book. At least the continuations wwere listed in the ToC. I need to check the state of those in the ISFDB soon -- I do own copies. -DES Talk 15:38, 12 May 2008 (UTC)

(unindent) I have copied most of this thread, and continued the discussion, at Rules and standards discussions#Collections in Omnibuses. -DES Talk 21:19, 12 May 2008 (UTC)


I was wondering what technical/philosophical reasons there were to only link to external sites? It's not exactly difficult to have an automatic "load from url" system, and it would be smack dab in fair use. Does it have to do with bandwidth concerns or just the database structure making that impractical? Circeus 20:44, 11 May 2008 (UTC)

Nothing to do with the database structure. As it stands, any editor can land us with a major legal problem as "Fair Use" varies a lot by country. As to bandwidth concerns: until recently, we had to stick to TAMU rules. And if we added several Gigabytes of other data to their creaking server we might have had a problem. I know people expressed a concern about the size of the backups that we worked with anyway, even if TAMU were OK with the extra space: WE couldn't cope. We're a bit more free to do what we want now, but we're also at more legal risk and it's easier to keep linking to friendly sites for images. If we do want to host our own images, we're technically capable but legally a bit wary still. I'm not sure if Al got a warning message or not, but today I had to take down all but one of our Photobucket links as they were broken or moved - and it's not an easy task, so I'd rather we got our links permitted in the first place. BLongley 23:25, 11 May 2008 (UTC)
I see. For the record, as far as the legal aspect goes, the only laws that are relevant are those of the country where the primary servers are located AFAIK (which is why despite anything. WikiMedia Commons is bound by U.S. copyright laws, IIRC). <MUSINGS>I've never understood quite how book covers in situations like this could create copyright issues, given that as far as books are concerned, the part really relevant for any publisher is the book's content. One of the four elements used in determining whether or not a work constitute fair use or not is "the effect of the use upon the potential market for or value of the copyrighted work" which are arguably virtually absent when it comes to (moistly) low-quality book covers.</MUSINGS> Circeus 02:13, 12 May 2008 (UTC)
On the legal issue, I think that Circeus is absolutely correct, only US law matters to the ISFDB as an organization, and there should be no legal risk. Wikipedia routinely allows book and album cover images to be uplaoded by any logged-in editor, they have tens of thousands of such images on display, they are a very visible target, and to the best of my knowledge they have not so much as gotten a take-down request over any such image. As long as we moderate submisisons to be sure that uploaded images are in fact book covers, i don't think we would be at any legal risk. (we do want to make sure that no one thinks we are a handy storage spot for their porn collection). I think the wiki software would allow limiting uplaod privilages to "admisn" (aka moderators). I think it can also be configured to require a captcha image test for each uplaod.
As to the issues of bandwidth, storage space, and backup size and duration, I am not in a position to speak. We could probably back up any image store separately from other backups, and perhaps that aspect would not need to be backed up as often. We would need to asses what the bandwidth and storage costs would be. Surely it would make things a lot easier for editors if they could uplaod their own images directly and then link to them. -DES Talk 15:03, 12 May 2008 (UTC)
I was not thinking of hosting it on the wiki (though that is an obvious option), but more like the web-capture function found on LibraryThing, which download the image directly from the servers and keeps it in their database (I have no idea how that works or how to describe: I have never touched the technical side of web database handling). As far as external sites are concerned, this is just one more download of the image, and could not be construed as "bandwidth theft" (which I've always a pretty ridiculous expression, but now's not the time for more musings), Circeus 01:29, 13 May 2008 (UTC)
If we do start hosting our own images, I want to be able to upload scans i make myself (which i am unwilling to upload to Amazon, long story). Enabling the wiki's upload feature and Images/Files section (currently disabled) would require zero additional coding, and seems the obvious way to implement such a feature, if we do. It would also provide a place to store source info along with every image. Images are not often re-written, but do have rather more bits than text pages. What this would do the the backup size and speed I have no idea -- obviously it would depend on how many images got uploaded or captured. -DES Talk 05:29, 13 May 2008 (UTC)
Book covers are copyrighted works with the rights usually owned by the artist or publisher. Any use of covers by ISFDB needs to be under the fair use rules. While Wikipedia is cited as an example of people using covers they have taken down many thousands of images as people were unable or unwilling to provide legally defensible rationales. That said - ISFDB should be able to host and use cover images as part of bibliographic records. However, I would not be in favor of enable image uploads on the wiki side until we have copied, and know how to use, Wikipedia's rationale templates. Any images uploaded would need to be with a rational and I'd want to hold them to even higher standards than Wikipedia. Marc Kupper (talk) 07:08, 13 May 2008 (UTC)
You are quite correct that any use of book cover scans would need to be under Fair Use rules. Wikipedia has taken down lots of such images in the course of self-policing -- what I said upstream was that so far as I know they have never received a demand from a copyright owner for a takedown, they have always gotten any legally questionable images in the course of self-policing first.
As for a rationale template, I'll be happy to copy and modify the Wikipedia templates and document them -- we won't need all of them of course, only two or three: One for author photos, one for book covers scanned by the uploader, and one for book covers from an external source. Wikipedia has dozens if not scores, but we won't be dealing with any of those. Note that the templates themselves include images (the copyright symbol, mostly) and so I can't have them fully working until image upload, for admins at least, is enabled. -DES Talk 14:07, 13 May 2008 (UTC)
See User:DESiegel60/Cover Image Data and User:DESiegel60/Cover Image Data Example for a draft of a fair use rationale template for book covers. modified from the templates in use at Wikipedia. -DES Talk 20:46, 13 May 2008 (UTC)
See also: User:DESiegel60/Cover Image Data-PD Renew /User:DESiegel60/Cover Image Data-PD Renew Example and User:DESiegel60/Cover Image Data-PD Old /User:DESiegel60/Cover Image Data-PD Old Example

(unindent) I am all for hosting our own cover scans if we can do so legally (and hassle-free), but first we will need to address the backup angle.

Keep in mind that now that we are on our own server(s), the way we handle the backups is about to change. We may want to wait a day or two until the new methodology/limitations have been posted and then we can decide how it may affect image hosting. There are quite a few options for backing up large files, e.g. will sell you online storage for $0.10/Gb/month and there are lots of tools to automate backups. Ahasuerus 21:17, 13 May 2008 (UTC)

I quite see the point, and I was not proposing rushing ahead until the storage and backup issues are solved. If hosting outside the wikiis less hassle on those ends, fine. The templates I created I have carefully left in my userspace -- they are examples of what we could do if we host images within the wiki, and besides, they were a useful exercise in template building for me. They are not ready for use until and unless Al says "go" -- not could they be used meaningfully until/unless wiki file uploading is activated. Do take a look at my examples, and see if they give comfort on the legal issues, and whether the metadata they capture seems like a good set for our purposes. -DES Talk 21:44, 13 May 2008 (UTC)
Oh yes, the pages do look like very workable templates, I just wanted to make sure that we were all aware of the fluid nature of the backup situation before spending a significant amount of time on an area that may be significantly affected by new developments just around the corner :) Ahasuerus 21:55, 13 May 2008 (UTC)

Upload enabled

Speaking of new developments: image uploading is now enabled. Now would be a good time to experiment with the templates. We also need some guidlines, such as restricting the maximum dimension (height or width) to 500 pixels. Alvonruff 23:49, 13 May 2008 (UTC)

Ok then, I have moved the tempaltes out of my user space, and declared them "live". They can be found at:
Each temnplate has documentation visible on the template page, and an /example sub-page. Please let me know what should change about these as you start to use them. -DES Talk 12:07, 14 May 2008 (UTC)
I have added the proper tempalte to each of the test uploads done earlier today. Uploaders, please check if I got the data correct. Also, please try an upload or two using the template yourselves, and see how much of a problem it is, and where I can improve it. Thnaks. -DES Talk 12:50, 14 May 2008 (UTC)
OK, my views so far:
  • The uploads are as easy as uploads to Amazon. Plus point.
  • What the templates are actually FOR was not clear. Minus Point.
  • If I've got it right, what they ARE for is lots of credits and/or disclaimers surrounding the image itself, before you've even used it anywhere. Currently, we get some of that automatically when we use a Visco image, for instance: the privilege of uploading our own images seems to demand the responsibility of a lot more work to justify it, and a lot more user knowledge. E.g. I didn't dare approve some previews as they seemed to want to REPLACE the image rather than add to it, but eventually tried it anyway on my sample Cover-Image upload and it seems to have worked. Although the example seems to suggest we should nest templates and I could add a credit (blame?) to myself as scanner if I wanted to. Minus point for clarity of purpose (actually that's covered above, ignore it), minus point for clarity of editing, minus point for having to reenter details already in ISFDB. More help needed here.
  • I haven't actually USED any of the images yet, in anywhere I'd consider them useful: I think I understand how I COULD use a coverart image on the "Bibliographic Comments" page, but that's not as good as what we already have on the pub itself. The templates don't yet cover the areas I want most (sample Artist signatures and Logos) but I'm sure those will come. More help needed? BLongley 21:54, 14 May 2008 (UTC)
If I'm missing the point, please let me know - but so far it looks like more work for things I can do easily already, and we're not yet at the stage where I can do things I couldn't before. I'm sure this will change eventually, though. BLongley 21:54, 14 May 2008 (UTC)
Yes, the source parameter can be used to credit the person who scanned the image. My idea in creating the templates was that they would sit on the image description page of images uploaded to the wiki. Such images could be linked to the display pages in the usual manner, by direct URL, but if the viewer followed the links, the description page would appear. A change of coding might make this both easier for and more obvious to the user (so that a direct click on such an image would go to the image description page) but even without that I think they have some value. Part of the advantage is that Image URLs would be safely stable, we need not worry about a host site taking down the images or changing the URLs. I will be happy to write additional templates and a help page on how and why to use them.-DES Talk 14:58, 15 May 2008 (UTC)
Stability and under our control is good. I see that although images get put under there's a subdirectory structure that doesn't seem so explicable: e.g. Asfaug2007.jpg is under "/3/32/" and my Panther logo under "/d/da/" - these seem automatically assigned at upload, are we sure they'll never be reassigned? (I always mistrust software that does things for me that I didn't ask it to.) And yes, it would be good if clicking on the image automatically took you to the Image page rather than just a larger image, and yes I'm sure that needs coding changes within the ISFDB logic for replacing our usual Coverart: I'm not so sure about use on Wiki pages. BLongley 19:51, 15 May 2008 (UTC)
There are a variety of techniques available for display of images on wiki-pages -- i can direct you to a tutorial on Wikipedia or copy relevant sectiosn here. In general an image can be displayed raw or in a frame with a caption, and at a specified pixil size (which the software will scale the iamge up or down to meet), and centerd, or on the right or left of the page, with text wrapping around it. For more complex effects it is often best to put the image in a cell of a table. In all of these cases, clicking on the image as displayed on the wiki page takes you to the image description page, which usually has metadata about the image and its copyright status. In image may be displayed on many wiki pages, but has only one description page. From the description page it is possible to link to the raw image, and in some cases to scaled down versions of it as well, depending on the raw size. I think the path to the raw image is stable, but I can swear to it.
If it seems that templates to simplify image dispaly on wiki pages are desired, I'll be glad to write them. See my response on Al's talk page. -DES Talk 20:04, 15 May 2008 (UTC)
Oh, another purpose of the templates is to document, clearly, what an image is an image of, rather than relying on what publication record an image's URL is used in. That way if an image is use on multiple publications (as does happen) we know what the uploader thought it was an image of. -DES Talk 15:51, 15 May 2008 (UTC)
This could be useful if we ever agree what we're using Coverart records and coverart images for. Use of the same coverart URL is fine by me if the publications DO have identical covers, or so similar that it only takes a minor note to explain. If we're using high enough resolutions to discern publication prices and serial/ISBN numbers then it might be better to demand one image per pub - although we'd have a lot of gaps then. In that case, knowing what the original was from might help if someone doubted a lazy cloning job. If we're just recording the cover ART on Coverart records rather than the entire cover, then we have to question why people are so reluctant to merge Coverart records - I have done in the past where covers are really identical (British pubs tend to have prices and ISBNs on the rear instead, so the front changes a lot less frequently) and would be tempted to merge more if the pub-level images are going to stay separate: the Coverart record can still record the first use of the image and the artist, the cover art images can record when the lettering or price or ISBN changed, maybe even when an image was cropped a bit, if that's not considered a major change in the Coverart record. Currently, Coverart records are a pollutant in the Simple Search results IMO, and should be looked at sooner rather than later (unless we can just have them moved to Advanced Search results only for now). There's still an option to record the ART at Coverart Record level and cover images at pub level, but that might be more work than we want. BLongley 20:17, 15 May 2008 (UTC)
I agree with msot of the above. I agree that the purpose of the coverart record should be reconsidered. I agree that cover art records should be either optionally or absolutely ommitted from simple search results. If we document on a wiki image description page the actual source of a cover art iamge, we could then link to it from other pubs with teh same art but different prices etc, with a pubn-level note indicating that this has been done, I think.
Upload of a new image with an exixting file name repalces the previous image. -DES Talk 15:54, 15 May 2008 (UTC)
That I do not like. If Random A. Newbie thinks my cover is wrong just because it doesn't match his edition from the same publisher and year, he can upload his image separately and use that. If he's overwritten my UK publication with a US cover, I want to be able to separate the two with a simple clone and put my image back. I do NOT want to have to go find and upload my cover again. Can we turn off over-writing? I want to separate USE of images from PRESENCE of images. I've already encountered one editor that was happily adding/changing cover images to what Amazon was currently showing: it was only when I saw him trying to change one of MINE that I realised he was assuming that an ISBN would change if the cover did. (It doesn't help that the default images that Dissembler supplies are ones that will change to the latest image Amazon has - which is why I use the ZZZZZZZ urls only as a last resort.) BLongley 20:36, 15 May 2008 (UTC)
Currently, the file history of an image page shows the image up-loader, (by ISFDB User ID), shows anyone who edited the description page (just like with any wiki page) and shows any later up-loader who overwrote the image (most often legitimately done when a better version of an image is uploaded to replace an earlier version), and shows the auto-display resolution (which may not be the full resolution for large images). All that we get automatically. But of course the up-loader may not be the scanner. -DES Talk 21:32, 15 May 2008 (UTC)
The over-writing is a feature of how the wiki file-upload works, but there are several safeguards. First, it only happens if the second up-loader specifies the exact same file name as the first -- unlikely unless an up-loader is intentionally trying to replace an existing image. The file name is not automatically derived from anything except the file on the up-loader's local system. Second, if an upload would overwrite an existing image, the up-loader is asked to confirm this. Third, it is possible (I think) to revert to a previous version of an image and thus (though it takes a bit of work), to split the images. Fourth, we can, I think, restrict uploads beyond the restriction to logged-in users - i think we can limit it to admins/moderators, and possibly to a different, separately set group. -DES Talk 21:41, 15 May 2008 (UTC)
In short, J Random newbie will have to work to overwrite your (or anyone else's, even his own) images, and if he does it, we can undo it even if you no longer have the file or the book, albiet with some work. -DES Talk 21:41, 15 May 2008 (UTC)
editing the image description page is just like editing any other wiki page.
If we think that some of the info on the image templates is not needed, it is easy to edit the templates so that the info they use is different. These are still experimental versions, and changes in info or wording are only to be expected. -DES Talk 15:54, 15 May 2008 (UTC)
The "frame with caption", "specified pixel size", "scaling image", "centred, right or left" and "text wrapping around" are the sort of things I'm looking for. All the "how to use an image nicely" things - I know I've messed up a few talk pages (my own included) just by blindly inserting an image URL. I'm not so bothered with pages for the images themselves, although if those could pull in a few details automatically that might help: e.g. the uploader, doing it from his own account, should be credited automatically as the provider: that should be possible on the Wiki side alone I think? But automatically using the first USE of an image here to provide default details of what the image is OF, and FOR, should also be possible - but probably would need ISFDB programmers (we seem to have more than one now!) and Wikimedia experts to work together to do it. As I say, I don't know enough about either yet to be sure of what's possible EASILY: I'm techie enough to know what's possible THEORETICALLY, but also know that I have to turn down such requests every week at work. Anyway, many thanks (yet again!) for trying - I think the replacement of Amazon Coverart with our own is now possible and maybe even desirable (if Al and Ahasuerus aren't worried about backup sizes so much now) so we could already see some benefits, even if already I'm moaning about new problems and resurrecting past discussions that never went anywhere... :-/ BLongley 21:01, 15 May 2008 (UTC)
For information on how to place a picture on a wiki page, see Wikipedia:Wikipedia:Picture tutorial. I'll do an adapted version of this for the ISFDB when i can --I'll be offline most of tomorrow and this weekend. Automatically sucking in info from an ISFDB record into a wiki page is, I suspect, likely to be tricky, but I don't know for sure. I'm sure there is no current, automatic way to do this -- well no, I'm not sure, I'm only sure I have never heard of one.-DES Talk 21:32, 15 May 2008 (UTC)
I'm not programer and for the most part you guys are talking way over my head but is anyone familiar with the book collector software at This is what I use for my collection and among other things it collects and stores both front and back cover images. Adding a new image doesn't delete the old one. If you are not familiar with this software you might want to take a look at it. It might not work as is but it might give you some ideas. I have lots of cover art that I would be happy to share.Rhschu 21:00, 29 May 2008 (UTC)Rshchu

Magazine Project Page

I have added a Magazine_Project in order to co-ordinate magazine editing projects. If anyone will be doing any magazine work on an ongoing basis please add a section so we will be able to avoid any overlaps. Also, Talk:Magazine Project can be used to discuss any policy adjustments and data entry consistency standards that are primarily specific to magazines. Input from all is welcome.--swfritter 23:17, 11 May 2008 (UTC)

I've noticed an overlap with the Publisher project - "Best Books, Inc." or "Best Books, Inc" or "Best Books, Inc. N.Y. City" for instance? As always, I'm reluctant to enter that side of editing but Universal Tandem are going to be a Book/Magazine crossover for one example, and I'm sure there's going to be others. Some inkling about how "exactly as recorded" works with "regularization" of publishers would be good sooner rather than later. BLongley 00:24, 12 May 2008 (UTC)
There are a number of magazines that include city of publication in the publisher field. I'd rather not lose that data and figure it will require a project of some sort to methodically move it from the publisher field to the pub notes or someplace else. As far as "regularization" - it seems like whatever method works for books will work just as well for magazines.--swfritter 18:35, 13 May 2008 (UTC)
Oh, I'm definitely wary of losing information, and don't touch things I can't be sure about, especially verified Geographical additions. So "Columbia Publications" still exists for "Holyoke, MA", "Springfield, MA" and "New York" among others which may or may not be "Inc". But I added a comma to one pub's "Inc" publisher to bring it line with the ", Inc" version, that seemed safe enough. One variant down, possibly thousands more to do still... BLongley 21:18, 13 May 2008 (UTC)

Mixed series (fiction with non-fiction)

I'm experimenting with Summary bibliography pages and series. I wonder how should we display mixed series in the Summary bibliography. Examples are: Peter_F._Hamilton (The Confederation Handbook... - NONFICTION title in the Confederation Universe series) and Paul_Di_Filippo (titles in his Plumage from Pegasus are sometimes recorded as ESSAYs and sometimes as SHORTFICTIONs). There are two option:

  1. display such series complete, in the Fiction Series section of the Summary bibliography; show which titles are essays/nonfiction using [E], [NF];
  2. display the 'fictional' part of the series in the Fiction Series, and essays and nonfiction books in the Essay Series and Non-fiction Series sections, respectively. Similar 'split series' solution is used for mixed 'novels and anthologies' series (see George_R._R._Martin and Wild Cards series - but there it is done, I think, due to different roles 'author vs. editor'.

Note that if we choose (1), we can still move non-fiction titles to a subseries (which would be a real nonfiction series and displayed separately) and if we choose (2), we can still see complete series after clicking on the series' name. --Roglo 16:58, 14 May 2008 (UTC)

I ran across a similar situation with a series that mixes genre and non-genre titles. At present ISFDB displays them under Fiction Series without an indicator. I was happy with that result as the non-genre is fiction though agree that some sort of indicator would be nice. In that sense I'm leaning towards #1 and to show the complete series but in playing with the series display code I see that it's rather complicated with logic to test for and to deal with various combinations of mixed fiction, non-fiction, non-genre, shortfiction, collections, anthologies, etc. The results are not always consistent and so a proposal would be to eliminate all of this and instead the logic would be
  1. If the authorship is the same for all items in a series then include it in the Series section and any titles that are not FICTION would have indicators. Rather than a codes such as [E] and [SF] I'd prefer they be spelled out as "(non-genre)" "(shortfiction)", etc.
  2. If the authorship for a series is mixed then display it in the Series contributed to section using the same indicators as Series. Marc Kupper (talk) 23:53, 14 May 2008 (UTC)
If I recall correctly, the Series display logic was fairly straightforward when it first went live in May 2006. However, I kept finding issues with various unusual permutations and Al kept patching them, which is presumably what made the logic so convoluted.
One thing to consider is that we will eventually want to have a separate database field for "Non-genre" so that we could finally enter non-genre anthologies and short fiction properly. It's probably easier to add this field (and any other ones that we need, e.g. "jvn"/"audience") first so that we don't have to re-do the Series logic for mixed genre/non-genre series like Solar Pons again. We will also want to add logic to auto-delete empty series, something that is not as simple as it sounds due to the fact that we support nested series, but that's another issue. Ahasuerus 04:06, 15 May 2008 (UTC)
I agree that genre/non-genre field is necessary (as is deleting empty series), but it won't solve display problems. The current display logic is quite flexible, even though the code has to select which ttypes are fiction and which are not, or what labels to display for them. We can choose which ttypes form the 'seed' of each series section and then which ttypes we want 'dragged into' this section. Superseries are added, but subseries are not. The limitation, as far as I can see, is that in the bibliography only the works of the single author are considered (so we don't know if there are works by other authors in the series; checking it would probably slow down the page generation, and a single story contributed by another author would change classification of the series; perhaps we need something like 'canonical author' of the series, so that titles by other authors are displayed as 'contributed'?).
From Marc's proposition, it would be quite simple to change the ttype indicators and to display all series in one section. I will see if I can cheaply add some indicator like 'There are more titles in the series than shown'. --Roglo 08:41, 15 May 2008 (UTC)
I tend to agree with Marc above, I feel that the default should be that all items in a series be displayed together. There should probably be an option to display a version of the bibliography without series indications at all, with each item strictly in its chronological or alphabetical place. I also agree that type designations ought to be spelled out rather than initials. In particular, there is already an enhancement request to change the designator [SF] to something else, as it is possible for a user to think that it means "science fiction" (perhaps as opposed to fantasy) rather than "short fiction". -DES Talk 14:46, 15 May 2008 (UTC)
We have alphabetical and chronological bibliographies but currently they don't display title variants, just canonical titles. E.g. George R. R. Martin. --Roglo 15:08, 15 May 2008 (UTC)
A patch to display title variants and pseudonyms in Chronological bibliography is now online. --Roglo 21:25, 17 May 2008 (UTC)

Update by me that I didn't enter

In there is a pub update of The Santaroga Barrier that I didn't create. How did it get listed as by me?

Hm, the submission text appears to have been created on behalf of "Dcarson:
<?xml version="1.0" encoding="iso-8859-1" ?>
<Subject>The Santaroga Barrier</Subject>
<Note>9th printing of the Oct 1968 edition.<br>
Cover signed ’Lehr’<br>
Printed in the United States of America</Note>
but it was verified just yesterday by User:Holmesd, so I wonder if it was actually Holmesd's submission? I will leave him a note and see what he can recall. Ahasuerus 01:29, 16 May 2008 (UTC)
Confirmed by Holmesd. Doesn't look good, notifying Al... Ahasuerus 03:11, 16 May 2008 (UTC)
I need the value of the cookie isfdbUserId for both Dcarson and Holmesd. Alvonruff 11:50, 16 May 2008 (UTC)
Dcarson has 2021 with an expires of Tue, Sep 8, 2037 5:00:00 PM. isfdb_mw_UserID is also 2021 and expires Mon, Jun 2, 2008 4:06:54 PM. Dana Carson 19:34, 16 May 2008 (UTC)

New Backups Location

As stated elsewhere, our backups have automated and the resulting files moved to . The daily zipped up file is a little smaller than what we used to generate because we no longer include some history tables in the file. If you find that anything important is missing, please yell and we will put it back. As far as backing up images goes, assuming 300K per image, 10,000 images should be about 3Gb uncompressed or 1.5Gb compressed, which should be doable from the backup perspective, although bandwidth may become an issue at some point. Ahasuerus 01:29, 16 May 2008 (UTC)

Status of "Our" Images and Images Available on Other Sites

Thinking back to the comment about cover art scanners having no copyright status, I wonder what the legal status of "our" images" will be? Will anybody be able to download them and them post them on their site? Conversely, can we legally download and post other sites' images? (Whether we want to do it and how we could get permissions from other sites to do it ethically would be the next question.) Ahasuerus 01:29, 16 May 2008 (UTC)

If we can legally display a copyrighted iamge under fair use, we can do so no matter where we find it. We would be legally in the clear to download every amazon cover image we are now linking to, and store all such images on our site, provided that our use of them was legal under fair use (which I think it would be).
Wether such mass downlaods would be in good taste or ethical might be debated, but since such sites don't normally own the copyright either (unless it is a publisher's or artist's site), they have no legal rights to control what we do with the images they post. Evan a publisher's or artist's site could not prevent copying an image if it is within the limits of legal fair use. What we want o do or should do is of course another question.
We could perhaps claim a copyright in the assoiation of our images with our bibliographiuc data, or in their layour on our pages and thus restrict people from copying our pages with iamges attached (although our Creative Commons licensce would confict with that) but we could not legally prevent any user from downloading a copy of any image we host and displaying it on some other site. If that use was not legsl under fair use, it would be up to the copyright owner to take action, ther would be nothign we could do, except that we could cancel a user's registrstion if that user violated a policy of ours against some particualr kind of reuse, and we found out. -DES Talk 05:16, 16 May 2008 (UTC)
I'm seeing more and more bookseller images done either as photographs of the book which qualifies as "an original artistic composition" and thus is copyright the photographer or are watermarked/tagged with the book store name. I believe what happened is the booksellers saw their images being used by other sellers or being uploaded to Amazon and there was no recourse.
That said - If I upload an image to ISFDB or anywhere I have no problem at all with someone else using it on their site or even to sell their books. My main issue is with people who indiscriminately download all of the images using a tool such as wget but then don't use them. I've toyed around with making very limited images available on a web/ftp server and that you'd e-mail an auto-responder for a high resolution image. The e-mailed response would include the image as part of a bibliographic record - meaning you'd request a publication record and it's best-available image. This would keep the images within fair use context while also making it easier to throttle indiscriminate downloaders. Marc Kupper (talk) 01:03, 18 May 2008 (UTC)
I'm seeing more and more bookseller images done either as photographs of the book which qualifies as "an original artistic composition" and thus is copyright the photographer or are watermarked/tagged with the book store name.
This would not affect us anyway as the additional copyright created by derivative work does not erase the original copyright of the illustrator, nor does it affect our fair use. It would make the images significantly less useful for us though.
We cannot control fully (anything we can do can be somehow bypassed) other's use of our images, and cannot be held liable to it anyway AFAIK. Besides, the original publishers can not control on a technical basis further willful fraudulent or fair use of their images any more than we can anyway. Circeus 03:29, 18 May 2008 (UTC)
Regarding images done as photographs - Those ARE artistic works in their own and we are not not allowed to use them in their entirety under "fair use" on ISFDB. If we wanted to write an article about these photos of books then we could use one of these images to illustrate the article but it needs to be reduced/blurred so that the image is clearly just an illustration supporting the article and it would then qualify as a fair-use illustration for the article.
The fair-use rules do not have specific guidelines on which portion of a work can be extracted, or which resolutions, are allowed under fair-use but for all sense and purposes if someone has a photo of a book then it can't be used on ISFDB unless you want to extract something very small, and very relevant, such as the catalog #. I would not be comfortable with reducing one of these photos, even down to a postage stamp size, and claiming "fair-use" of it to support a bibliographic record.
In summary - if someone wants to upload a photo to ISFDB then it needs to be accompanied by a full release of the image to the public domain and sufficient documentation that the submitter was in fact the copyright holder. I suspect "sufficient" would be an original high-res image that includes the camera meta-data though ideally it would be a written release sent to ISFDB galactic headquarters. This is why nearly all of the photos of people were removed from Wikipedia as they clearly are copyright photos and had not been accompanied by public domain releases with supporting releases sent to Wikipedia.
Cover scans are "slavish copies" of copyright works and the image/file copyright belongs to the publisher and/or cover artist. The scans are not "artistic compositions" in themselves and so there's no copyright, or other rights, given to the scanner. Thus the fair-use rules are simplified as we are dealing with the publisher and/or cover artist's rights and intend to use their work under fair-use. There still are rules that must be followed. 1) Any images MUST be used to illustrate biographic or bibliographic material. 2) The fair-use rules state that you can extract a very small portion of a work and the portion extracted must be directly relevant to your own work. There seems to be an established precedent that a reduced size image of the entire front cover of a book qualifies as "small portion." It's possible this has not been tested in courts and that sites such as Amazon are using another area of fair-use which allows for images of products being sold.
There does seem to be some established precedent on Wikipedia of cover images being used to support articles about a book. From a legal view it's an interesting area as the images are not a "small portion" of the cover artist's work and the article is nearly always about the book's content whose copyright is usually with the author or publisher and thus unrelated to the cover artist's rights. Thus as far as ISFDB goes, I'd monitor the situation carefully and would remove the book images if the legal winds ever start blowing in favor of the artist's rights. Sites such as Wikipedia and ISFDB are unusual in that they contain biographic and bibliographic content and are not selling the items pictured. Marc Kupper (talk) 18:23, 18 May 2008 (UTC)
I strongly suspect that a "photo of a book", if cropped so that only the front cover is visible, and thus the image cannot usefully be distinguished from a "slavish" scan, would not have any rights remaining to the photographer, and be back to the same issues as a normal cover scan. You say above "The fair-use rules state that you can extract a very small portion of a work and the portion extracted must be directly relevant to your own work" Taht isn't quire coirrect. What the law says is that the "amount of the work used" is one of the factors to be considered in whether a use is fair use. In msot court decisons, only uses that are a small portion have been tested. But anothe factor is "does not ahrm the market for the original" If the original cover art is sold only on a book cover, any site such as the ISFDB would easily pass this section. Wikipedia's rules err on the side of caution, as we probably should also. -DES Talk 00:44, 19 May 2008 (UTC)
You wrote "I strongly suspect that a "photo of a book", if cropped so that only the front cover is visible, and thus the image cannot usefully be distinguished from a "slavish" scan, would not have any rights remaining to the photographer, and be back to the same issues as a normal cover scan."
That is not true - Only the owner of copyright in a work has the right to prepare, or to authorize someone else to create, a new version of that work. Accordingly, you cannot claim copyright to another's work, no matter how much you change it, unless you have the owner's consent. See Copyright Registration for Derivative Works. If a photograph were modified so that it was 100% indistinguishable from from a "slavish scan" then it's still a copyright violation and the person doing so is a pirate. Obviously, if someone uploads what appears to be a "slavish scan" to ISFDB we pretty much have to trust that's what it is and that it was not someone who pirated a photo and edited it. Marc Kupper (talk) 22:32, 25 May 2008 (UTC)
In general you are correct. But an absolute requirement for copyright protection under US law is originality. If, starting with a copyrighted work, soemone creates something so lacking in originality as not to be copyrightable at all, then neither the creator, nor the owner of copyright in the "parent" work, have any copyright, because no one does. A "slavish copy" -- however achieved -- of a flat work of art, is not an original creation. it is not a derivitive work because it is not a "work" at all within the meaning of US copyright law, at least if Brigeman vs Corell is good law (and Feist strongly suggests that it is). It is merely a copy of the original flat work. Passing through a spearate stage will not, I think, change this -- because all the original aspects of the photo that made it a separate copyrightable work have been removed. You can't get a copyright on a "slavish copy" by starting with a "photo of a book" and cropping.
All that said, a true scanned image is easier to make and likely to be a better image anyway, so we won't want cropped photos if there is any alternative in any case. -DES Talk 22:51, 25 May 2008 (UTC)
I just entered/verified some pb editions of Larry Niven's Destiny's Road. I saw that the hc 1st edition did not have an image in ISFDB and so took a look at Amazon to see if one is available. 0312851227 offers a scanned cover image (which I used) but there's also a nice customer photo at I would consider this photo copyright Paco Alameda "ballisticrose" (the person who uploaded and also presumably the photographer as to upload he checked the box that says "I own the rights to this image.") as it demonstrates originality in the book's setting, lighting, etc. It's not a "slavish copy" of the book or its cover. These are the sort of images that I'm referring to that should not be uploaded to or used on ISFDB without a full release from the copyright holder (Paco Alameda).
The second point I was making was that someone can't, without permission from the copyright holder, take this copyright work and to make derivative works. We can't, without permission, rotate/crop the image so that it looks like a "slavish copy" of the cover.
If someone wants to use this photo under fair use then the fair use guidlines need to be considered for the entire work and not just the book that's the main subject. To really make sure I don't infringe on that photo's "potential market for or value" I'd need to reduce or otherwise degrade the image so that it does not infringe on the value of the original work.
The scanned "slavish copies" of book covers we have been using on ISFDB are reduced to 200 pixels high and presumably that, and the way we use them on ISFDB, is sufficient to pass the four fair use tests.
ISFDB does have an aspect where the image is reduced to 200 pixels for the publication display but if you click on it you are taken to the original image. Within the publication display we had a number of fair-use elements that we lose when the image is displayed alone. With that in mind I'd follow Amazon's standard where they reduce images to 400 pixels and do not make higher quality versions available. Marc Kupper (talk) 20:22, 28 May 2008 (UTC)
I think that you are mistaken on the question of creating slavish copies by cropping. I think that if you take an original work and so crop it that nothing but a slavish copy is left, you have removed the photographer's originality, and what is left is not a derivitve work in the copyright sense. I would have no hesitation in uploading such an image to wikipedia, or placing it on a personal web site (in a situation where a cover scan would be fair use, or if the underlying image was PD). But I don't think we normally need such images here in any case, so i hope the point is moot.
I agree with your other points. I think that Al had asked for a limit of 500 pixels for technical reasons anyway, which is probably limited enough, as it is likely to be under 100 dpi for even a MMPB cover, which is well below print quality. -DES Talk 21:14, 28 May 2008 (UTC)
I don't recall Al asking for 500-pixel limits at all: I remember Marc pointing out that's what Amazon reduce images to. Personally I scan at 300 dpi, upload to Amazon and let them deal with it. BLongley 22:22, 28 May 2008 (UTC)

(Unindent) I'm not too worried about the copyright issues really, I just want to stay on the safe side. I'm not going to upload anything I didn't scan, am not going to post big images, and don't care if they're copied elsewhere. All I want to know is how we are going to use and organize them. I don't see a huge rush to use the new upload functionality: I've uploaded some examples and used them. I've tried the templates but frankly it's easier to upload to Amazon still. I appreciate that's not acceptable to many others that might post images, but it works for me until Amazon shaft all the links. If someone wants to make the effort, feel free to transfer all my Amazon images from the UK and US sites to here. BLongley 23:52, 25 May 2008 (UTC)

I have uploaded omnly for test purposaes, because to the best of my understandign Al had not given the go ahead to actually use uploaded images as links in pub pages. Did I musundeerstand? -DES Talk 05:33, 26 May 2008 (UTC)
I have asked Al. on his talk page, if he considers images ok for live use, or only for on-wiki testing. No response to date. -DES Talk 21:14, 28 May 2008 (UTC)
Well, I've used two for cover-art in the DB, whether you consider that for testing purposes or not doesn't really matter: at the moment it's still easier for me to use Amazon. As it stands, I'm not going to be using ISFDB for cover-art images unless there's no entry I can use there, or it proves too unstable: if people can sort out what the Image Upload rules here are, and if they're no more onerous than elsewhere, I'll use it. But at the moment Image Uploads here seem only useful for other uses - none of which is really leading to much of a response. Someone make some decisions and make this feature useful, please! BLongley 22:22, 28 May 2008 (UTC)

Helix Magazine

We currently have at least 7 issues of Helix magazine in the db: 867624, 867719, 867684, 867648, 867702, 867665, and 868842. There is a submission pending to edit one of them. As I understand it, Helix is a completely online web magazine, not a downloadable e-zine like Jim Baen's Universe or The Grantville Gazette. This is supported by what I read at today. As I understand our current Rules of Acquisition, these should be out. Since deletion is not currently easily reversable, i want to confirm that my understanding is correct before I delete these items. -DES Talk 22:27, 20 May 2008 (UTC)

Some of the authors have a number of other stories from acceptable sources. When it comes down to it the issue of including webzines is totally dependent on the willingness of moderator(s) to totally verify the contents of submissions as they are added and to monitor their stability - webzines without reliable archives would be out. I think we are doing a great disservice to the sf community by excluding them but it would be up to us, and there are more of us now, to be willing to share the load. Who knows, we might even attract some webzine savvy moderators. The other issue is that some sites charge subscription fees but most of those still display a table of contents.--swfritter 22:53, 20 May 2008 (UTC)
Indeed several of the authors and both of the editors are fairly well-known, and they apaprently limit submisisons to SFWA members and associates, or those qualified for membership. Perhaps we need to revist this issue? I could verify contents at time of submission, but how could stability be monitered reasoanbly? -DES Talk 23:00, 20 May 2008 (UTC)
The first requirement is that at least a few moderators want webzines and are willing to take responsibility for specific webzines. There is no point in an extensive discussion of methodology if that is not the case. Since everything is online it shouldn't take a long time to periodically check the entries - maybe only a couple of times a year. If a site goes black - no problem - we have verified the data. If there is no interest in including webzines I will help you remove the Helix entries. It can be done fairly quickly by 1) Unmerging titles from the pubs, 2) Giving authorship of all entries to a bogus name, 3) Deleting all entries for the bogus name and making sure that Editor and Coverart records are gone.--swfritter 13:38, 21 May 2008 (UTC)

Using the new Web API (safely)

Now that we have a shiny new Web API, we may want to think of the Warious Wonderful Ways in which we can use it. One thing that comes to mind is that there are a lot of missing, duplicate and otherwise deficient records in the database that are easy to identify programmatically, but could be tedious to correct. For example, there are a few thousand magazine issues with a missing EDITOR title and they are easy to find with a simple script, but it would take a lot of man-hours to pull each one up, add an EDITOR Title, file, approve, etc. I figure that the Web API could make this process much simpler by letting us create automatically generated submissions (via XML:PubUpdate) with an EDITOR Title reconstructed from Publication level data. Similarly, we could programmatically create submissions to merge any Title pairs as long as all of their fields are exactly the same.

Does this sound like a reasonable approach? If it does, I could try generating a few dozen sample submissions and see where it gets us. If it works, then I will probably have to create a new bot account, e.g. "Fixer", so that other moderators would know which submissions can be freely approved. Of course, if a bot screws up, we will want to put the submission on hold and notify the bot owner to prevent similar problems in the future. Ahasuerus 23:15, 22 May 2008 (UTC)

This sounds good. I'll probably create some XSD schema... sometime relatively soon, I guess. This can be used to validate the XML before it is ever placed in the submission queue. Alvonruff 00:05, 23 May 2008 (UTC)

Editing Briefly Offline Tomorrow Morning

I'll be putting the changes in place tomorrow morning to support linking reviews to titles. This requires running an SQL script to create entries in a new table, as well as updating/adding numerous apps. The publication submissions will autolink reviews to titles where possible.

Since the autolink apps can't go online until the table has been created and populated, and we don't want any submissions to be approved until the autolink apps go online, I will disable editing during the update. This should happen around 6am CST. Alvonruff 13:34, 24 May 2008 (UTC)

And there were wild celebrations in the streets! Ahasuerus 17:55, 24 May 2008 (UTC)
Ta Da! Is there anything different we need to do when entering reveiws in future to make sure that they are linked properly? Is there anything we should or can do to link reviews that the automated scripts miss, if there are any?

Linking reviews

Here are how things will be different:

  • When a publication with reviews is approved, an attempt will be made to match up with a title that has the same title and author. If a match is found, a mapping is made between the review and the title.
  • When printing a title, reviews are displayed if a mapping exists. So far this is no different than the current setup (it's still lexical matching), there's just a table to support it.
  • If you look at the original publication, each review that had a mapping to a title will be linked to that title.
  • That means if you see a review refer to a title that isn't underlined, it needs some work. These are pretty easy to spot now.
  • There is also a link for the review itself in the pub listing. Once you find/add the correct title, you can click on the review, click on "Link This Review...", and add the title's record number to link the two.
  • This also means that the title of the review can differ from the title of the reviewed work, as sometimes happens in a publication.

The script I have that creates the initial mapping table found matches for about 90 percent of the reviews. Some of the titles are missing, some have small lexical variations (like different single quote characters), and some reviews still have the old caret annotation for variant titles. It also possible that some of the mappings will be wrong (right title, right author, wrong work, matching to a short story instead of a collection) - at least now they can be corrected. Alvonruff 21:45, 24 May 2008 (UTC)

This all sounds good. I take it that reviews link to titles, rather than publications? So if a review is explicitly of a particular publication (for example if it discusses presentation, as some reviews of Gregg Press reissues do) this can only be indicated in notes? -DES Talk 22:12, 24 May 2008 (UTC)
It does link to titles. Review linking is now online. I checked through a few magazines and have already hit an interesting issue that we'll probably need to make a policy for. The Jan/Feb 2008 issue of Analog has a review "Hurricane Moon" as by Alexis Glynn Gardner, while the actual title is attributed to Alexis Glynn Latner. Presuming that the data entry on the review is correct and the author was listed incorrectly in the magazine (don't know if that's the case), do we want to preserve the review titles and authors as those listed in the magazine?
Now that we have review linking, I'd personally keep the titles (and did so in the May 2008 Locus on Celebration, The Starry Rift, Where Angels Fear, and Deamons Are a Ghoul's Best Friend), but the authors aren't of much value if the names are mangled. Alvonruff 10:42, 25 May 2008 (UTC)
This seems to have created MORE work in some areas. E.g. I added "Angles & Other Stories (Audio)", the omission of which was flagged by a review, but now I have to manually link the review (which currently isn't working anyway) whereas before the lexical match would give me the link automatically. Is there a plan to eventually check new Pubs (not clones) against reviews, as well as reviews against titles? BLongley 13:23, 25 May 2008 (UTC)
Also, when fixing reviews directly there's no attempt to find a link. For instance, the missing reviewees when the reviewer is a pseudonym, e.g. here. BLongley 13:37, 25 May 2008 (UTC)
If you add a review to an existing pub, it will look for the matching title and make the link. If you edit an existing review it will NOT look for the matching title. This is by design - we don't want to lose review links because someone updated the review title to match its appearance in a magazine, which happened to differ from the actual title.
Can't you make it so it ONLY looks if there currently is NO link? Although in this case of pseudonymous reviewers, we may want to leave them unlinked anyway if the canonical reviewer's version IS linked - but we still have a bit of a display problem there. BLongley 14:46, 25 May 2008 (UTC)
If you add a review and it wasn't autolinked, then there is a difference between the titles. So far the common ones I've seen are differing versions of single quotes, and ASCII '&' vs '&amp;'.
If you add a new publication (and thereby title), there is currently no search for a matching review - matching is currently triggered on review changes. I'll need to think through if that's desirable. Alvonruff 14:14, 25 May 2008 (UTC)
I use the reviews to find missing pubs, and at the moment the more evidence there is that we ARE missing the pub, the more work it is to add it and link it correctly. I guess I can take the benefits (easier finding, thanks for that!) and leave the problems (reviews still need linking at some point) to everyone else but it seems best to sort it all out in one go. BLongley 14:46, 25 May 2008 (UTC)
Seems to work fine. One complaint, it was easier to spot the reviews when they were italicized. Now that the linked reviews are in plain text, maybe they could be indented?--Rkihara 16:01, 25 May 2008 (UTC)
And perhaps indent the interior artwork; along with sorting artwork on same page as story after the story.--swfritter 22:25, 26 May 2008 (UTC)
Seems to work fine for me also. I agree with Rkihara about intenting the review listings. Also, the "Link review" screen seems to be based on the "Make varient" screen, but without the "Create new parent" section. If we get a review of a work not in the DB it would be handy to be able to create the missing entry and link the review to it all at once, as the info needed may well be right at hand in the review. -DES Talk 16:34, 25 May 2008 (UTC)
That only creates titles though, whereas a review normally contains enough information for a pub, or maybe more than one. (Simultaneously published hc and tp versions for instance.) I'm not keen on creating empty titles (those get picked up by data quality scripts as an error) and suspect it would be a fair bit of work to create a "New Pub based on Review" page? BLongley 18:30, 25 May 2008 (UTC)
It occurs to me that we now need a data cleanup page for unlinked reviews. -DES Talk 17:11, 25 May 2008 (UTC)
When I last looked, there were far too many to list. I did create Authors that only exist due to reviews to clean up some of the worst offenders though, if someone wants a project to finish off. I'm finding it pretty easy to find stuff to work on just by looking at a magazine series though: e.g. I've done three years of Campbell Analogs already. BLongley 18:25, 25 May 2008 (UTC)

(Unindent) I'm leaving this for tonight. Yes, there's a fair number of submissions of mine on the queue still - feel free to approve or reject those, but please explain the reasoning. For example, there's two with a missing exclamation mark in the title, many missing or added middle initials in the author, a "Jr." suffix, a "Sir" prefix, etc. I've submitted and approved several minor variations like "III" versus "3" or "Three" in titles but am wary of Author pseudonyms. Is it OK if we have the Title AND the Author here, even if we don't have that exact combination already? If you're not sure, feel free to leave YOUR edits for Peer-Review. BLongley 00:42, 26 May 2008 (UTC)

FYI, I just ran across a magazine that was recently edited by an editor who seems to be confused about the new review linking process. Several reviews were entered as essays, with "Review:" appended to the title. I wasn't able to figure out who it was, so if a new editor submits an edit on a magazine, please check.--Rkihara 04:33, 26 May 2008 (UTC)

Log-in problems

Is it just me or do other editors have to log-in several times in one session when switching between the database and the wiki? This also happens when I've posted a comment on the wiki, then the edited page returns and I discover that I'm no longer logged in. I usually open three windows in Firefox. The first is the wiki where I initially log-in and check the box asking to be kept logged-in. The next window is the front page of the database where I log-in. Then I open the moderator page for which I don't have to log-in since the system knows that I'm there. But when I get a submission that needs clarification from the editor, I click on the submitter's name and am sent to their User page (ideally, I think I should be sent to their Talk page, but that's a matter for another discussion.) When I try to post a comment, I discover that I'm not logged in. After I log-in once in that window, I'm fine for the rest of the session. Should I have to log-in to the wiki in every open window, or is this a Firefox thing? MHHutchins 15:07, 2 June 2008 (UTC)

I find that i can open new windows or new tabs and stay logged in just fine (unless the session times out, which it will after I think 24 hours of inactivity) with one exception. If from the moderation page I click a user's name to go to that user's user page, I usually arrive in a logged-out state. If instead i copy the user name, go to an open wiki window/tab (or open a new one from an open one, by right clicking on "main page" say, with "Open in New Window" or "Open in New tab") and type in "User:" (or more likely User talk:) followed by the pasted user name, I arrive in a logged-in state. I'm not sure whether having to click log-in, re-type my pw and click ok, is more trouble than having to open a new window/tab and type "User talk:" and do a paste. -DES Talk 17:04, 2 June 2008 (UTC)
And I use IE7, so this isn't a Firefox issue. -DES Talk 17:05, 2 June 2008 (UTC)
Check to see if 'www' was dropped from the URL. If so, prepend www and you will find yourself logged in again. Browser cookies are URL sensitive, and we don't have a server rewrite rule for any dropped www. Alvonruff 17:27, 2 June 2008 (UTC)
That is it, at least for me, that link goes to "<UserID>". That explains why a single log in covers the matter -- there are then two cookies on file, one for each URL stem. -DES Talk 18:08, 2 June 2008 (UTC)
So is it possible to have the link from the submission page contain the "www" as part of the URL? David's method of copy and paste would take more time than just logging in again. :) MHHutchins 20:05, 2 June 2008 (UTC)

Review links to pubs

copied from User talk:BLongley‎ -DES Talk 22:24, 2 June 2008 (UTC)

Bill, I think we should go ahead and approve those submissions in which you link reviews to pubs. Whether or not there's an exact match between the title as published in the review and the title recorded in the ISFDB, it's rather clear that the review refers to a specific title (not necessarily a specific pub). I think that's the whole point about Al giving us the power to link review to pub. We don't (and, IMHO, shouldn't) create variants of a title simply because there exists a review that doesn't EXACTLY match the title as published. I'm interested to hear if you have any argument with this view. I suspect that you agree otherwise you wouldn't have gone to the trouble of submitting those edits, but held off on approving them in case someone should come along later and create variants (God, I hope not!) MHHutchins 14:53, 2 June 2008 (UTC)

Look closer - I didn't leave them because the Titles mismatched, but because the AUTHORS mismatched. Here Al said "the authors aren't of much value if the names are mangled" and I tend to agree - but the reviewees aren't really mine to correct in the pub itself. If people now want to leave the pubs with reviews as stated, then I want the stray authors created ONLY from reviews to be set up as pseudonyms, as when you come across an empty author in the author directory there's no way to find why it's there. But so far nobody's voting for fixing reviewees (I suspect that's Al's intention but it's not clearly stated) or leaving them and creating pseudonyms: I just don't want to leave them stuck halfway between those options. (That's why I created "Authors that only exist due to reviews".) Do you have a preference? BLongley 17:57, 2 June 2008 (UTC)
I have been happily linking non-exact-match titles when the reviewee matches the right author, but I'm not sure about linking them to pseudonyms, see this question. BLongley 17:57, 2 June 2008 (UTC)
Oh, there is one other type I left: "A Spectrer is Haunting Texas" is surely either "Specter" or "Spectre"? There I suspect the review could be cleaned up. BLongley 18:15, 2 June 2008 (UTC)
Whether the differences be author or title, I don't feel that variants should be created. For instance, there was a work reviewed by "E. Hoffman Price" in one of my verified pubs. I corrected the review record to the correct author's name "E. Hoffmann Price". It made the author record and summary page for the single-n Price disappear, as there are no pubs actually published in that name. Let's take your first submission as an example. The Universe Makers has never been published as by "Donald Wollheim", regardless of what the review states (whether it be an editor input error or reviewer error.) By approving this submission, you're only linking this review with the publication under review. You're not even changing the title record of the review (though I'd do it if I were the original verifier of the magazine in which the review appeared.) MHHutchins 20:01, 2 June 2008 (UTC)
Not the best example, as we already have the "Donald Wollheim" pseudonym set up. (See below DES's comment for a better one.) I'm not that bothered about Title discrepancies - ones that look like typos can usefully be queried, most are just ones where the reviewer includes a series name as well, or adds a series number or such - not important, and I've let dozens of those through. BLongley 21:49, 2 June 2008 (UTC)
My feeling is that the review record should record what was present in the review. Formerly we had to correct to canonical author and title to make the lexical link work. Now we can instead simply link to the correct title (as I understand it reviews cannot link to a specific pub) no matter what name was used in the review itself. The only problem comes when the review is used as a source to create the initial publication record. Then if the title is incorrect or the author's name is misspelled, we will want to correct it in the pub record (not in the review), but this is no different from any other case where a record is constructed from secondary data -- when we can determine what the actual state of the published work was, we correct the record. As for variant reviews, the only reason i see for those now is when the reviewer used a non-canonical form of his or her name -- then we want a vt to put all the reviews onto the primary page for that author, just as we would for an essay, story, or novel published under a non-canonical name. Am I missing something here? I'm about to go off and do a little experimenting. -DES Talk 20:46, 2 June 2008 (UTC)
A better example of my concern is probably this one. Have a look at the Author page for "Dr. E. E. Smith" and see if you can get any use out of it. The usual "Click on Titles" solution will NOT help. An Advanced Author search gets you to the same page. Advanced Title and Pub searches by that name just hammer the database. It's THAT sort of problem I want to avoid. BLongley 21:49, 2 June 2008 (UTC)
I suspect (but have no evidence yet) that pseudonymous reviewees don't display correctly if we don't have the reviewed pub with exactly the right title AND pseudonym - but Al may have done a better clean-up than I expected. I do know there was a problem with pseudonymous Reviewers though - see here. One of the reviews lost its reviewee. BLongley 21:49, 2 June 2008 (UTC)
Anyway, glad to see there's some discussion at last, I just wish it was where it was supposed to be rather than my talk-page! As I said originally "feel free to approve or reject those, but please explain the reasoning" - from what you've both said here, it seems Mike favours fixing the reviewee, and DES wants it left alone? BLongley 21:49, 2 June 2008 (UTC)
I just did a little testing, and I'm rethinking the above. More exactly, the above as what I think we would do in an ideal world, but...
Take a look at {{P|264574}. In this test publication, you will find a review of Test Novel by Test_Author. However, in the review, the author has been identified as Testy_Author, this may be a typo in the review, or a pesud not actually used on the particular pub in question. This creates an empty author record for "Testy Author", which is probably a bad idea, unless we already have this variant recorded as a pesud. Note that the empty author record is created even though the review is properly linked to the title that shows the correct author name.
Given the above test, i now think that we need to alter author names recorded on reviews to match the name on the actual title being reviewed, unless the name used in the review is a variant we already have on file (presumably because works were in fact published under it). In any case, we ought to link reviews to the proper title records whenever possible. The name of the reviewer ought to be recorded as printed at the head of the review or review column. If this is not the canonical name of the individual, then a variant record should be set up to put the review on the proper summery author page.
What do the rest of you think about this? -DES Talk 21:26, 2 June 2008 (UTC)
Will look, but Edit conflicts here are making it a pain to do anything at the moment. BLongley 21:49, 2 June 2008 (UTC)
OK, had a quick look and added the extra complication of this review: see where the reviewee (Book Author) is lost? We have several of those to fix still (and Al has to fix the "Make This Title a Variant Title or Pseudonymous Work" code at some point too before we get more). BLongley 21:55, 2 June 2008 (UTC)
And now it looks like "Mike favours fixing the reviewee, and DES wants it left alone IF it's a pseudonym already, or corrected otherwise."? BLongley 22:02, 2 June 2008 (UTC)
I am going to copy this to the community portal, and respond there, This is too general and too important to be left on an individual's talk page. -DES Talk 22:24, 2 June 2008 (UTC)

End text copied from User talk:BLongley‎ -DES Talk 22:24, 2 June 2008 (UTC)

My view is that we should record what actually exists, to the extent feasible, unless that interferes with presenting the data usefully. I don't see any problem with leaving the reviewer's name as it appears, except for the display issue Bill highlighted with this review, and i hope and trust that that is merely a bug that will be fixed promptly. I do see a problem with creating empty records for misspellings and variants of author names that were never used on an actual publication, and that leave no easy way to get to the canonical author name. If a review link would remove record for the author listed in the review, (when different) I would be happy, but I suspect that would be much harder to do. None of the problem comes, as I see it, from approving or not approving review links, the problem comes from correcting or not correcting the names of the authors of works reviewed to their canonical names, or to the names actually used on the works as published. That issue arises whether the review is linked or not -- indeed it arises even when the work reviewed is not (or not yet) included in the ISFDB at all, so there is nothing to link to. For example, suppose that Joe Blow is a well-known author of science fact books, who has also published a few SF stories, so he is in the DB. Suppose Analog reviews his latest science fact book. We are not going to put that in the DB, according to the Rules of Acquisition. Suppose that the review describes him as "Dr. Joseph Blow", but all his SF (and indeed his non-fiction books) are published as by "Joe Blow". If the review record records this as published, it will create an author record for "Dr. Joseph Blow", an empty and useless record. That is a Bad ThingTM. But, on the other hand, suppose that half his fiction had been published as by "Dr. Joseph Blow". Then we would already have the record, which would properly direct the user to the "Author:Joe Blow" page. That would be a good thing. Correcting a review to the canonical name in the 2nd case would not be very harmful, but is not required either. That is my view, at the moment. -DES Talk 22:24, 2 June 2008 (UTC)

If it's an obvious misprint in the review, yes, I believe it should be fixed in the submission. (That would get rid of the stray authors at least.) But even if the submitter doesn't fix it or is unaware of the error to begin with, he should go back to see which reviews have not been linked automatically in his submission (Al's layout makes it clear which reviews aren't linked). Then he should search the database to determine if the book under review is in the database, and when the title is discovered he should create a link. This leaves the original review exactly as it appears in the publication, but still links it to the title under review. This is how I've been fixing the reviewzines that I entered months ago. As for pseudonymous reviewers, it's a moot point, because it was never part of my original post to Bill. MHHutchins 22:51, 2 June 2008 (UTC)
I agree that when a review does not auto-link, a search should be made and a link created if the title reviewed is in the DB. But the question arises as to what is a "misprint". I have seen reviews that used informal versions of an author's name exclusively (made up example: Foundation by "Ike Asimov"), particularly when a review column is discursive and the reviewer knows the reviewee. And then there was the case I ran into a few months ago where a review referred, three times, to "Ursula LeGuin", no space. That is a far from unheard of variant, and ideally we would record that this review uses it, but it creates a useless and empty author record. But is it a "Misprint" or a variant? -DES Talk 23:35, 2 June 2008 (UTC)
When a reviewer attributes a book to "E. Hoffman Price" that's a misprint. When a reviewer attributes a book to "Ursula LeGuin" that's a misprint. When a reviewer attributes a book to "Ike Asimov" that's just plain goofy. When a review is for "A Spectrer is Haunting Texas" or "Gather Darkness" that's a misprint. None of them are variants. Variants should not be created. It should be easy to link the review to the correct publication or create a new publication if there's none in the database. What's the big deal if there's a blank summary page for "Ike Asimov"? Anyone who goes there looking for a publication, would clearly see that no publication in that name EVER existed. If we create a variant for the review, how many people will mistakenly believe that the publication may have existed. Honestly, no one has made a valid case for creating variants for misprints in reviews. As the person who has entered thousands of reviews in this database, I think I know a thing or two. This is an area where the "record exactly as shown" policy requires the need to use a little common sense. What happened to past suggestions that we can note any obvious mistakes in the notes field of the periodical's pub record? MHHutchins 03:16, 3 June 2008 (UTC)
Perhaps I am restating the obvious, but just to be on the safe side let me try to summarize where I think we are. Now that Al has added support for links between review records and regular Titles, we can enter reviewed titles as they appear in magazines regardless of whether they use subtitles, series names or other previously headache-inducing additions/subtractions. Whether to record typos in the review record itself or in the Publication's Notes field is still debatable, but as long as we link the review record to the correct Title, we should be OK either way. The last outstanding problem, if I understand it correctly, arises only when the author's name is printed differently in the review vs. the book/story being reviewed.
For example, as far as we know, Davy and the Goblin; or, What Followed Reading "Alice's Adventures in Wonderland" (1884) has only appeared as by Charles E. Carryl. However, a recent review of the book by F. Gwynplaine MacIntyre used Carryl's full name, "Charles Edward Carryl". The question then is whether we want to record the author's name as it appeared in the review or as it appeared in the book. If we decide to record it as it appeared in the review, then the next question is whether we want to set it up as a pseudonym -- the way Charles Edward Carryl is currently set up -- or leave it hanging there without any Titles or other ways to easily determine why we have it in the database.
As far as my personal opinion goes, I am torn here. On the one hand, I can see how it would be quite useful to let our users search for "Charles Edward Carryl" since, if they have read the review and not the book, it's probably the only form of the author's name that they are familiar with. If we don't record it, then there is a decent chance that a user looking for Carryl will never find him. On the other hand, I can see how adding every single form of every name ever used by every reviewer can lead to thousands of additional pseudonyms in the database. If you search on "Simak" and the search returns "C. D. Simak", "Clifford Simak", "Clifford D. Simak", "Cliff Simak", "Clifford Donald Simak" and perhaps a few misspellings, does it get to be too much? (BTW, do we have an outstanding request to mark pseudonyms as such on the Search Results page?) It's a tough call, but the former approach seems to be, on balance, a little better.
If we do decide to record the names that are used in review records, then I would be inclined to record them as pseudonyms, i.e. the way "Charles Edward Carryl" is currently set up, since running into an Author record and not being able to tell why it's in the database or what it relates to can be quite irritating. Ahasuerus 05:11, 3 June 2008 (UTC)
I think a lot of the problem is that the canonical author is fixed. A way out might be to link all forms of the author's name to the "author page" and let the canonical name be determined by automatic numeric polling from pubs and reviews. This might be indicated in either "tree" or "outline" form at the top of the author page. In outline form the legal name could be under the main heading (I.), and the canonical name (II.), with pseudonyms as subheadings (.a, .b, .c, . . .), subheadings again ranked by numeric polling. All rankings would be fluid. Shared pseudonyms could be placed under a third major heading (III.). Everything will now be self-adjusting, and misspellings or mistakes will be easily spotted. It will probably be a semi-major programming job though (sorry Al!). Just my two-bits worth.--Rkihara 06:09, 3 June 2008 (UTC)
I think we are talking about two different problems: (1). Do we want to record variants used in second sources only (perhaps in Internet reviews too), which helps to find an author / a work you see mentioned somewhere. The same policy should be used for titles as for author names. (2). Now, that we have reviews linked to the reviewed titles, do we need the old 'reviewee link'? Perhaps we should remove 'reviewee links' from linked reviews and just record in the review's title the book's title and the reviewed author as stated 'Book X by Joe S. Author'; you can click on the link to get to the real title and author. --Roglo 22:52, 4 June 2008 (UTC)


I stumbled across [3] while looking for an ISBN. It looks like either they're using an incredibly similar system (compare the authors borne today list!), or outright mirroring an older version of the website without attribution. The Bernard Werber page is EXACTLY as it was here before I got to adding all the missing content... Circeus 22:26, 2 June 2008 (UTC)

There IS a credit at the "About" page: "Bibliographic information was stolen (that's such a nasty word) adapted (why didn't I become a lawyer?) in large part from the Internet Speculative Fiction Database (ISFDB) by Al von Ruff et al. as well as, the Library of Congress Online Catalog, and Wikipedia, among numerous other print sources and member contributed data." I guess that qualifies, although "Al et Al" makes us sound like a Middle East airline or something... :-/ BLongley 23:12, 2 June 2008 (UTC)
They should really reconsider that copyright symbol on the front page... Circeus 23:40, 2 June 2008 (UTC) is another Web site that cuts and pastes our bibliographies without attribution and claims a "2006-2007" copyright, e.g. this bibliography of Henry Kuttner. Ahasuerus 19:31, 19 June 2008 (UTC)
I just found that site after looking at David Drake to see how other people deal with the "Venture SF" series/imprint/publisher. Combine that with the thousand-to-one chance of finding duplicated covers that happened today as well, and the arrival of two identical books (one ordered and one error) and it's a been a VERY coincidental day for me. I just wish it was a lottery draw day, so my numbers could coincidentally match the winning ones. BLongley 20:05, 19 June 2008 (UTC)

Error in Site Map

in Site Map, the "Top Contributors", "Top Moderators", and "Top Verifiers" pages are listed as requiring Moderator rights. IME, this is not correct, anyone logged in can go to to Moderator page, wher a notice about needing mod rights is displayed, but the links to the various "Top" pages are present and working. Is this an erro in the documatation, or is the intent that non-mods hsould not be able to see these pages? -DES Talk 16:58, 6 June 2008 (UTC)

apps in cgi-bin can be seen by anyone; apps in cgi-bin/edit can only be seen by people logged in; apps in cgi-bin/mod can only be seen by moderators. The "Top X" stuff above are all in cgi-bin, so they should be viewable by anyone. Alvonruff 17:06, 6 June 2008 (UTC)
Thanks, I have corected the site map to match the above. -DES Talk 17:43, 6 June 2008 (UTC)

John Grisham

We seem, at John_Grisham, to have a fairly extensive list of John Grisham's novels. I don't think any of these is speculative fiction, in any plausible sense. I see he has edited one anthology, with Peter Straub and Stephen King. Both of them have written enough speculative fiction that even their non-genre work should IMO be included, but I don't see that as a reason to include Grisham's novels. A hypothetical Internet Crime Fiction Database or Internet Thriller Fiction Database would surely want his stuff, but i don't think we do. Is there anywhere that this data ought to be placed/preserved, before it is deleted from the ISFDB? Is there any reason i am overlooking for retaining it here? I have read and enjoyed most of his work, but he is no more an SF writer than Agatha Christie -- indeed less so, there is the occasional ghost in Chrstie's work. -DES Talk 23:05, 6 June 2008 (UTC)

If you are going to start deleting the non-genre works of all the primarily mystery authors in the ISFDb this may well be a larger project than you expect. I don't know how some of the authors got into the database in the first place. Mysteries are, of course, almost a sister genre to sf because of their common pulp backgrounds and crossover credits. Grisham has no pulp credentials and has zero (or close to it) fantastic literature to his credit so I see no reason for him to stay but there are a few mystery writers who have minimal fantastic credits but have fairly extensive nongenre entries. Thank goodness nobody decided to put in all of Ed McBain's 87th Precinct novels--swfritter 23:36, 6 June 2008 (UTC).
I don't plan to exhastively search for such. I happend to notice the Grisham entry because I was doing a title search on another book with a title similar to one of his. There is also a difference (IMO) between an author who has written some SF or Horror and some Mystery, and one who has written pretty much nothing but legal thrilliers and mainsteam novels (and one non-fiction work about a criminal case). -DES Talk 23:49, 6 June 2008 (UTC)
As my post stated, I am not talking about authors who have done some and some. Ed Gorman, for instance, has done very little spec fiction and a lot of mystery but during his youth he was also an active sf fan and still has connections to the genre. I think there are also a number of other similar cases. No such case can be made for Grisham.--swfritter 20:08, 7 June 2008 (UTC)
I wouldn't miss John Grisham here, so long as any reference to him isn't a dead end: e.g. reviews of his work in SF magazines (unlikely) and entries in works with some SF relationship (I think you found that here). I never bothered adding my "A Time to Kill & The Pelican Brief" Omnibus (which I will be donating to the T-Mobile book club on Monday anyway) as I don't think he belongs here, and currently we can reduce him to one title. Agatha Christie keeps cropping up during my work on British Horror Anthologies though, and I think we need to keep her for the Ghost Stories - but please don't accept any OTHER works by her! And if anyone wants to clean up her current "novels" that are obviously collections, please do so. Strangely enough, Agatha Christie is SF enough to have appeared in Doctor Who in the last few weeks. BLongley
It's rare to have someone going through all the work of deleting an author back down to the relevant entries unless they're purely comic book/manga/RPG though: the last I can think of was Clive Cussler. But if someone wants rid of someone that much, I'll let them do the work. BLongley 21:10, 7 June 2008 (UTC)
Very well, I have started to delete Grisham. a good many edits will be needed. Amazing how many editions someone entered and approved. -DES Talk 22:36, 7 June 2008 (UTC)
They were entered in a by-gone era when their presence was not considered offensive. Alvonruff 01:48, 8 June 2008 (UTC)
Just trying to follow the standards as now written. Deletes complete. It took 103 edits. The by-gone era appears to have ended no earlier than late 2004, because some pubs dated 2004-12-00 were included. The earlier works generally have the most editions recorded, however. -DES Talk 06:05, 8 June 2008 (UTC)
The current approval mechanism wasn't implemented until the ISFDB-2 redesign in 2005-2006. One of the reasons why it was added was that a significant proportion of the old data was "dirty" in various ways. We are still cleaning it up, but overall things are looking up :) Ahasuerus 18:57, 8 June 2008 (UTC)

Variants and the Vance Integral Edition

I have been entering books from the Vance Integral Edition (see here). Their website ( is a key page) lists lots of useful data about these books. In many cases, when they reprinted a novel or story, they did so under a variant title which, they say, is the author's original (before editorial changes) or preferred title. It appears that they consulted extensively with Vance in preparing this edition. My question: in such cases, should the VIE title be considered the "canonical" title, particularly for works that were not frequently reprinted and are not well-known under their other titles, or should the first published title be used? Also, in some cases the VIE site indicates that a work was revised for their publications (often restoring changes made by an editor for initial publication), and in those cases i will probably be constructing "(revised)" variants. Any thoughts? -DES Talk 18:44, 10 June 2008 (UTC)

Is there any indication how significant the changes to the stories are? How many such stories would this affect?--swfritter 19:04, 10 June 2008 (UTC)
Here are the relevant notes:
  • Abercrombie Station -- "Restored from late ms: typescript fair copy of Monsters in Orbit text with holographic changes for The Best of Jack Vance"
  • Anome, The -- "Restoration sources: late ms material; 1st book publication (Dell, 1973); 1st magazine publication (F&SF, 1971); author’s map sketches. The latter explain some minor textual inconsistencies, corrected with author. The most extensive replaces this original line: ‘And one day it’s Pagane and Amaze, the next Garwiy, then off over the Great Transverse Route to Pelmonte and Whearn and the Blue Ocean’ with: ‘And one day it’s Dublay and Mourmouth, the next Garwiy, then off over the Great Transverse Route to Eye-of-the-East and the Blue Ocean.’"
  • Blue World, The -- "draft title: King Kragen. Restored from late ms, typescript expansion of The Kragen, with holographic amendments."
  • Chateau d’If -- "known as: New Bodies for Old. Restored from tearsheets from Thrilling Wonder, 1950, with holographic corrections."
  • Crusade to Maxus -- "known as: Overlords of Maxus. Draft tiles: Crazy Crusade, The New Crusade, Call for the New Crusade, Crusade to Alambar. Resolved from more than one published text, particularly Thrilling Wonder, 1951, and authorially revised, and editorially modified, version in: The Augmented Agent, Underwood-Miller, 1986."
  • Cugel: The Skybreak Spatterlight -- "known as: Cugel’s Saga (editor’s title). Restored from late typescript with holographic corrections."
  • Dark Ocean, The -- "Restored from published edition, Underwood-Miller, 1985, and final typescript with holographic corrections."
  • Dead Ahead -- "known as: Ultimate Quest. Restored from all published texts: Super Science Stories, 1950; Super Science Stories, 1951; Dark Side of the Moon, Underwood-Miller, 1986, and final draft typescript."
  • Deadly Isles, The; vol.14, 1966? 1969. Restored from setting copy for Bobbs-Merrill, 1969."
  • DP! -- "known as: DP. Restored to best available published text: Avon Science Fiction & Fantasy Reader, Avon, 1953. Spurious ending removed."
  • Dream Castle -- "authorial revision of I’ll Build Your Dream Castle. Restored from: Great Stories of Space Travel, 1963"
  • Face, The -- "Restored from late ms."
  • Fader’s Waft -- "see: Rhialto the Marvellous. Restored from late ms."
  • Flesh Mask, The -- "known as: Take My Face, and: Mask of Flesh. Single published source: Mystery House, 1957, not properly edited. Restored with participation of author. Story incorporates material from Cold Fish, 1948, an unpublished lost novel.
  • Four Hundred Blackbirds-- "Restored with tearsheets from Future, 1953, with holographic corrections."
  • Gold and Iron; -- "known as: Planet of the Damned, and: Slaves of the Klau. Restored to best available text: Space Stories, 1952. Editor’s marriage ending removed at author’s direction for VIE."
  • Hard Luck Diggings -- "Restored from: Startling, 1948; and typescript representing different and possibly earlier version."
  • House Lords, The, known as: House Lords. Draft title: A Domestic Tragedy. Restored from ms.
  • House on Lily Street, The; -- "Text never properly edited. Resolved from only published edition, Underwood-Miller, 1979, authorial errata sheets and guidance."
  • I’ll Build Your Dream Castle -- "known as: Dream Castle. Restored from tearsheets for Astounding, 1947, with holographic corrections. Story revised twice by author, both times from 1947 text. See: Dream Castle, vol.44, for 1962 revision. Second revision (Lost Moons, Underwood-Miller, 1982) less extensive."
  • Insufferable Red-headed Daughter of Commander Tynnott, O.T.E., The -- "known as: Assault on a City. Restored from setting copy typescript for Universe 4."
  • Killing Machine, The -- "Restored from typescript for Berkley Medallion, 1964"
  • Kragen, The -- "Restored from: Fantastic, 1964, and typescript pages of The Blue World. Frank Herbert discussed his desert-world story idea Dune with author, who reacted to Dune’s eventual success with this novella, later expanded into The Blue World."
  • Languages of Pao, The -- "Resolved from: Satellite, 1957, and: Avalon, 1958. Differences between these sources are authorial."
  • Magnificent Showboats of the Lower Vissel River, Lune XXIII South, Big Planet, The -- "known as: Showboat World (editor’s title). Restored from setting copy for Pyramid, 1975."
  • Man in the Cage, The "Restored from: Random House, 1960 and partial non-final draft typescript with holographic corrections on verso of The Miracle Workers typescript."
  • Man Who Walks Behind, The -- "known as: The Madman Theory. Partially restored from ms material on verso of other mss."
  • Milton Hack from Zodiac -- "known as; The Man from Zodiac. Restored from typescript for: Amazing, 1967."
  • Miracle Workers, The -- "Restored from setting copy for: Eight Fantasms and Magics, Macmillan, 1969."
  • Moon Moth, The -- "Restored from late ms."
  • Narrow Land, The; -- "Restored from typescript for Fantastic, 1967"
  • New Prime, The -- "known as: Brain of the Galaxy. Restored from late ms with holographic corrections."
  • Night Lamp -- "Restored from author’s electronic ms and errata sheets."
  • Noise -- "known as: Music of the Spheres. Restored from ms and setting copy for Eight Fantasms and Magics, Macmillian, 1969."
  • Nopalgarth -- "known as: The Brains of Earth. Restored from setting copy for Ace, 1966."
  • Palace of Love, The -- "Restored from setting copy for Berkley, 1967. Galaxy editor, Frederik Pohl, removed epigraphs which disgusted author causing delay of following two books in series."
  • Pnume, The -- "Restored from setting copy for: Ace, 1970."
  • Space Opera --- "Restored from setting copy for: Pyramid, 1965."
  • Star King, The -- "known as: Star King. Resolved from: Galaxy, 1963-4; and: Berkley, 1964"
  • Strange People, Queer Notions -- "known as: Strange Notions. Restored from typescript for: Underwood-Miller, 1985."
  • Strange She Hasn’t Written -- "Known as: The Four Johns, and: Four Men Called John. Partially restored from manuscript material on verso pages of other stories."
  • Suldrun’s Garden -- "known as: Lyonesse, and: Lyonesse: Suldrun’s Garden. Restored from setting copy for Berkley, 1986, and early ms."
  • Telek -- "Resolved from: Astounding, 1952; and: setting copy for Eight Fantasms and Magics, Macmillan, 1969"
  • To B or Not to C or to D" -- "known as: Cosmic Hotfoot. Restored from tearsheets from Startling, 1950, with holographic corrections."
  • View from Chickweed’s Window, The -- "Restored from typescript for Underwood-Miller, 1979."
  • Wannek, The -- "known as: The Servants of the Wankh; and: Wankh!. Restored from setting copy for: Ace, 1969. New authorial title for VIE, in reaction to British slang connotation. "
To save space, i have ommitted entries that mention title changes but not revisions, or that simply indicate that a specific "best published source" was followed. -DES Talk 20:12, 10 June 2008 (UTC)
Most of the time it is hard to find a justification for doing variants based on source differences without physically analyzing the stories but these are well documented. I always wondered if the use of Wankh was some kind of smarmy pun but, if so, at least it wasn't perpetrated by the author.--swfritter 20:56, 10 June 2008 (UTC)
Some further comments, from this page:
" An interesting problem is raised by ‘Guyal of Sfere’. This story, originally part of ‘The Dying Earth’ (published in the VIE as ‘Mazirian the Magician’) and written around 1944, was revised by Jack for its re-publication in the collection ‘Eight Fantasms and Magics’ in 1969. The revision was quite drastic, shortening the story appreciably and clearly affecting its colour and texture. It must represent the author’s final intention—but that presents the textual editor with two problems. First—frankly—not all would recognise it as an improvement on the original. To some eyes it represents an uncomfortable compromise between styles separated by over twenty years. Second, should we publish ‘Mazirian the Magician’ with one part heavily revised and now stylistically inconsistent with the rest? Our conclusion was to treat ‘Mazirian’ as the “text” in question, allowing us legitimately to include the original ‘Guyal’, with the revised version considered as a separate text and published in the VIE as such."
"But we have not been consistent in the way that we have treated these revisions and it would be difficult for us to articulate a single credible rationale: rather, we propose to make it clear, in each case, what the treatment of the text has been. Thus, ‘The World-Thinker’, in its revised (1981) version, and ‘I’ll Build Your Dream Castle’ in its original version, appear in the main volumes, with the substantially different version of the latter, published in ‘Great Stories of Space Travel’, 1963, appearing separately (in this volume), and the not substantially different original of the former not included. Crusade to Maxus was heavily revised for its re-publication in ‘The Augmented Agent’ in 1986, with editorial changes as well as the author’s; the VIE version has elements of both, with a major episode from the original reinstated. ‘Noise’, ‘The New Prime’ and ‘The Men Return’ similarly represent a set of judgements as between the originals and the 1969 revisions in ‘Eight Fantasms and Magics’, but take as their bases the originals. In the same volume ‘Cil’ was changed very slightly from the first version published only three years earlier, and in the VIE the story appears in its context in ‘Cugel the Clever’ (originally published as ‘The Eyes of the Overworld’) in its original form. Telek contains many of the textual changes in the revised version but restores some major excisions. ‘When the Five Moons Rise’ largely restores the original, but the differences are very minor in any case. ‘The Miracle Workers’ generally follows the revised text, though again the level of revision is fairly low."
"Where the author is known to have re-edited his text, usually parsimoniously and with the aim of eliminating the occasional awkwardness or mistake in the previously-published text, we have had no difficulty in adopting the revised text....Throughout, we have been concerned to ensure that the texts presented in the VIE are authentic Vance, and especially where later versions are known to have been the subject of editorial intervention, we have often preferred earlier versions. This is also the case with three of the stories in ‘Future Tense’, including ‘Ullward’s Retreat’ but also ‘The Gift of Gab’ and ‘Dodkin’s Job’. Although the author is known to have revised these, mainly by deletion, there are also substantial editorial changes. However, ‘Sail 25’ was considerably added to by the author for ‘Future Tense’, and we have therefore preferred that version, though we have also attempted as far as possible to eliminate editorial intervention."
"The stories collected in ‘The Dark Side of the Moon’ and ‘The Augmented Agent’ (1986) were heavily edited in an attempt to modernise the texts. With the exception of ‘Crusade to Maxus’, revised by the author, and the removal of the last line in ‘DP!’ (see below), we have not used these texts."
"Towards the end of the process of writing a book or story—probably after the end as he saw it—Jack would often comply with a publisher’s request for changes, but without enthusiasm, and minimally, making the fewest changes to the text that he could consistent with satisfying the publisher’s requirement. In the case of the text for which we have the fullest evidence—‘Emphyrio’—including correspondence with the publishers, this is very clear. In ‘Gold and Iron’, the ‘happy’ ending has been disowned by the author, to the extent of suspecting (we should remember, fifty years after its composition) that it had been added by another hand. However, there is not much doubt that the offending material is by Vance himself, and a reasonable explanation is that the happy ending was required by the publisher and provided by the author with a minimum of enthusiasm and a maximum of despatch. For the VIE appearance, the text is based on the original, and suspect, appearance in ‘Space Stories’ but, in order to respect the author’s intention, a slightly revised ending approved by the author has been incorporated. In contrast, in its first appearance (‘Avon Science Fiction and Fantasy Reader’, 1953), a grotesque final line was added to ‘DP!’ by its editor (“But I am seized by an irresistible urge to ‘tell-off’ a rotten, inhumane world…”). This was removed by the author himself for its re-publication in ‘The Dark Side of the Moon’ in 1986." —The preceding unsigned comment was added by DESiegel60 (talkcontribs) .
Help:How to record a variant title currently states:
If your two titles have not yet been linked as variants, you now have to decide which one is going to be the parent, and which one is the child. This is referred to as the "canonical title" of the work. For the Asimov example, "Foundation and Empire" is clearly the canonical title: not only is it much better known under that title than as "The Man Who Upset the Universe", but the first book publication was as "Foundation and Empire". It won't always be clear which is the canonical title, but it's a choice that can be reversed later, so it's not absolutely critical to get right first time. When in doubt, pick the title used on the first published version of the story; but if there is any question at all, make a note on the author's project page on the Wiki as to why you selected one title as canonical. If the titles match, but there is a pseudonym on one of them, use the one that is the canonical author name by preference. If neither do, then use the earlier one.
Admittedly, not the most exact guidelines, but the bias is to use the best known version, which, in Vance's case, would be typically one of the pre-VIE versions. If and when the VIE becomes better known, we could make the case that its titles should be "canonicized", but until then I suspect that they would confuse most readers. Ahasuerus 01:01, 11 June 2008 (UTC)

Images Redux

I've uploaded an image, and linked it to a pub as an experiment. Image height reduced to 561 pixels, then compressed for web. Final size <20 kB. I have hundreds of images that I'd like to upload, but I'm not clear on the Fair Use issues, server issues, and what to do with, and how to use the "fair use" templates. Anyone care to enlighten me?--Rkihara 07:20, 11 June 2008 (UTC)

  • On Fair Use issues, it is pretty clear that using scans of cover images on pages of bibliographic data about the books is Fair Use under US law. Exactly what we need to do in the way of attribution is unclear. The templates I created are based on what the English language Wikipedia does, and may well be over-kill for our purposes. It would be a good idea if images were uploaded with reduced resolution, say down to under 100 dpi or a max size of under 5-600 pixels on the long side. This will also reduce server demands.
  • On server and backup issues, i am not the person to speak. My understanding is that Al and Ahasuerus don't seem to think this is a problem for reasonable sized images.
  • On "what to do with": Using the current setup, you can upload images using Upload file. You can then capture the URL of the uploaded image, and add it to a pub record just as you would the URL of an image from amazon or visco.
  • On current fair use templates, the main one is Template:Cover Image Data. To use it in its current form, follow these steps.
    1. Navigate to an image description page, such as Image:IAS 1980 09 Schomburg.jpg (By the way if we are to upload many images, we should probably agree on conventions for image file names.)
    2. Edit the page. Insert the following skeleton:
      {{Cover Image Data
      |Title=<publication title>
      |Edition=<date and/or publisher>
      |Publisher=<Name of publisher>
      |Artist=<name of cover artist>
      |Source=<origin of image>
    3. Fill in the data. For any item where you don't have/know the data, delete the line and that parameter will default if possible.
    4. save the page
  • The current templates have a "resolution" parameter, but the wiki already displays resolution in pixels, so this is redundant and i will be taking it out.
  • If we determine that the above is more data than we want to display on the image description page, the templates can be edited to reduce the info displayed. Existing uses of the templates will still work.
  • The other currently available templates are Template:Cover Image Data-PD Old and Template:Cover Image Data-PD Renew. Both are for rather unusual situations, and no book printed more recently than 1964 will use either.
  • All three templates display documentation, and link to examples, on the template pages. Comments are welcome. -DES Talk 15:21, 11 June 2008 (UTC)
The above long response edit-conflicted with Rkihara's 2nd edit to this section, it was composed before that edit. -DES Talk 15:22, 11 June 2008 (UTC)
In the case of the image linked to at the start of this section, the filled in template call would be soemthing like
{{Cover Image Data
|Title=Isaac Asimov's SF Magazine
|Edition=Sept 1980
|Publisher=Davis Publications
|Artist=Alex Schomburg
|Source=Scan by User:Rkihara
For a display of this, see the image page linked to above, or [[Template:Cover Image Data/Example.
I hope this is helpful. -DES Talk 16:09, 11 June 2008 (UTC)
Thanks! I got it. One question, where is the dividing line between high and low res? The scans I'm uploading have been reduced to 72 dpi.--Rkihara 16:45, 11 June 2008 (UTC)
I would say that anything less than 250 to 300 dpi would probably count as low res. For print purposes, a commercial publisher generally wants images of 600 dpi or better, as I understand it. The real dividing line is "Would this image, if printed at the full size of the the original cover, be of commercial or near-commercial print quality?" if the answer is clearly no, then it is low-res. -DES Talk 16:57, 11 June 2008 (UTC)
On image labeling, I think that using the pub tag might be best, so my first uploaded image should have been ASFSEP80.jpg, or .gif, etc.
I'm not sure that is the best idea. One problem with it, it is not always possible to predict the pub tag knowing the name and date of the publication. Another problem, in some cases multiple publications have the same cover (this is more common with books than magazines) and so the pub tag becomes ambigious. yet another, i think that on more recently entered pubs the pub tag has been replaced with the purely numeric pub record number, which probably will not make a good file name (although it could be used.
However, image files can always be moved if needed, and a naming convention, while sueful, is not strictly required, and can always be modified if need be. -DES Talk 17:24, 11 June 2008 (UTC)
It would be painful to modify after use though - even if there's some Wiki feature to mass move a whole set of images and links automatically, it won't help change the ones in the ISFDB database itself. I presume we don't want to do 15,000 individual pub edits so it would be wise to set up a future-proofed directory structure and naming standards BEFORE starting to use them, probably even before starting to upload them. BLongley 17:36, 11 June 2008 (UTC)
I'll be pretty much uploading only magazine covers, and I think using the magazine tags for images will not be a problem. Books are probably a lot less predictable, and I'm not sure how you would address reused covers. With magazines recycled cover art would be uniquely identifiable by the overprinting. In any case, I won't be uploading any more cover scans until this plays out.--Rkihara 18:02, 11 June 2008 (UTC)
Reused covers for books are either identical and we can link several to the same image, or close and can be dealt with in notes (e.g. when a cover is identical apart from a price/serial number in the small print) or very different and need a new scan. FINDING a cover to reuse will be the problem if we let the Image directories get too big or the naming standards too wide. BLongley 19:33, 11 June 2008 (UTC)
Don't be too sure about Magazine Tags either: there could be a lot of work to regularize those before you can predict what the image name would be. E.g. "SF Impulse" seems to have led to Tag Prefixes of "IMPULSE" and "SFMPLS" at different times. Manual entry over automatic generation maybe? BLongley 19:33, 11 June 2008 (UTC)

(unindent) For magazines, there is not likely to be much reuse, so using tags as filenames could work, but it makes it hard to find images without first searching on the pub, and pretty much impossible to upload an image before entering at least a placeholder for the pub. If those limits are not a problem, this could work. For books, reuse is more likely, and having several pubs (with different tags) using the same image makes it seem dubious to me to use any one of those as the image file name. The same limitations would also apply as for magazines. However, please remember that as long as file names are unique (which the wiki software will enforce) having a few which violate conventions is not fatal -- we can always set up redirects or index pages. You mention "file directories" -- as long as we are using the wiki's file upload, it sets the storage path, and the entire "Image:" namespace is in effect a single directory, and any file name must be unique across the entire namespace. (we could use name including slashes to simulate directories, but we don't gain all of the benefits of true directories). We could, however, use categories to group images by artist, by date, by uploader, or in whatever way we like (and are willing to record) -- remember that a wiki page can belong to multiple categories, and that cats can have sub-categories, which can be organized into a hierarchy or partial hierarchy. Wiki cats, unlike file names, can also be added or changed at any time. See Category:Fair use images for an example. -DES Talk 23:11, 11 June 2008 (UTC)

OK, I've tried looking at Category:Fair use images and they're ALL broken thumbnails. Please, suggest how we're going to make this usable as I support the idea of taking control of our images in general, but cannot see how it's going to work. I can't "remember" "that a wiki page can belong to multiple categories, and that cats can have sub-categories" as that's never been explained to us in the first place. Today I concentrated on my verified pubs without coverart, and it is still far easier to fix those via an Amazon upload than it is here. BLongley 21:51, 14 June 2008 (UTC)
Although the thumbnails are broken, the images are still listed, and a click still takes you to the correct image page. I suspact that the wiki software is making assumptions about where it can create temporary thumbnail image fiels on demand, and we need to either create the proper directories on the server, or change the wiki's configuration settings to tell it whare it can create thumbnails. i will send a note to Al about this.
As to how setting up images on the wiki could work, it depends on how we want to organize things. We could sort the the images into categories by artist, by year of publication, by publ;isher, by author of book illustrated, or by whatever criteria we please. (We could even sort by dominant color, but I see no point in such a classification.) The image data templates could be modified to put the images into any category (or group of categories) based on data that is collected for thsoe templates, so a separate step would not be needed. Ctegories could then be used to find the relevant images, and be manually organized into whatever super-categories seem helpful. I'll go into more detail about possibilities shortly. -DES Talk 18:35, 15 June 2008 (UTC)

(unindent) Al has fixed the thumbnail problem -- for the time being, thumbnails do not appear, only links, but at least no ugly error msgs appear. See Category:Fair use images now. In addition, i have modified Template:Cover Image Data so that it places any page it is used on into a publisher category, and an artist category. This would allow finding all covers published by a given publisher, or created by a given artist. To make these work well, use of canonical publisher (and artist) names will be helpful, but not essential -- variants can be grouped into super-cats if this seems worthwhile. It would be easy enough to add a year published category, if anyone thinks that would be a useful way to sort images, and indeed if that is done a cross-cat for artist by year would be easy enough. Do people think that sorting by year would be worth entering the year for each image uploaded (with a default of "unknown" for items not entered)? None of this has any impact on the choice of file names for images. The pub tag is one obvious choice, or a semi free format in the form of title-artist could work in many cases. But I am amenable to whatever choice others make on this issue. -DES Talk 21:41, 17 June 2008 (UTC) For examples of the new categorues, see Category:Publisher:Pamar Enterprises, Inc. Images and Category:Artist:Stephen Hickman Images. The conventions i have used are for the artist categorys to be named "Artist:<Name of Artist>" and the publisher ones to be named "Publisher:<Name of publisher>" But this can easily be changed at any time by editing the template. -DES Talk 21:46, 17 June 2008 (UTC) On the actual file names, the Mediawiki help page says:

Whenever an image is uploaded, several things are created:

  1. An article in the namespace image with the exact name of the file, e.g. Image:MyPicture.png. This article is stored and behaves like any other article.
  2. The file itself is stored in a folder of the hosting (unix) system.
  3. If the file is wider than 800px or higher than 600px, a thumb with either 800px width or 600px height will be created. The thumb is stored in the folder pathofwiki/images/thumb/x/xy/MyPicture.png/MyPicture.png. Each thumb gets its own folder by the name of the original picture. Any time you create a thumb or resized image within an article, another thumb is created and stored here, prefixed by its width in px, e.g. 800px-MyPicture.png.

x and y are the two first characters of the md5 hash of the final image filename.

Apparently a (possibly) separate folder is created for every image, with parallel folders for thumbnails (if created) and over-written files, if any. This meas we will samve some server space if all imgaes are less than 800 px high and less than 600 px wide, as this avoids the thumbnails. -DES Talk 21:53, 17 June 2008 (UTC)

Circulation statements

Hi -- haven't been around for a long time, but I thought I would stop by to comment on the magazine data. I've been writing articles for Wikipedia on magazines (Authentic, If, Imagination, Fantastic Universe, Beyond, Venture, Space Science Fiction) and one thing I've done on a couple is dig up the circulation statements. Starting in about 1960 the law required circulation figures to be posted yearly, and these can be revealing -- take a look at the histogram I put together for the article on If, for example, which clearly shows how poor a job Ejler Jakobsson did. Anyway, these circulation statements are not indexed by the ISFDB. If I recall rightly, I even wrote up some notes on the ISFDB wiki asserting they should not be. I now wish they were -- currently I have to dig out every issue from my shelves, and page through each one -- they weren't published reliably at the same time each year, and they're not in the ToC, so they're hard to find.

From the point of view of a researcher, this is information that would be great to have indexed. If the ISFDB intends to support researchers, it would be terrific to add this. (If you do, please do Amazing next -- I'm working on that this summer!) Just a thought. Keep up the good work, by the way; I use the ISFDB all the time and it's a terrific resource. Mike Christie (talk) 02:21, 13 June 2008 (UTC)

Welcome back stranger! :-)Kraang 03:33, 13 June 2008 (UTC)
I'm working on filling in Amazing stories from 1960 to present, and I've filled in a lot of the pre-fifties pulps. Swfriter and I did the fifties, so the issues should about 85% done in a couple of months (people with deeper pockets can do the rarer issues). On the circulation statement, I can put the page numbers as I run across them into the notes. I think putting them into the index would not go over well.--Rkihara 04:40, 13 June 2008 (UTC)
As an afterthought I started putting them into If. As an example see the note field although I did not make title entries or enter the page number. Once I had gotten done I wished that I had. I see no reason why they should not be listed. Mike, it's been a pleasure adding a little bit of polish to the fine foundations you built on a number of mag titles. I will be doing a secondary verification run on 60's Amazing and I can at least add page numbers to the notes.--swfritter 17:24, 13 June 2008 (UTC)
Thanks for the compliment, but I think even by the time I'd left you'd done more than I had, and I know by now you've long passed me. I would like to have more time to work here, but for some reason Wikipedia has claimed my attention for a while now. As for the circulation statements, maybe another option is to make a note of them on something like the magazine title page in the Wiki? What a researcher wants is the numbers, and the date of the magazine in which those numbers appeared. The page number is nice but not crucial. I believe all statements in different magazines relate to the same time period, so there's no need to note the "sworn" date (it wasn't a calendar year, it appears). Mike Christie (talk) 00:35, 14 June 2008 (UTC)
That's to say, I don't need to go to the magazine if I have the numbers, so this data is somewhat different from a story, where ISFDB is just an index. Here you'd be listing the whole of the (relevant) contents, so there's nothing really left worth indexing too. It's more like listing the editor of the magazine. Just a thought. Mike Christie (talk) 00:37, 14 June 2008 (UTC)
The database might possibly last longer than the wiki so I don't think it is inappropriate to also make a brief notation in the pub notes.--swfritter 21:20, 14 June 2008 (UTC)

[unindent] Locus has been printing the average circulation of all major SF magazines since 1967 in its annual wrap-up issues. If there's any in particular that you'll looking for just ask, and I'll look it up. MHHutchins 23:01, 23 June 2008 (UTC)

"Bonus" stories

For context, please see this thread. When a novel is published bound with a single work of short fiction, we often leave the pub record as of type novel, not collction or omnibus, calling the short fiction a "bonus story". But we have not really documnted the practice anywhere, and a strict reading of Help:Screen:EditPub would lead one to think that thees should be collections, or perhaps omnibuses. Is this practice standard enough that it should be documented in the help screen? are there any limits or restrictions that should be applied? -DES Talk 22:12, 16 June 2008 (UTC) See also Talk:Data Consistency/Novel-Collection Mismatches. -DES Talk 22:16, 16 June 2008 (UTC)

Some other books that trigger a Novel/Collection mismatch are the ones with an excerpt or preview of a forthcoming title as well as the main book. These aren't always just excerpts - e.g. here the preview is of a collection, and the excerpt is a complete short story. One bonus story, or preview, is fine in a NOVEL by me. BLongley 06:10, 18 June 2008 (UTC)
Then should we change the relevant help page to say so? Any objections to doing that? -DES Talk 15:29, 18 June 2008 (UTC)
I have no objections, except that things aren't always easy to find in the Help Section. I would like to see notes about this placed in the Novel, Anthology, and Collection Help areas.CoachPaul 20:39, 18 June 2008 (UTC)
I will alter the template that covers the "type" field, which is used by several help pages. -DES Talk 21:02, 18 June 2008 (UTC)
Done. Are there any other help pages you think should be changed to match? -DES Talk 21:15, 18 June 2008 (UTC)

Change in author link template

I have changed Template:A so that it will automatically convert spaces to underscores, and the user no longer needs to do so. -DES Talk 21:48, 18 June 2008 (UTC)

policy on bio pages

I have made a proposal on Help talk:Contents/Purpose to clarify our policy on the use of the "Author:" and "Bio:" pages in the wiki. Please take a look and indicate your views. -DES Talk 15:56, 19 June 2008 (UTC)

Any opnions on this? -DES Talk 03:53, 21 June 2008 (UTC)
It looks reasonable, but the section which says "For those authors who do not have and are not eligible for Wikipedia articles, or whose Wikipedia articles might plausibly be deleted in future ..." assumes that the user is at least vaguely familiar with Wikipedia and its policies, which may not always be the case. Perhaps we could add something about notability/notoriety or lack thereof as the main criterion? Ahasuerus 04:52, 21 June 2008 (UTC)
Fine, I can do that. -DES Talk 05:16, 21 June 2008 (UTC)
Discussion of wikipedia notability criteria, with quotes from and link to the relevant guideline page, have been added. -DES Talk 05:31, 21 June 2008 (UTC)
I don't know whether we have been slashdotted, wikipedia'd, or what not, but looking at the Author and Bio pages created in the last few days, e.g. Bio:Michael James Makled and Bio:Ren Holton, it strikes me that these pages are often way too meandering and sometimes cutesy. Can't really blame their creators since "writers write" and what more appealing subject to write about than oneself? :) Still, a modicum of moderation (no pun intended) seems to be in order, but how do we phrase it in a polite way that wouldn't alienate potential contributors? Ahasuerus 05:02, 22 June 2008 (UTC)
I was partly getting at that when I wrote "a brief, neutrally-written biographical article" in my proposed text on Help talk:Contents/Purpose. But perhaps the matter can be made plainer. Then we can simply refer people to the help text. And if you want the extreme case, see the (deleted, but mods can still look) page Bio:Saab Lofton. -DES Talk 13:47, 22 June 2008 (UTC)
It may be partly the change in robots.txt. For example, run a Google search on "Michael James Makled". As of a few minutes ago, his summery biblio of the ISFDB was the top listing (he has one book published by AuthorHouse, a POD/Vanity press, and is apparently a grade-school child.) It is hardly surprising that if he is checking his own Google results he would find his page, and follow the obvious "Bio" link. -DES Talk 14:58, 22 June 2008 (UTC)
I have added a bit on the subject of the tone of Bio pages to my proposed text on Help talk:Contents/Purpose. Do you think this handles the matter? Does anyone object to putting my proposed text on the Help:Contents/Purpose page, and considering it a statement of policy? -DES Talk 15:25, 22 June 2008 (UTC)
At one point I noticed that our pages had been slowly rising through the search engine ranking systems ever since they re-indexed us after the TAMU migration. However, it looks like Google (and only Google) decided to give us a much higher default score a few days ago, so we comes up first if you search Google for "Michael James Makled", but other pages are ahead of us on Yahoo and MSN. Can't complain about free advertising, of course, but it means that we will be getting more newbie visitors who will need to be gently chaperoned in the right direction. Ahasuerus 17:40, 22 June 2008 (UTC)
I agree, we seem to have been Google-Ranked (which is sort of Slash-Dotted in the effect it has on us). "Satan's Moon" by Kae Charles has Wikipedia and us in the Top 10, and we've both had to remedy it (us politely adjusting, Wikipedia demanding speedy deletion). I suspect this is going to become a common problem. Time to review what new users see here first and encourage more "say hello and read the advice before editing", maybe? BLongley 20:50, 22 June 2008 (UTC)
As far as the proposed Help changes go, they look like a good start. We can always revisit the subject if this becomes a bigger issue. Ahasuerus 17:40, 22 June 2008 (UTC)

More on Image Uploads

I have added several relevant templates (see Category:Image License Tags), and I created Help:How to upload images to the ISFDB wiki‎. The various Cover Image Data templates now all place the image in categories by artist and publisher. I hope with these facilities in place, image uploads are in better shape. -DES Talk 03:52, 21 June 2008 (UTC)

Frankly, that puts me off uploading ANY images here. I just want to upload fair-use images, to a place they'll be useful, and that's it. Amazon resize Cover images for me, and give me a link within a couple of minutes. Does anyone here really want the extra "justification" steps? Does anyone want to resize their images in advance? (As it's still simpler to upload to Amazon, download the result, and upload the result here - but we can just link to the upload we did to Amazon anyway.) BLongley 22:34, 23 June 2008 (UTC)
For all the other uses (artist sigs, publisher logos etc) it may be desirable to justify them, but we've got no standard places to put them yet. I know I provided some of the test cases, used them, have we had complaints? BLongley 22:34, 23 June 2008 (UTC)
It was this same licensing/justification/etc. issue that made me stop uploading image scans to Wikipedia. Why can't it be understood that when we upload cover scans that the licensing/justification/etc. be automatically added to the page where the image is stored? Are we that afraid of being sued by the law firm of Powers, Freas and Gaughan? I agree with Bill. All these steps to add images have cooled me to the idea that excited me once I discovered we could store images here on the ISFDB. MHHutchins 23:10, 23 June 2008 (UTC)
Well I would like the categorization we get (or can get) with storing our own images. I would prefer automatic thumbnailing, but it seems that that requires the "ImageMagic" library, which doesn't come installed by default, and which Al says has a long list of other packages that it calls, some of which probably need yet others, etc, so that installing it might well take a couple of full days of his time.
Unfortunately there is no easy way to "automatically" add licensing info to the page -- the best I think we could do is allow the uploader to pick from a list, as wikipedia does -- I'll look into just how tricky that would be to implement, possibly not very.
Of course, I won't upload to amazon at all, because their customer image boilerplate makes me lie. Also, if I am going to the trouble of scanning for myself, I don't want to bother finding the page on amazon with the exact edition I am scanning -- to me that is more work than filling out the image template.
Amazon's "Terms and Conditions" for customer images say, in part:

You may only submit Materials to the Service for which you hold all intellectual property rights. In other words, if you submit a digital image to us, you must own all rights to such image or you must have the authorization of the person who does own those rights.


You represent and warrant to Amazon and its Affiliates that ...(c) the Materials submitted to Amazon by you, and Amazon's and its Affiliates' exercise of their rights hereunder, do not and will not violate, misappropriate or infringe any intellectual property right, including but not limited to trademark rights, copyrights, moral rights and publicity rights of any third party, (d) you possess all rights necessary for the reproduction, distribution, transmission, public performance, public display, and other exploitation of the Materials by Amazon and its Affiliates as permitted hereunder ... You agree to indemnify, defend, and hold Amazon and its Affiliates harmless from all claims, liabilities, damages, and expenses (including, without limitation, reasonable attorneys' fees and expenses) arising from your breach of any representation or warranty set forth in this paragraph.

Those statements cannot be true for any image of a book cover that is still under copyright. Posting it may be fair use, but the scanner does not have "all intellectual property rights", nor the "rights necessary for the reproduction, distribution, transmission, public performance, public display, and other exploitation" unless the scanner is the publisher or perhaps the artist. -DES Talk 03:41, 24 June 2008 (UTC)
I am sorry that the image templates seem like too much trouble to you. What level of effort would you be willing to accept? The templates can be changed.
If I can get a legal opinion on what the minimum we need to do to avoid suits would be, should I do so? I would have to spend some cash, I'm sure, but perhaps not more than a few hundred dollars. -DES Talk 02:45, 24 June 2008 (UTC)
Before we start spending any kind of non-trivial money on the project, we may want to make sure that we are all on the same page.
First of all, the image upload Help page looks somewhat intimidating at first, but if I am reading it correctly, it just lists multiple options and the most commonly occurring template seems to be reasonably easy to fill out.
Second, I see that a number of bibliographic sites are using recent -- and high profile -- images without explicit disclaimers associated with each one, e.g. this 1993 Playboy cover (which is safe for work) linked from this page, which has lots of Playboy thumbnails, all leading to larger images. As far as I can tell, the only disclaimer they have is "Cover images are provided by and Galactic Central" on the front page, so I wonder how they manage to get away with it? Is it because some of the relevant UK laws are different? Or have they been simply lucky that Playboy and other big boys haven't found them yet?
Third, sometimes things turn out to be less complex than they appear at first glance (we have seen it happen here), so perhaps Al may be able to install the requisite software in the next few weeks once he has a couple of hours to spend on research. Ahasuerus 04:44, 24 June 2008 (UTC)
I like your idea of a checkbox at the time of upload. That would keep the uploader from having to go back and update the page with the proper licensing and justification. Sounds like a simple enough solution for me, since I set my scanner for the required resolution so that I would not have to resize before uploading. Thumbnailing isn't an issue IMHO because the system's resizing on the pub page looks OK to me, and with just a click you get the 500 pixel image. MHHutchins 03:49, 24 June 2008 (UTC)
I did say a pull-down, not a check books, i.e. a choice from multiple options. That might be made to work, although we would probably lose the catagorization of images. -DES Talk 04:57, 24 June 2008 (UTC)

Two different covers for Galaxy issues circa late 1973

I've been entering issues of Locus and came across an interesting note in Issue 151 (December 22, 1973): "Collectors please note: you're not going mad. Galaxy has been using two different covers on the same issue as an experiment to see what happens to sales. One is the normal Brian Boyle Studio covers and the other is a conglomeration of Jack Gaughan interiors surrounding a contents page." All of my issues are the Boyle covers. Has anyone ever seen (or even heard about) any of the Jack Gaughan issues? MHHutchins 01:07, 23 June 2008 (UTC)

Hm, I think I recall some weird looking Galaxy covers in that time frame, but I'll have to check my collection on Friday. Ahasuerus 01:25, 23 June 2008 (UTC)
My November 1973 issue seems to be such a creature. The table of contents credits the art to Brian Boyle Studio and states the it illustrates "The Dream Millenium" while the cover actually has a series of black & white illos that are reduced from Gaughan's interior work; they surround a contents page. The other issues I own from this time frame are Boyle. If I remember correctly some of the later issues of 50's Future had a similar layout strategy and in neither case is it very appealing.--swfritter 14:08, 23 June 2008 (UTC)

Weisskopf cannonical name

According to this locus interview and this wikipedia article, Toni Weisskopf and T. K. F. Weisskopf are the same person. Most of her publications seem to have been under "T. K. F. Weisskopf", but she is by far better known as an editor and now publisher with Baen Books, all of which work seems to have been done under "Toni Weisskopf". In view of this, should "Toni" be the cannonical name? -DES Talk 15:25, 23 June 2008 (UTC)

I think she is usually referred to as "Toni" by writers and editors, but as far as I can tell she uses the "T. K. F." form of her name in print -- see the linked page above -- so it looks like our users will most likely know her as "T. K. F. Weisskopf". Not a big deal either way, though :) Ahasuerus 18:30, 23 June 2008 (UTC)
But when she is mentioned on the Baen web site (which is often) and when Baen books have carried an "Edited by" credit (which few prople read anyway) she is "Toni". Here Wikipediaz page is under "Toni". But you are right that it doesn't really matter which way we go here, we just need to pick one. -DES Talk 18:46, 23 June 2008 (UTC)
It's currently less work to go with T. K. F. but I think it can wait till we see if "Edited By" support appears anytime soon. I do feel a bit guilty when I demote an editor of a collection to notes. BLongley 19:52, 23 June 2008 (UTC)
At least then the data is not lost, and if we ever do have support. a proper pesud relationship will at least have all the data in one place. -DES Talk 02:11, 24 June 2008 (UTC)
I suspect that we will get the ability to delete/reverse Author Pseudonyms before we are able to add editors to novels and single author collection, so I think that it would be better to link the two Author records (one way or the other) to ensure that our users can see the data in its totality rather than to leave them separate. Ahasuerus 03:36, 24 June 2008 (UTC)

Worldcat issues

This "image" discussion above prompted me to check Worldcat's terms of service, which state, among other things:

The following activities are prohibited and you agree not to engage in (or permit) such activities:
(ii) use of Data for cataloging;
(vi) permanent or long-term storage of Data (including, but not limited to, creation of or repackaging in a database containing material amounts of Data).

This is obviously a Bad Thing from our perspective and, if it means what I think it means, may require a fair amount of work to correct by replacing Worldcat data with similar data from other, less restrictive catalogs like Melvyl, Toronto, etc. (Once we have confirmed that they are indeed less restrictive, that is.) On the plus side, the process of identifying Worldcat-derived records in our database can be easily automated and I have a half-baked library catalog crawler which may help with finding similar records in online library catalogs. Ahasuerus 04:21, 24 June 2008 (UTC)

P.S. I have run a couple of quick scans and it looks like we have over 1300 unique Notes which mention OCLC and 186 (mostly overlapping) Notes which mention Worldcat. This isn't too bad, really, and can be replaced with data from other sources over a few weeks. I am sure we will still be checking Worldcat for verification and research purposes, but it looks like we won't be able to able to use it as the sole verification source. Bummer! :( Ahasuerus 04:49, 24 June 2008 (UTC)
I just read the Worldcat terms from your link above, and I think we are ok as we stand. First of all, individual data are facts and facts are not subject to copyright under US law. Collections of facts may be protected, but only insofar as there is a creative, non-trivial organization, and the individual facts are still not protected, if taken out of the protected organization. See Wikipedia:Feist Publications v. Rural Telephone Service.

The Court ruled that information contained in Rural's phone directory was not copyrightable, and that therefore no infringement existed.

It is a long-standing principle of United States copyright law that "information" is not copyrightable, O'Connor notes, but "collections" of information can be. Rural claimed a collection copyright in its directory. The court clarified that the intent of copyright law was not, as claimed by Rural and some lower courts, to reward the efforts of persons collecting information, but rather "to promote the Progress of Science and useful Arts" (U.S. Const. 1.8.8), that is, to encourage creative expression.

O'Connor states that copyright can only apply to the creative aspects of collection: the creative choice of what data to include or exclude, the order and style in which the information is presented, etc., but not on the information itself. If Feist were to take the directory and rearrange them it would destroy the copyright owned in the data.


Assessment Technologies also created new law, stating that it is a copyright misuse and an abuse of process if one attempts to use a contract or license agreement based on one's copyright to protect uncopyrightable facts.

Also, what we do is not cataloging, in the library-science sense of the term, as I understand things.
Insofar as Worldcat attempts to claim more than copyright law allows, without negotiation or express agreement, their terms are a contract of adhesion, and not binding. They have published their data, and cannot restrict reuse in that way. Resale or repackaging of the service as a whole would be a different matter, or storing what was in effect a copy of their database, or major sections of it, but we don't do that -- we extract certain specific data elements, combine them with other data from other sources, store them in a different organization, and display them in a different format. And we do this guided by individual human judgment, not in an automated manner. I think we're fine. -DES Talk 04:55, 24 June 2008 (UTC)
I might add that nothing above convinces me not to continue adding notes referencing Worldcat, or creating records derived largely or indeed solely from its information. -DES Talk 04:59, 24 June 2008 (UTC)
Well, I've never used Worldcat as "sole verification source" yet (I don't even access it via that site, so have never seen those conditions before), and although I've been quoting OCLC references often they've only been used to get a better page count or ISBN than Amazon. If they get us into trouble, remove the notes - but I think my use has been fair and compliant so far. I don't think you can claim ownership of a number - and if you can, I want 42.994. BLongley 18:51, 24 June 2008 (UTC)
I mostly go in via "first search" which has different terms, but some of the same provisions. I don't think I have ever exactly used Worldcat as a "sole verification source, but I have used it as the "sole data source" for a number of entries. By this I mean that I have constructed numerous publication records entirely from OCLC/Worldcat data. For example THRMNTHDRG1872, THTWNTNBLL1947, LSBLLNSHKS1968, THTWNTNBLL1997, THGNTBNGMZ1954, MNCNGLTTLM1998, BWLFSCHLDR1996, TLSFHRRRFR2006, BRDTCMQCVP1980, NGLSSDRKLP2005, THHSBTHCHR1992, and many others including the vast majority of my entries of 19th and early 20th century works, including those of LeFanu and Dunsany.
In some cases, I have combined Worldcat data with data from amazon and/or other online sources, such as THFRBDDNFR1978, SQRRLHTLTM1979, THWRLDTHNK2005, MDMCRWLSGH2000, BRDTCMJZHX1975, BRDTCMCQGH2001, JRRTLKNSND1977, and many others, including all my entries of works for the Vance Integral Edition.
Google finds 29 items on with the text "Entry based on data from OCLC/Worldcat". I suspect that most of these are my entries.
While it might be possible to use the Library of Congress site for some of the same work, I find the OCLC site much better, even if I had to avoid the "First search" site. -DES Talk 21:11, 24 June 2008 (UTC)
Well, the "half-baked library catalog crawler" that I mentioned above will eventually do some of that for you, including finding new publications as well as checking page counts, ISBNs, etc against a variety of library catalogs, the Library of Congress among them. Unfortunately, I am not very good at this programming thing (certainly not in Al's league) and I don't have as much time to spend on it as it requires, so it's not clear when it will be ready for prime time.
Back to the issue at hand, David's "phone directory" example sounds convincing, but I wonder whether its applicability to the world of library records has been tested in court? To use a randomly selected message posted by a professional librarian, "When an organization is a member of a bibliographic utility [OCLC was previously mentioned], that organization has permission to copy the cataloging of other members. ... The copyright for the record belongs to the bibliographic utility. The Library of Congress does not copyright its records. This is not just altruistic; the records are government publications in the public domain." Granted, even if he is right, the copyright may only cover the "presentation" part of the record, something that we are clearly not copying, but I would sleep better if we could find a court decision one way or the other. Ahasuerus 02:17, 25 June 2008 (UTC)
I don't think there has been a court decision specifically on library records, but the Feist decision has been taken in practice to apply to all kinds of directories, collections of information, and databases, not just telephone books. To again qute the wikipedia article: "The ruling has major implications for any project that serves as a collection of knowledge. Information (that is, facts, discoveries, etc.), from any source, is fair game, but cannot contain any of the "expressive" content added by the source author." and "For example, a recipe is a process, and not copyrightable, but the words used to describe it are; see Publications International v Meredith Corp. (1996). Therefore, you can rewrite a recipe in your own words and publish it without infringing copyrights."
The biggest followup was in the area of legal casebooks. To quote further:

"Although one might assume that the text of U.S. case law is in the public domain, Thomson West had claimed a copyright as to the first page citations and internal pin-point page citations of its versions of court opinions (case law) found in its printed versions of the case law ("West's citation claims.") West also had claimed a copyright in the text of its versions of the case law, which included parallel citations and typographical corrections ("West's text claims.")

"West was found by the Second Circuit in 1998 not to have a protectable copyright interest in its citations; neither to the first page citations nor to its internal pagination citations. See Matthew Bender v. West, Citation Appeal. ... In the same case, but in separate decisions in which Matthew Bender was not involved, HyperLaw successfully challenged West's text claims. Matthew Bender v. West, No. 94 Civ. 0589, 1997 WL 266972 (S.D.N.Y. May 19, 1997), aff'd, 158 F. 3d 674 (2nd Cir. 1998), cert. denied sub. nom. West v. Hyperlaw, 526 U.S. 1154 (1999). West lost to HyperLaw in its appeal to the U.S. Court of Appeals for the Second Circuit and certiorari was denied by the U.S. Supreme Court.

The court noted that West had:

"concede[d] that the pagination of its volumes - i.e., the insertion of page breaks and the assignment of page numbers - is determined by an automatic computer program, and West [did] not seriously claim that there is anything original or creative in that process. … Because the internal pagination of West's case reporters does not entail even a modicum of creativity, the volume and page numbers are not original components of West's compilations and are not themselves protected by West's compilation copyright." 158 F.3d at 699.

The court also held that a different database which held the same information as the origianl, and from which the original could, with effort, be reconstructed, was not a copy of the original work, and not a copyright infringment as such. 158 F.3d at 703. See Copyright in Cyberspace: Requirements for Copyrightability

"Another case covering this area is Assessment Technologies v. Wiredata (2003)[7], in which the Seventh Circuit Court of Appeals ruled that a copyright holder in a compilation of public domain data cannot use that copyright to prevent others from using the underlying public domain data, but may only restrict the specific format of the compilation, if that format is itself sufficiently creative. Assessment Technologies also held that it is a fair use of a copyrighted work to reverse engineer that work in order to gain access to uncopyrightable facts. Assessment Technologies also created new law, stating that it is a copyright misuse and an abuse of process if one attempts to use a contract or license agreement based on one's copyright to protect uncopyrightable facts.

From Copyright for Indexes, Revisited comes the following:

Hot on the heels of the Feist v. Rural decision, The Wall Street Journal (Sept. 24, 1991, B1) reported that "a federal appeals court ruled that listings from yellow-page phone directories can be copied by competitors as long as changes are made in the way the material is organized.... The federal appeals court in Manhattan ruled that no copyright infringement occurs when the competitor lifts some of the information but, for example, puts the telephone numbers under different headings.... Michael Epstein, an attorney in intellectual property law at Weil, Gotshal & Manges in New York, said that the decision could give publishers 'free rein to use the facts that are located in a compilation...if the arrangement or selection is different.'"

See also Lee A. Hollaar on Misuse of Copyright. When a copyright holder conditions licenses to use copyrighted content on a contract that attempts to give the copyright owner more rights than the copyright law affords, the courts will find "Misuse of copyright" and will decline to enforce even the otherwise valid copyright, in effect placing the work into the public domain. Note that this may apply even when the accused infringer never signed such a contract, it is enough that the copyright holder had routinely required others to do so. The terms of use stated by Worldcat, in purportign to restrict the use of uncopyrightable facts as a condition of access, probably fall into this case, thus completely voiding their copyright.
See also Sourcebook on Intellectual Property Law By Groves, Peter J Groves pp327-328.
See also Supreme Court Narrows Copyright Protection for Factual Compilations (newsletter of a law firm] where it is said:

It is clear that this [Feist] decision has implications far beyond the particular subject matter of the case, and applies to all factual compilations, regardless of their medium. Some of the most obvious implications are in the area of computer databases. Take, for example, Lotus's recently abandoned CD-ROM database product, "Lotus Marketplace: Households," which was to have provided information on demographics and prior purchasing behavior for millions of Americans. Under the Feist decision, Lotus would have had only a weak (if any) claim to copyright protection for this work. At the very least, other companies would have been free to copy the facts contained in the database, so long as they did not copy Lotus's arrangement of these facts.

In short, there are lots of situations (I have quoted only a small number) where reusing facts from a directory or compilation of factual data has been held not to be an act of copyright infringment, and where it has also been held that an attempt to precent such copying by contract cannot stand either. -DES Talk 16:38, 25 June 2008 (UTC)
The part that they can probably claim copyright on because it is creative is the part we don't care about because we've decided that ourselves. Thats the classifications the book is in. Just don't copy all those into notes or tags. Dana Carson 01:45, 26 June 2008 (UTC)
Thanks, David, it sounds like we would have an excellent chance if the issue had to be decided by the courts, although I am not sure that we would be able to afford an adequate level of legal assistance. Let me leave Al a note to see what he thinks... Ahasuerus 03:12, 26 June 2008 (UTC)
Having been on our patent committee for a few years, as well as being involved with numerous IP transfers, I have some familiarity with this topic. While I am not a lawyer, I have spent many hours with them, and I'm fairly sure that the first thing they would say is that this discussion should not be happening in a wiki since it's discoverable. Since that cat is out of the bag - there are actually 3 topics here: what copyright law says about the redistribution of facts, the licensing terms associated with using WorldCat, and the appearance of whether of not the ISFDB is in compliance with both copyright law and the licensing terms.
While US copyright law is clear on the redistribution of facts, European law is not so clear (see the Wikipedia article on Directive 96/9/EC of the European Parliament). Even with US laws, some care needs to be taken in which bits are facts and which are not. Data which is easily measured from the publication are clearly facts: isbns, page counts, authorship, reprint information, etc... Items such as classification data, synopsys details, and call numbers are not knowable from measuring the book, and are all examples of value-add information which could be copyrighted.
The WorldCat Terms and Conditions make it clear that data redistribution violates the licensing terms of utilizing the web site. In some cases the reason for this is quite clear - they are redistributing 3rd party database information which IS copyrighted, and these are their terms of use. In other cases, it is either because they are making copyright claims on the WorldCat database (as they would have the right to do so in certain European countries), or they don't want to go to the trouble of specifically listing the specific rights on each record. In any case, it doesn't matter whether the data can be copyrighted or not - WorldCat is expending resources to make that data available to the public for a specific use, and have given us their terms for using their website.
Appearance of compliance is always important - unless you're the ACLU and specifically want to go to court. A notice in an ISFDB record that noted: "all data taken from WorldCat" or attributions such as "page count and isbn taken from WorldCat", would appear to be in violation of their Terms and Conditions. An ISFDB record that noted: "verified against WorldCat", might be okay. An ISFDB record that noted: "compared against WorldCat, and it's the same as our data" would not violate the Terms and Conditions. Alvonruff 11:01, 27 June 2008 (UTC)
Curiouser and curiouser... Well, given all of the above it would appear that the easiest thing to do would be to (a) identify the 1,300+ records that currently mention "OCLC" and (b) see if we can get the same data from other sources and rephrase the Notes fields of the affected records along the lines that Al suggested. The "identification" phase is pretty straightforward and anybody with a working MySQL database should be able to create a baker's dozen Wiki pages (100 affected records per page) with two just columns, the first one containing a link to an ISFDB Publication record and the second one a free text "disposition" column. The second phase, i.e. "getting the same data from other sources" may be a little tricker. We could start with the Library of Congress, which has a reasonably easy to use online catalog and whose bibliographic data is in public domain. Not only is their catalog quite large, but it can also be used by any editor without having to resort to special tools. This step should (hopefully) take care of most "OCLC" records. Once that is out of the way, I can run my top secret queries against the 2,000 library catalogs that I have ready access to and fill in most of the rest of the lacunae. At that point our "purely OCLC" records should be down to the low double digits and we can decide what to do about them.
If this sounds like a good plan, then we need volunteers to help with the "identification" phase. I could whip something up reasonably quickly, but now that I have been reunited with my collection, I would prefer to spend my "ISFDB time" on verification since there is no telling when I may have to resume my wanderings. If we don't have volunteers in the next 48 hours, I will start working on it on Monday. Ahasuerus 01:02, 29 June 2008 (UTC)
I take it that your goal is to remove "OCLC" from all db records, even where it merely records that the data has been checked against the Worldcat system, or even where it merely points the user to the corresponding record in the Worldcat system. I think this is a very bad idea, it is a case of copyright paranoia, and of yielding to an illegal and unenforceable set of "terms of use", and moreover, it is going much farther than those terms even claim to demand. If there is a consensus not to mention OCLC/Worldcat in future I will cease doing so, and simply not cite a source if some of my data was derived from or checked against their records. (Note that I do not see such a consensus yet.) I will not, however, aid in this removal project. I urge that actual legal advice be gotten before undertaking this. I am willing to spend significant amounts (say up to $500 or so) on getting such advice, from my personal funds, if people are interested in receiving such advice, and would be willing to consider it before deciding on action. But there is no point in paying for an opinion that will not be used. Let me know. (At least I will find out what an opinion would cost.) -DES Talk 01:25, 29 June 2008 (UTC)

(unindent)Oh no, I was not asking that we remove all references to OCLC. I was merely proposing that we "rephrase the Notes fields of the affected records along the lines that Al suggested". As Al wrote:

  • An ISFDB record that noted: "verified against WorldCat", might be okay. An ISFDB record that noted: "compared against WorldCat, and it's the same as our data" would not violate the Terms and Conditions.

Naturally, we would need another source of information, e.g. the Library of Congress, aside from WorldCat that we could compare with WorldCat's data. For example, our current record for the first (1915) edition of George Allan England's The Air Trust currently says "Data from OCLC record 5299578", which, as per Al, is a Bad Thing. If we changed it to "Data from Library of Congress record 16000099, compared with OCLC record 5299578, which has the same data", we should be fine.

One added advantage to this approach is that OCLC doesn't own any books per se (that I know of) and their bibliographic data comes from their member libraries' cataloging librarians, from public domain resources like the Library of Congress or from printed bibliographies like Tuck and Reginald. Note how similar the LOC record is to the OCLC record in The Air Trust case -- any guesses where OCLC's data had come from? :-) By going to the source of their data, we will not only avoid potential legal unpleasantness, but will also get closer to the underlying physical publication.

Another advantage to using library catalogs as opposed to OCLC is that they often have additional information that we are interested in and which OCLC drops. For example, OCLC record 943253 has bare bones information about the first (1891) edition of F. Marion Crawford's Khaled, but if you go to MELVYL, you will find that there were actually two editions of this book in 1891, one in May and one in November.

Having said that, OCLC is still a wonderful research tool because their data is generally more comprehensive and better indexed than any individual library catalog. When I want to know what other SF books a particular publisher is responsible for, I go to OCLC first because it only takes a minute to get a reasonably complete list. Ahasuerus 02:20, 29 June 2008 (UTC)

Well that is better than what I thought you had in mind. I still disagree, but if that is how you want to do things, so be it. (Checking OCLC data against other possibly more thorough library catalogs would IMO be a good thing, refraining from using OCLC data to create pub records is IMO a bad thing.) I take it that there is no point in my paying for an actual legal opinion? Or would any attention be paid? Similarly, (and possibly at the same time and from the same source) is there any interest in an actual legal opinion on what we have to do about fair use in images? I repeat my offer to at least investigate the costs of such an opinion, and to contribute a significant amount if the costs are at all reasonable. -DES Talk 04:05, 29 June 2008 (UTC)
I suspect that this is primarily a question for Al. If OCLC sends its legal eagles after us, the whole project will suffer, but Al's name is on the front page, so he will be Target Number 1.
As far as spending a non-trivial amount of money on the project goes, the problem is not just finding the money (which some of us could do easily), but also how that very fact would affect the project. It's been my experience that when money becomes involved in a previously all-volunteer project, it can lead to subtle shifts in contributor attitudes, not always for the better. Granted, it may be inevitable at some point, but we will want to be careful with it. Ahasuerus 06:14, 29 June 2008 (UTC)

Discussion of digest sized books

If anybody hasn't had a chance to review this discussion on the Standard page, could you please take a peek? We are discussing possibly helping Help in an area (bindings) that affects pretty much everybody. Ahasuerus 16:41, 26 June 2008 (UTC)

Reprinted Stories and Artwork

We've been having an extended discussion about how to handle reprinted stories and artwork in magazines. The problem we have is that both stories and artwork are often altered in reprinting. If the story is altered, then the publication date is sometimes changed to the date of reprint, to distinguish it from the original, otherwise it's assumed to be identical and merged. The artwork is a different story, in that the merging of artwork causes more than a few problems, so the consensus is that it's better not to do so. If we don't merge the art, we find it necessary to distinguish the reprint from the original to avoid the proliferation of identically named entries which may not be identical, as only some of the illustrations may be reproduced in the reprint, and in that case there will be no one-to-one correspondence. For example the original might have illustrations labeled "Art [1], Art [2], Art [3], . . ," while the reprint has only a single illustration, which could correspond to any of the three. It would be difficult to determine which illustration in the original the single illustration corresponds to without having the original pub. We've decided the best way to proceed is to label all of the reprinted art as a "reprint" to escape from these difficulties. Does this seem like a reasonable idea?--Rkihara 02:30, 28 June 2008 (UTC)

I might add that the main source of this problem is the Ultimate reprints from Amazing/Fantastic that appeared from mid-1965 for about a decade. This does not become a vexing issue until interior art from both the original magazine appearance of a story and a subsequent reprint are entered into the system. Cover art is a separate issue primarily because only one piece of artwork per pub is involved rather than multiple pieces per story.--swfritter 15:14, 28 June 2008 (UTC)

Under the Moons of Mars by Moskowitz

This is an important work showing the early history of magazine publishing. It uses the title of an Edgar Rice Burrough story. The problem is immense and I will need guidance. Moderators need warning. The first entry is duplicated and almost everything needs correction. I am new and most leery of what the effects could be. I have listed on my talkspace the problems as I see them. Obviously I need some one to make sure what I propose is reasonable. The work is anthology and has a last section with 19 sections which goes into depth about the stories. Those 19 sections, entered as separate titles, would give someone looking at the fiction history of the era a big break. I know, Why start something so big and difficult? The problem is that I have other material that needs to go in that cross relates. If there is an ERB person this will impact them. Thanks, Harry. --Dragoondelight 13:57, 29 June 2008 (UTC)

Many of these stories also appear in Famous Fantastic Mysteries and Fantastic Novels Magazines which were, at least initially, owned by Munsey. The complete run of Fantastic Novels is now in the system and verified and I own a large number of Famous Fantastic Mysteries. Most of the stories have the original date of publication and some of them state the original source of publication in the notes which may be an adequate methodology for our purposes. The notes can be updated with any pertinent data. The variance between the text of the original appearances and those of the reprints varies. In at least one case A. Merrit approved the publication of a version of a novel with his preferred ending. Other stories may have been substantially cut or otherwise modified. If you have any questions about the reprints in the above magazines, let me know.--swfritter 17:56, 29 June 2008 (UTC)
You may want to look at the much longer discussion of this at User talk:Dragoondelight#nder the Moons of Mars by Moskowitz General Plan -DES Talk 18:33, 29 June 2008 (UTC)

New source site

I have to recommend this site. For photographic/scanned evidence of British covers of the 1940s and 1950s, and pricing at least. I've also found corrections to authors typoed here, a missing E.C. Tubb book, miscredits, not so sure on years. The ISFDB seems a bit weak on this era and although I wouldn't offer it as a definitive source it's certainly worth a look in the same way that Noosfere seems to be pretty good at American pulps even though it's French. BLongley 21:10, 29 June 2008 (UTC)

Looks like a pretty good site, I have added it to Sources of Bibliographic Information, thanks! Ahasuerus 13:14, 30 June 2008 (UTC)


While looking for the new Contento site recently, I noticed that several editors still had his old site linked on their Talk and User pages. I'm sorry, but I don't remember which editors they were, but at least one belonged to a Mod. So if you link to Contento and haven't changed to the new site...CoachPaul 14:45, 30 June 2008 (UTC)

One of them was me. I updated my own browser's bookmarks, but forgot to change the link on my talk page. Thanks for the reminder. MHHutchins 16:31, 30 June 2008 (UTC)

Asimov's SF Magazine

Since this magazine has been woefully neglected, I've started updating and verifying from issue one (yes, another long-term project). Because there seems to be no entry standards for this particular magazine, I've started listing them on the wiki page. Please feel free to comment about what I've done so far, and make suggestions if you feel strongly that I'm off-base. I see that the magazines guys have finally arrived in the 70s (Galaxy, If, Amazing/Fantastic, F&SF seem to be coming along nicely). Their expertise in the matter of magazine entry will be very much appreciated. MHHutchins 01:48, 1 July 2008 (UTC)

Asimov's was the very first magazine I indexed (borrowing a co-workers collection) in the ISFDB. This was back in 95, before the ISFDB was available online, and numerous fundamentals (like essays) were introduced while it was being indexed, and others (like reviews and interior art) were introduced later. Much has happened since then, so expect lots of inconsistencies. Alvonruff 12:28, 1 July 2008 (UTC)
Looks good.--Rkihara 16:04, 1 July 2008 (UTC)
Looks fine. I placed a couple of comments on the Magazine talk:Asimov's Science Fiction. Doing long runs of a specific magazine title can get very tedious and repetitive. The great thing about the 50's mags from that perspective is that most had short runs.--swfritter 16:46, 1 July 2008 (UTC)

Entry Type for Letters to the Editor?

Letters to the editor are currently given an Entry Type of 'Essay'. The entry of letters for prozines has been done on a restricted basis but in order for fanzines to be fully documented it will require a loosening of restrictions. Creating an Entry Type of 'Letters' would reduce ambiguity and allow Letters to the Editor to be displayed in a separate section on the author biblio pages. If we continue to classify them as 'Essay' we will be dependent upon our editors entering data in a standardized manner or using a kludgey series mechanism to identify letters. The downside - creating a new entry type will result in the need to make manual edits to existing Letters to the Editor. I intend to hold off on continued entry of my small collection of fanzines until we have a standard for entering letters.--swfritter 17:52, 1 July 2008 (UTC)

I guess the question is: how far do we take this? As an example, Contento uses the following possible entry types in a magazine listing (see
ad 	adaptation
ai 	author's introduction to story
an 	anthology
ar 	article
av 	advice column
aw 	afterword
bg 	biography
bi 	bibliographic material
br 	book review
cf 	true confession
cl 	column
cn 	contest
cr 	criticism
cs 	comic strip
ct 	cartoon
ed 	editorial
es 	essay
ex 	extract
fa 	facetious article
fw 	foreword
gr 	game review
gs 	graphic story
hu 	humor
ia 	illustrated article
il 	illustration
in 	introduction
ins. 	insert
is 	introduction to story
iv 	interview
ix 	index
lk 	linking material
lr 	magazine review
lt 	letter
mm 	memoir
mp 	map
mr 	movie review
ms 	miscellaneous
mv 	movie
na 	novella (46-100 pages, 17,500-39,999 words)
nf 	non-fiction
nv 	novelette (21-45 pages, 7,500-17,499 words)
pi 	pictorial
pl 	play
pm 	poem
pp 	prose poem
pr 	preface
pz 	puzzle
qz 	quiz
rd 	radio review
rr 	round-robin
rv 	review (misc.)
sa 	story adaptation of a play/screenplay
sf 	special feature
sg 	song
si 	section introduction
sl 	serial segment
sp 	speech
sr 	story review
ss 	short story
sy 	symposium
tc 	true crime
te 	true experience
th 	theater review
tr 	translation
ts 	true story
uw 	unfinished work
vi 	vignette (under 4 pages, under 1,000 words)
If we're going to add letters, we might as well add any other needed types at the same time. But what level of detail are we shooting for? Having all of the above in a pulldown menu may be a tad awkward, and I would be hard pressed to tell the difference between a poem and a prose poem, or a true story/true crime/non-fiction/true confession, etc...Alvonruff 23:34, 1 July 2008 (UTC)
If I were making the choice, i would keep all our current types, and add
introduction to story
story afterword
facetious article (used for "in-universe essays" and the like)
linking material
section introduction
Oh, and I would change Chapterbook to Chapbook.
If I had to pare the list of types more, but still could add some, i would add:
comment on story (used for both itro and afterword)
biography (also used for memoir)
facetious article (used for "in-universe essays" and the like)
and keep all our current types. -DES Talk 01:40, 2 July 2008 (UTC)
I would add "play" to the list.Kraang 02:50, 2 July 2008 (UTC)
Changing "Chapterbook" to "Chapbook" is a no-brainer because it was an error to begin with. I believe that the only question at this point is how long it will take to change the software since "CHAPTERBOOK" is embedded in a number of places, including some table definitions.
As far as the newly proposed types go, that could be a major project. Adding them to the appropriate dropdown boxes would be the easy part, but it would also require a substantial revamp of the display logic to account for all possible display permutations. That could keep Al busy for a good long time (especially if we keep adding new types every few months) and I am not sure we would want to fragment our bibliography pages that much even if we could do it easily, e.g. if we had a GUI design tool for layout modifications. Ahasuerus 07:56, 2 July 2008 (UTC)
Well Al asked what types might be desirable, so i gave my view. Some of those types, if implemented, would in my view be handled exactly like "shortfiction" for display, except that they would get a marker indicating what type of item they were, and could be searched for by type in an (improved) advanced search. I do think that in displaying a pub having a type for story intros would make it clearer what works of non-fiction appertained to what stories. I do think that having a type for in-universe essay or whatever we called it would help with a number of items otherwise hard to categorize (they aren't really stories, but they aren't fully non-fictional as essays should be, either).
I don't think any of this would be my top development priority, nor would adding a type for letters. But it might well be that if new types are to be added, it would be easier and quicker in the end to do them all at once, which would mean discussing what we want. I see your point about fragmenting the display page -- if we were to add a bunch of types, perhaps we would need/want to add some sort of preferences item on which types are to be grouped together or displayed separately, if that wouldn't be too much of a coding nightmare or performance hit. I'm still in a brainstorming stage on this one, not a "lets implement it now" stage. -DES Talk 14:42, 2 July 2008 (UTC)
I would keep it at an absolute minimum. Those that can be implemented as STORYLEN fields will have minimum impact: Vignette for shortfiction and perhaps something for short filler pieces - not sure what to call that yet - miscellaneous might get overused. Implementing a specific cartoon category would result in having to modify the existing cartoon records so that the titles would not have the redundant nomenclature that would result. The current implementation for cartoons is workable. Implementing a Letters entry type would probably require modifying the existing titles but I think it is worth the effort. At some point in times letters should appear in their own section on author biblios. Perhaps we need also a generic non-linking review type that could be used for movies, games, etc. Series are an adequate way to support many of the suggested types.--swfritter 17:50, 2 July 2008 (UTC)
I'm in the middle of implementing WikiMedia-like history, which is fairly gigantic, so I wouldn't get around to new title types for a while. Alvonruff 17:52, 2 July 2008 (UTC)
My vote would be to leave out Vignette until they give out Hugos and Nebulas for them. Also, if we were to add Introduction to Story, and they all were added, no one would ever be able to view the Asimov page, as it would take several lifetimes to load. Right now, we pretty much group many of these items on the list into ESSAY. I'm OK with leaving it like that.--CoachPaul 18:07, 2 July 2008 (UTC)
A MediaWiki-like history? If my understanding of the way the application works is even remotely accurate, "gigantic" may be an understatement! Ahasuerus 03:12, 3 July 2008 (UTC)
Ah, but I have a cunning plan. (Update history/diffs are working, deciding on what to do with delete and merge). Alvonruff 13:15, 3 July 2008 (UTC)
While I sometimes run into the need for an extra type/code I don't like the idea of many types/codes. Bill Contento's list works fine for him as he's one person making evaluations of which code to apply based on rules that he understands. I like how publication binding is handled where we have a limited set of easy to remember codes and when something comes up that does not seem to fit we are free to invent a code and explain it in the notes. Marc Kupper (talk) 21:08, 6 July 2008 (UTC)

Database-Wiki integration

Considering how hard it has been for many new editors to find the Wiki, I wonder if we may want to ask Al to change the current post-submission page to something like "Submission accepted for review, please check your [Talk page] frequently for moderator feedback".

Magazines are another area where the database relies on the Wiki to a considerable extent and I have often been asked by rec.arts.sf.written posters how to go from one issue to the next/previous one. Since it's unlikely that we will be able to provide this kind of linkage within the database in the foreseeable future, do we want to add a generic link from all Magazine Publications to our Wiki-based list of magazines? Something like "For other issues of this and other magazines see the Magazine page of the ISFDB Wiki"? Ahasuerus 17:14, 4 July 2008 (UTC)

Those would be good. What would be even better, for the first case, is if when there is an unread msg on the user's talk page, a banner or other indicator displays on db pages if the user is logged in. This had beem mentioned before, I thought it was in the list of requested features, but i don't find it there. -DES Talk 17:48, 4 July 2008 (UTC)
Now i've found it: Feature:90094 Talk alert on ISFDB. -DES Talk 17:52, 4 July 2008 (UTC)
I'm still figuring out how this works on MediaWiki; after that it should be easy to implement. Alvonruff 18:40, 4 July 2008 (UTC)
That's precisely why I proposed a static link, which doesn't require querying the Wiki :) Ahasuerus 19:26, 4 July 2008 (UTC)
Even if the wiki-query is to be implemted eventually, the static link would be a good improvement in the meantime. -DES Talk 01:53, 5 July 2008 (UTC)

"T. Windling" vs. "Terri Windling"

It would appear that T. Windling the artist and Terri Windling the editor/artist are the same person. Would somebody happen to know for sure before we create a pseudonym and a bunch of VTs? Ahasuerus 02:11, 12 July 2008 (UTC)

I'm pretty sure they're the same person. I added most of the records for "T. Windling" as that was how she was credited for the artwork in Basilisk edited by Ellen Kushner, a well-known collaborator of Terri Windling. I'd say go ahead and create the variant. MHHutchins 04:12, 12 July 2008 (UTC)
Thanks, linked! Ahasuerus 00:04, 14 July 2008 (UTC)

Phillip José Farmer's The Dungeon, Book I: The Black Tower

I have this submission on hold at the moment and I wonder what we should do about it. You can only see it if you are a moderator, but basically it includes 16 new Interiorart Titles with titles like "Sketch of the Gong at the Black Tower", "Sketch of a Strange Star formation", "Sketch of NRRC' KTH", etc. We typically do not enter this level of detail unless it's a magazine or an art book, but perhaps we could profitably discuss it here first. Ahasuerus 01:32, 15 July 2008 (UTC)

I would also be hesitant about approving it. Not because of the level of detail, but because it's not clear whether the pieces are actually a "sketch" (and given a title by the submitter) or are actually titled "Sketch of...". I also don't exactly know what is meant by "ep" as the page number. Since this is a novel published in book form, I believe a single INTERIORART credit for Gould would be sufficient. Just my two-cents-worth. MHHutchins 02:02, 15 July 2008 (UTC)
My vote would be for the single INTERIORART credit. To save the work we could move the individual listings into notes.Kraang 03:27, 15 July 2008 (UTC)
From Help - "ep -- unpaginated pages that follow pagination (although generally we would expect people to count forward to find a page number)." It might be nice if there were some indication in a single entry as to how many illustrations there are; something like "The Dungeon, Book I: The Black Tower [16]" or ""The Dungeon, Book I: The Black Tower (16 illustrations)". If this was someone's favorite artist I'm sure this information would be of value to them. If we were starting over again or if there was a willingness to make the change I would suggest using a similar notation for multiple pieces of artwork for stories in the magazines.--swfritter 15:55, 15 July 2008 (UTC)
I questioned this amount of detail when I did it. It is a group of sketches preceded by a title page and a one page explanation. This section is un-numbered and immediately follows the last numbered page. The title of this section is "Selections From The Sketchbook of Major Clive Folliot". The one page explanation is uncredited but was clearly written by the author. Is there a way that the artwork can be grouped under one entry? If so, what about the one page explanation? This issue is particularly pertinent because I have other volumes of The Dungeon series, each with a similar grouping of sketches. Rhschu 18:58, 15 July 2008 (UTC)Rhschu
We have examples of other books like this and this with this level of detail. I don't actually mind INTERIORART entries in books when they have completely different captions to the fiction titles, I actually find the COVERART/INTERIORART records more annoying when they're the same titles, as they pollute the simple title search results. But we've got so many of those already I'm forced into advanced search until Al makes simple searches simple again, so it doesn't make much difference to me whether it's notes or INTERIORART. BLongley 19:39, 15 July 2008 (UTC)
I have Dungeon vI and II. The artwork is excellent. I think that it is supposed to be viewed at one time. The artist was clearly hired to do them as a set. Some of the drawings overlap pages. The commentary on a two page illustration is in one case of 5 different subjects and specific to each. The whole folio set is an example of the old time explorer sketch book with descriptions. My inclinations would be to list it as "Selections from the Sketchbook of Major Clive Follot (vol1)" as interiorart by Robert Gould. That way researchers know that they are looking at a folio collection that goes through the series. Thanks, Harry. --Dragoondelight 20:59, 15 July 2008 (UTC)
It doesn't look like we have a consensus at this point. Approve and revisit if/when the issue comes up again? Ahasuerus 05:29, 20 July 2008 (UTC)
Part of the thought process I use is to ask myself "would someone ever look for this work by its title?" If the answer is "yes," let's say it's a famous illustration or a key quote, then I'll create a title record for it. 99% of the time though illustration captions are more part of the story rather than an official "title" for the illustration. In the past I would sometimes document the actual captions as a publication note or wiki-text but more recently I've just added a note that there is a list of illustrations on page X and that the illustrations are captioned. The actual captions are unimportant and are often useless outside of the context of the story in the area immediately before or after the illustration. Marc Kupper (talk) 04:50, 21 July 2008 (UTC)

The Willows Magazine

I have a submission on hold which seems to fall within our policies and guidelines since the magazine is apparently downloadable. However, I can't seem to find its contents anywhere on the Web and the submission is also contents-free. Any ideas? Ahasuerus 01:54, 15 July 2008 (UTC)

It is actually a physical magazine. I went through the initial stages of subscribing and there is a form for the subscriber's address and shipping location. Their advertising is a bit misleading. It states that a one-year subscription beginning in January 2008 is only $10.00 but at checkout time there are $15.98 worth of shipping charges tacked on - at which time I punted.--swfritter 15:25, 16 July 2008 (UTC)
Thanks, I will go ahead, approve the submissions and add the other issues that are listed online. I am sure we will come across their contents at some point... Ahasuerus 01:22, 18 July 2008 (UTC)

All publications for one author

In case anybody is wondering why I created a big Wiki table over on User talk:Crofton99 earlier today, this new editor had requested a list of all publications by John Brunner so that he could print it out and work on verifications off-line. I put something quick and dirty together this morning for Brunner, but the same script could be run for any other author on demand. It doesn't print the cover artist (although it could be added easily) or the books' contents, but perhaps someone else may find it useful. Ahasuerus 19:08, 15 July 2008 (UTC)

I did wonder, so thanks for the explanation. I'd have put it on his User page rather than his User_Talk though, like I've done when other people request lists that ISFDB doesn't generate - I wasn't sure where to add my words to him. The Contents section seems to have gone missing on his talk page now... does anyone know how to fix that? BLongley 19:20, 15 July 2008 (UTC)
The Wiki software generates a TOC automatically when the number of sections exceeds 3, so the next section to be created on Crofton99's Talk page should make it appear automagically. Ahasuerus 20:48, 15 July 2008 (UTC)
As for the script, what language is it in? Plain SQL or something else? BLongley 19:20, 15 July 2008 (UTC)
I wish it was plain SQL so that other folks could run my scripts, but it's grimy ancient stuff which makes PDP-11 assembly look like a Lamborghini. On the plus side, when the aliens finally launch a full blown e-attack on Earth, at least my scripts will be safe because no hacker could ever figure them out... Ahasuerus 20:48, 15 July 2008 (UTC)
I'm sure I've submitted PDP-11 Assembly Language for coursework before, and probably still have a PDP-8 handbook around somewhere. I probably over-simplified "Plain SQL" too - I've never found a decent script that will run on ALL of DB2, Oracle, Sybase, MySQL, Ingres and whichever other flavours I've used over the years: and of course they'd never work on DLI or IDMS or IDMSX. My computer language skills only go back to COBOL (1968 version) and CORAL (1966) and FIXPAC (196?) but I suspect you're talking about APL. Or you're trying to make LISP sound like a super-secret. Confess now, before the Aliens from ALGOL 60 invade to defend their good name! BLongley 22:27, 15 July 2008 (UTC)
Why is it that whenever an ancient and ugly computer language is mentioned, everybody always says "Is it APL?" On second thought, never mind, no need to answer that question :) Still, in the interests of plausible deniability, I will neither confirm nor deny that I have ever heard of it! Ahasuerus 23:24, 15 July 2008 (UTC)
APL is an ancient but not ugly language -- I work in it full time professionally, and am a past chair of ACM SIGAPL. I could expound on this at MUCH greater length, but it really is off-topic. -DES Talk 03:41, 16 July 2008 (UTC)
For all I know, Ahasuerus is using grep, sed and awk alone - there's many ways to manipulate the downloaded data, and I can't say any of them are WRONG if it gets the right results. I'm not against APL either, it's just that I no longer have keyboards with the special characters on. (Although those IBM keyboards were actually in better condition than most others I've disposed of, and would probably out-last me, they just don't connect to current PCs as easily.) I'm a bit resistant to change for change's sake (there's no need to convert ISFDB to Java or .NET or Ruby on Rails or whatever the fashion is this week), but if you can use the offline data usefully in something else I'd like to hear about it. I'm as Polyglot in computer languages as I am a Monoglot in spoken/written languages. BLongley 21:57, 18 July 2008 (UTC)
Some of my early scripts were done in Perl and then Python in conjunction with MySQL queries. However, the stuff that I have done since early 2007 has been written using the "grimy ancient stuff" mentioned above. I won't mention the name simply because there are about three and a half people who still remember it (well, perhaps four and a quarter depending on how you count), so it would come close to being personally identifiable. And then I would have to install additional traps around the box with Amazing Stories Vol. 1., No. 1 and who needs that?! Ahasuerus 01:26, 19 July 2008 (UTC)
I'm sure your Amazing is safe. FWIW - the scripts that generated the DAW lists are Visual Basic for Applications (VBA) within an Excel spreadsheet. As these can't generated newlines in their output the result is run through sed which converts \n into newlines. Marc Kupper (talk) 03:55, 21 July 2008 (UTC)

Archived Talk Page

Wikiadmins and password changes

Two of the robots are hitting pages that should not be protected

Does anyone know how to change someone else's wiki-password? Marc Kupper (talk) 18:42, 19 Mar 2008 (CDT)

From times when corrupted accounts showed up at wikipedia, that probably requires direct access to the wiki db. I know an ordinary admin can't do it. -DES Talk 06:02, 20 Mar 2008 (CDT)
Once we have a newer version of media-wiki, we should be able to block the robot accounts, and perhaps semi-protect such pages if need be. (semi-protection prevents an account from editing a page unlesss that account has exsited for a certian minimum period (4 days by default) and has made a minimum number of edits (I think 25 or 50 by default). Both numbers can be configured. -DES Talk 08:57, 20 Mar 2008 (CDT)
According to there is a script to do this. From what Al has said in the past about such scripts, he can't run them on the TAMU server, but he may be able to extract from the script the SQL code needed to do the same thing directly. -DES Talk 10:40, 20 Mar 2008 (CDT)
The mediawiki site also says:
You can also use the old way, by modifying the database directly. Assuming that $wgPasswordSalt is set to true (the default), you can use the following SQL query for MySQL:

UPDATE user SET user_password = MD5(CONCAT(user_id, '-', 
  MD5('somepass'))) WHERE user_name = 'someuser';

Where obviously "somepass" is changed to the password you want to set and "someuser" is changed to the user name as it is listed in the table "user".
Note: The user_id in the CONCAT string is a column name and is not meant to be replaced with 'someuser'
Note: If you obtain a 'dbname.user table does not exist' error, please check the LocalSettings.php file, and double-check the value for the $wgDBprefix variable. If that variable is not empty, try repeating the command, replacing $wgDBPrefix_user instead of user in the UPDATE clause of the SQL statement.
If using PostGreSQL, use this query instead:

 UPDATE mwuser SET user_password =
   md5(user_id || '-' || md5('somepass')) WHERE user_name='someuser';

You can also try the Extension:Password Reset
I hope this is helpful. -DES Talk 10:51, 20 Mar 2008 (CDT)

Reviews linking to shortfiction

Amongst all the verbiage I did not find a mention of this - in at least one case I found reviews linked to a shortfiction piece rather than a novel of the same name by the same author.--swfritter 22:17, 2 June 2008 (UTC)

I have been entering reviews (for example those in The Issue at Hand that are explicitly of shortfiction works, and linking them appropriately. i don't see any problem with this. Of course, in soem cases, mistakes will happen. -DES Talk 23:20, 2 June 2008 (UTC)
Just something to keep an eye on and perhaps projectize. And there are probably cases where there are reviews of both a short story and a novel with the same title. It also means it's a good idea to check any new reviews we put in.--swfritter 00:17, 3 June 2008 (UTC)

getpub.cgi shows less info than expected.

I've developed a simple book getter using the API but noticed that the getpub.cgi proc returns less data in the query than the example for submission shows.

Notably, the image link is left out (and there may be others that I am not aware of).

Will the image link be returned in the API eventually? I realise the links are quite often old or misleading but they would be useful.

Thanks, Gary.

P.s. I tested the getpub.cgi using the isbn from the submission example (which did have an image link) and getpub returned no image link. —The preceding unsigned comment was added by Fangles (talkcontribs) 01:56, 31 May 2008.

Image and tag info are now returned. Alvonruff 11:52, 31 May 2008 (UTC)
Thank you, this is much appreciated! —The preceding unsigned comment was added by Fangles (talkcontribs) 04:56, 1 June 2008.

Is there any chance that this script or a similar script could be extended to extract title information? My use case is as follows: When I get a new book, I scan in the ISBN and run a script that extracts the publication information for my local database of what I own. I would like to be able to include a series and seriesnum field in this db.

Maybe it could be a different script, accepting the record number of a publication as an index. Or the <publication> record returned by getpub.cgi could include a title field, with the index of the title information. Md5i 01:21, 18 June 2008 (UTC)

New WorldCat Tool

Has anyone else noticed the wonderful tool that Al's added to the menu on the publication listing page? If you click on the WorldCat link, you're taken directly to the WorldCat page for this pub based on its ISBN. Great tool! Thanks, Al. MHHutchins 00:22, 4 June 2008 (UTC)

Oh yes, I was going to add it to "What's New" once we reach 300,000 titles and have another reason to celebrate :) Ahasuerus 03:59, 4 June 2008 (UTC)
Yes, looks good to me too. Ahasuerus, which title TYPES are you counting? As we're at 315,635 titles in general already... BLongley 18:09, 4 June 2008 (UTC)
The ISFDB Statistics page has 299,570 non-variant Title records at the moment. We have about 15,000 variant titles, which presumably accounts for the difference :) Ahasuerus 19:47, 4 June 2008 (UTC)
Ah, OK. I don't find that page useful myself. Well, except that "None: 0" and "CHAPTERBOOK: 5" are good signs of Entropy. BLongley 21:57, 4 June 2008 (UTC)
I'm not interested in "COVERART: 42856" as it seems people are unwilling to merge them and I'm sure we really don't have that many actual artwork differences. "INTERIORART: 23619" - well, that's not Speculative FICTION, is it? Nor are "NONGENRE: 1018", by definition. "ESSAY: 41632" might be useful for SF purposes, but might not be fiction-related, as might "NONFICTION: 3989" - I know some are Bibliographies or Encyclopedias of SF and good things like that. "INTERVIEW: 1174" - good background material I presume, but not Fiction. (OK, I suppose it depends who you interview.) "SERIAL: 2739" - I suspect most of those are duplicated by actual Novels or Novellas at least, and even if they aren't, it probably takes two or three of them to make one proper SF work. "OMNIBUS: 1908" - well, those are just confirmation of proper titles mostly. Same with "REVIEW: 26343". "EDITOR: 2004" just confirms magazines, not contents, and I know we have loads of missing Editor Records. "COLLECTION: 4807" and "ANTHOLOGY: 4703" cover shortfiction (hopefully, although I appreciate they may cover shortfiction we haven't got yet). "POEM: 8985" - not my cup of tea really, but they count for short-fiction. BLongley 21:57, 4 June 2008 (UTC)
So it comes down to "NOVEL: 48423" and "SHORTFICTION: 85364". I keep finding Dups of shortfiction whenever I try to find the Dups I've just created by adding another anthology or collection, so that's probably overstated as well. Still, the figures suggest there's enough SF already created that I can stay happy for the rest of my expected lifespan, even if we don't separate out the Fantasy and Horror I'm unlikely to read, so don't be disheartened with these stats! I think I'll be happier with seeing "edit number 1,000,000" though, rather than "300,000 titles". BLongley 21:57, 4 June 2008 (UTC)

Section creation link

I have added the "+" link to create a new section to the top of ISFDB :Community Portal. This can be done for any page by placing the "magic word" __NEWSECTIONLINK__ on the page, preferably at the top, but anywhere that it will not be erased. -DES Talk 23:59, 17 June 2008 (UTC)

Very nifty, thanks! :) Ahasuerus 01:09, 18 June 2008 (UTC)
Very nice and thank you! I should read the manual too some day... :-) Marc Kupper (talk) 05:02, 18 June 2008 (UTC)
I only found this one jsut now myself, when I was looking though the help pages on the mediawiki site for info about image display in categories. -DES Talk 15:28, 18 June 2008 (UTC)