ISFDB:Community Portal/Archive/Archive46

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 January - June 2019

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

---

Expanded archive listing

Performance - 2019-01-01

The bad robot is back. Working on it. Ahasuerus 12:13, 1 January 2019 (EST)

Can you figure out who owns the bot and contact them? ···日本穣 · 投稿 · Talk to Nihonjoe 12:57, 1 January 2019 (EST)

I am afraid not. All I know is that it claims to be a Firefox browser running on a Mac, but that's trivial to forge. When the bot started its quest for our data, the queries were coming from the Czech Republic. Once I blocked a couple of Czech IPs, the bot started using IPs from all over the world. It's possible that the original Czech IPs were also fake, so we really have very little to go by. Ahasuerus 13:10, 1 January 2019 (EST)

Proposal: Advanced Search and performance

The reason that the bad bot that we are currently fighting is affecting performance is that the Advanced Search lets you page through the whole database if that's how you formulate your query.

How about we limit the Advanced Search functionality to the first 100 pages? Are there real life scenarios where legitimate users may need more than 100 pages worth of Advanced Search results? Ahasuerus 12:48, 1 January 2019 (EST)

Absolutely yes. I've often had cases when I've paged through search results. How are the results being returned? Are they all being returned and then paged on the display side? Or are they being paged via MySQL limits? The latter is the better way for performance. -- JLaTondre (talk) 13:00, 1 January 2019 (EST)

Advanced Search pages are currently built using MySQL limits.

What would you say was the maximum number of pages that you had to page through? 300? 500? (500 pages is 50,000 records, which is a lot of records to process manually.) Ahasuerus 13:05, 1 January 2019 (EST)

Not more than a handful. I'm trying to think of a specific example and failing ;-). Usually when I'm doing clean-up and trying to find something. However, I suppose in many of these cases, I could probably have found ways to do it with a more restricted query. I've just gone with that is there. However, how are you thinking this would help? Wouldn't the bot simply change it's query parameters to return fewer results and then do more requests? Since the queries are already using limits, the performance gain from eliminating the built in paging and having the bot do its own is unlikely to be much, is it? -- JLaTondre (talk) 13:11, 1 January 2019 (EST)

I have on a few occasions pulled up searches containing hundreds of pages and gone through them in a spot-check way just to get an overview of what is there. Trying to get that overview from more focused searches would have been complicated, though not impossible. Can you display a message like "The query returns 654,399 results; only the first 20,000 will be displayed"? Twenty pages plus knowing the total number would have been good enough for me. --Vasha (cazadora de tildes) 15:36, 1 January 2019 (EST)

There is an FR (FR 882) to "Display the record count in advanced search results". It's doable, but it would slow down certain types of queries, in certain cases significantly.

Perhaps we should add a new button next to the "Next Page" button. Something like "Show the number of matching records". Also, it may be useful to add a "Previous Page" button. Ahasuerus 18:13, 1 January 2019 (EST)

"Previous Page" will be nice - although next puts the start parameter early enough in the string so it is easily visible.:) And being able to get a count before I see the results will also be very useful... And I like the two separate buttons idea - maybe even make the new button to return only the number of records (with the parameters you searched with above so you can see if you did a typo or something?) and the current search button working as before. Annie 18:25, 1 January 2019 (EST)

The first thing that comes to mind is that if the bot maintainer needs to rewrite the bot, then there is a good chance that he or she will notice the recently added note about our backups being publicly available.

Also, I have run a few versions of the queries that the bot is using on the development server. Better scoped queries -- e.g. "title date starts with 1981" instead of "title date starts with 19" do not affect performance nearly as much as the queries that the bot is currently running. I haven't done a lot of digging, but I assume that the temporary tables that the MySQL engine has to build are much smaller when using tighter selection criteria. Ahasuerus 13:24, 1 January 2019 (EST)

I'd say the same - the only case when I needed to go over more than the first 3-5 pages had been during cleanup (mainly keeping an eye on how many OCLC records are still there - the number of pages was the easiest way:)). But I could have used better search queries if we had a limit (the hard 1000 limit on the Notes search made me very creative :)

I wonder if some FULLTEXT indexing (see this) Annie 14:48, 1 January 2019 (EST)

At one point I started investigating the MySQL implementation of FULLTEXT indices. Different MySQL versions support different subsets of the functionality and I need to do more digging. Ahasuerus 18:07, 1 January 2019 (EST)

and probably even indexing on the date columns (I did not see one last time I looked at the DB but I was not looking for it specifically) in the DB won't help more than limiting paging. Annie 14:48, 1 January 2019 (EST)

Adding date indices to titles and pubs may help with certain types of Advanced Searches. However, it's unlikely to help in this particular case since the bot searches for "title date starts with 19", not something that having an index would speed up. Ahasuerus 14:55, 1 January 2019 (EST)

I have added date indices to titles and pubs. It may help speed up some types of Advanced Searches. Unfortunately, the process caused the server to freeze for over a second. Sorry about that -- I failed to consider the effect of fragmentation on the live server. Ahasuerus 18:57, 1 January 2019 (EST)

Advanced Search changes

Part 1: A "Previous Page" button has been added to all Advanced Search pages. Ahasuerus 15:51, 2 January 2019 (EST)

Part 2: Advanced Search results have been tentatively limited to 300 pages or 30,000 records. Ahasuerus 19:19, 2 January 2019 (EST)

Part 3: "Submit Query" in Advanced Title, Author, and Publication Search sections has been replaced with two new buttons. The first one, "Get Results", is functionally equivalent to the old "Submit Query" button. The second one, "Get Count" gives you a count of the records that match the entered section criteria. For example, clicking "Get Count" after entering "Title Year is exactly 8888" results in the following output: "Count of matching records: 514". Ahasuerus 20:05, 7 January 2019 (EST)

Can we reverse the Count and Results buttons? I keep hitting Count all the time when I am trying to get to the results... Annie 16:52, 10 January 2019 (EST)

It would be easy to do, but let's make sure that other users won't run into the opposite problem first. Perhaps we could move the new button further to the right? Maybe under the AND/OR buttons? Ahasuerus 16:57, 10 January 2019 (EST)

Maybe... my hand gravitates to whatever is on the right - but maybe putting some space between them will help as well. Annie 17:11, 10 January 2019 (EST)

Part 4: Error messages and informational messages have been consolidated and cleaned up. Certain invalid Advanced Search URLs should no longer generate Python errors. Ahasuerus 18:51, 8 January 2019 (EST)

Part 5: As per FR 1251, Advanced Search has been changed to allow only one "AND" or "OR" value per search. As noted in the FR:

• Currently [Advanced Search] allows a mix of "ANDs" and "ORs", which results in unexpected behavior. This will become more important as we let users specify more selection criteria in Advanced Search queries.

Previously bookmarked searches should still work as long as they contain all "AND"s or all "OR"s. Ahasuerus 15:42, 10 January 2019 (EST)

Part 6: The number of selection criteria which can be simultaneously specified in Advanced Title/Author/Publication searches has been increased from 3 to 5. At the moment, each page section displays 5 blank lines. Once we confirm that everything is working as intended, I plan to shrink each section to 1 line and display the same "+" sign that other multi-fields have been using for the last year. Ahasuerus 19:53, 10 January 2019 (EST)

I don't know if it is related to the recent changes, but the Show All Titles link on author pages are producing a "Error: Invalid Advanced Search Parameters" message. --Ron ~ Rtrace^Talk 20:45, 10 January 2019 (EST)

Oops! Let me try to fix it real quick... Ahasuerus 21:50, 10 January 2019 (EST)

The bug should be fixed now. Sorry about that! "Show All Titles" and Advanced Search use the same code for historical reasons which no longer apply. It was a timely reminder that I need to separate them as part of this rewrite. Ahasuerus 22:10, 10 January 2019 (EST)

Thanks for being so quick! --Ron ~ Rtrace^Talk 22:23, 10 January 2019 (EST)

Also, I tried Tag (does not contain) "juvenile" AND Title Type (is exactly) CHAPBOOK AND Title Year (is exactly) 2018, and got only eighteen results -- what was wrong with that? --Vasha (cazadora de tildes) 20:50, 10 January 2019 (EST)

Vasha, keep in mind that if there are no tags at all, this query won't match either - "does not contain" means "exists and does not contain". Can that explain what is going on? (not too many people tag chapbooks I think - here are the ones that have the letter a - the ones that appear empty have private tags. Annie 21:06, 10 January 2019 (EST)

Ah, if the "empty" ones contain private tags, that makes sense. Ahasuerus, is changing the behavior of "does not contain" to also find nulls (or adding an "is empty" option) one of the things you have planned? --Vasha (cazadora de tildes) 21:19, 10 January 2019 (EST)

Yup, it's our old friend Bug 690; it also has cousins like Bug 571. I hope to address them as part of this rewrite once I get a bunch of other things sorted out and streamlined. Ahasuerus 22:18, 10 January 2019 (EST)

(unindent) The new fields (4 and 5) do not change their field type when you chose an appropriate field in the first one - Language list for example. 1-3 switches to the dropdown list; 4&5 remain as text fields. Annie 10:43, 12 January 2019 (EST)

Thanks, I'll take a look. Ahasuerus 11:24, 12 January 2019 (EST)

The bug has been fixed. You may need to do a full page reload (Control-F5 in most browsers) for the fix to take effect.

There is another, display-only, bug which can be reproduced by changing "Title" to "Title Type" on the first line. The drop-down list is displayed to the right of the AND/OR radio buttons instead of to the left. Ahasuerus 15:12, 12 January 2019 (EST)

OK, I think I finally got it to work correctly. Please let me know if you run into anything unexpected. Ahasuerus 19:14, 12 January 2019 (EST)

Lao added

The Lao language has been added to the list of supported languages. Ahasuerus 14:57, 3 January 2019 (EST)

Can we add the Canopus Award?

(Copied from the Help Desk)

Although it's not clear if it's still alive, it did make some relatively high profile awards for a couple years:

http://canopus.100yss.org/

Previous awards: http://canopus.100yss.org/?page_id=63

—The preceding unsigned comment was added by Deepsettpress (talk • contribs) .

Checking their list of judges from 2015, I see a wide variety of backgrounds, including an Analog author (Juliette Wade), a former Communications Director for the SFFWA (Jaym Gates), a few PhDs/MDs, a former Wall Street Journal reporter and a major general. Looks like a legitimate award. Ahasuerus 20:51, 4 January 2019 (EST)

Can we add it then? Deepsettpress 19:12, 6 January 2019 (EST)

If there are no objections, I will add it tomorrow. Ahasuerus 20:19, 6 January 2019 (EST)

I have added the award and entered all professionally published fiction nominees. We may need to revisit it if and when the award is revived. Ahasuerus 12:30, 7 January 2019 (EST)

Valerie Nieman canonical name

I'm about to switch the canonical name of Valerie Nieman Colander to Valerie Nieman. She's been publishing under the shorter name since the late 1990s and has a new speculative novel coming out next year under that name. --Vasha (cazadora de tildes) 03:05, 5 January 2019 (EST)

Viewing submissions as raw XML

As per FR 1223, the software has been updated to allow viewing submissions as raw XML. In the past this ability was restricted to moderators. If you encounter any issues, please report your findings here. Ahasuerus 20:20, 6 January 2019 (EST)

Thanks, that'll be helpful sometimes. --Vasha (cazadora de tildes) 21:02, 6 January 2019 (EST)

Silas Weir Mitchell

The canonical name of Silas W. Mitchell clearly should be changed, but to what? On the one hand, we have lots of publications from his own lifetime in the database, all of them under "S. Weir Mitchell." On the other hand, he is in SFE3 as "Silas Weir Mitchell," and that's the name that editors nowadays are using the most when they reprint his stories (3 in this DB). Should we use the "Silas Weir Mitchell" form because it is currently the most familiar, or "S. Weir Mitchell" because it is the most numerous? --Vasha (cazadora de tildes) 05:09, 9 January 2019 (EST)

Update: I have used Google books to look into 10 other recent nonfiction books about science fiction, apart from SFE3, and all but one of them use "S. Weir Mitchell." That considerably strengthens the case in favor of "S." --Vasha (cazadora de tildes) 05:49, 9 January 2019 (EST)

"S." seems most appropriate to me, given our standard for canonical names. --MartyD 21:43, 9 January 2019 (EST)

Yep. And the change is now done. --Vasha (cazadora de tildes) 21:47, 9 January 2019 (EST)

Google Search and regular Search

The regular Search software has been updated to leverage the Google Search functionality which was implemented a few weeks ago. If a regular name, title, series, publisher or publication series search fails to find any matching records, the results page will let you re-run the same search using Google.

Somewhat surprisingly, Google's search logic isn't always very robust when searching a subset of Web pages hosted by a single site (isfdb.org in our case), but it's better than nothing. Ahasuerus 19:02, 9 January 2019 (EST)

Galactic Journey, again

Hello, folks. I could use some assistance.

We determined earlier that, as a webzine with more than a hundred articles every year, it is too cumbersome to add Galactic Journey one article at a time. At the same time, it was apparently decided that its having been a finalist for the Hugo last year justified some level of existence at the ISFDB. Some kind soul accepted the 2017 entry of the Journey, listed at 15 pages (which is the number of pages it takes up when you put http://galacticjourney.org/tag/1962 in the URL window).

So, great! The Journey is on the ISFDB.

There are three things I'd like to do now that it's here:

1) Give author credits since I'm not the only one who writes for it (including several folks who are published authors with works elsewhere in the ISFDB). I have tried putting the authors in the top and in the bottom of the edit dialogue; both options have been rejected. I have been told sternly to add each article individually. We've already determined that's not necessary.

2) Each year of the Journey get an omnibus entry, like the 2017 one. So, 2013, 2014, 2015, 2016, and 2018.

3) The Hugo Finalist page link to Galactic Journey (2017), and the Galactic Journey (2017) page indicate that the Fanzine was a Hugo finalist that year.

However you would like that to be accomplished, please let me know. :) Thank you for your attention.

--Galacticjourney 09:43, 11 January 2019 (EST)

I think that you are mixing two types of entries in your head. If you want the other contributors added, you will need to add the individual articles as essays (and reviews) inside of the main yearly publications. What we determined is that we do not want a separate PUBLICATION per article (so a new webzine for each article) but we still want a separate title - a content item inside of the Fanzine publication - so it looks similar to how we record Daily Science Fiction for example (over here - we do these montly due to the structure; if you have more than 100 articles per year, we may also need to go for an yearly title and monthly publications as well).

So if you want, you can start by using what you have now as a 2018 entry and start adding the separate posts as content elements (and rename it for the year you want it to belong to). Then we can make a series from the yearly records. Once we have the yearly record for the year you need, we can then add them to the awards properly. Let me know if that makes sense and if you need any assistance. If you want me to, I can also create the yearly records for you and make the series so that after that all you need to do is to add the contents titles. Annie 16:58, 11 January 2019 (EST)

Yowza! Well, if that's the way it has to be done. I'm just confused because underneath the link you have, there's a simple entry for 2017: http://www.isfdb.org/cgi-bin/pl.cgi?700031 -- this is what I was hoping to model rather than individual posts. As for yearly records, any assistance you're willing to offer is tremendously appreciated. Thank you so much for your quick and warm response. :) --Galacticjourney 17:16, 11 January 2019 (EST)

Again - two different concepts. :) What you will have will be 1 fanzine publication entry per year (for now); and then 1 EDITOR title inside (a special container title for magazines and fanzines). Then when you open the publication record, you will see the articles as content items - how you see them inside of one of those monthly Daily SF pubs (or how you see under a collection for example). I can setup your publications so you just need to edit to add the content - let's start easily with the year that will be getting the award connection (2017?) as it is the only year that is clearly eligible anyway - plus that way we can see how it looks and we can use that as a base for conversation. How many articles are there in that year? If we have more than 200, I think we should split in quarters or even months - but I need some numbers (and yes, I know I can check on my own but do not feel like counting just now :)). Once I know that, I will get the structure created.

We are a bibliography, not an index. Having an empty container with a gazillion names attached to it does not describe your fanzine - it just says "here, it exists". Well - not what ISFDB is all about. :)

Alternatively we can leave the only record you have now and connect that to the awards -- but then you cannot add other writers. :) Annie 17:24, 11 January 2019 (EST)

That all makes perfect sense. You really are fantastic, thank you. :) And yes, you shouldn't have to do more than the minimum work! There are precisely 150 articles published in 2017. --Galacticjourney 17:48, 11 January 2019 (EST)

OK. Here is the series where all the yearly records will go. This is now the record for 2017 - notice how the name is different for the publication vs the title (this has to do with the series view - both in list form and in grid view). Now you can edit the 2017 record and add articles. I would advice you to do them in batches (10-20 at a time) and you can use the pipe format to give them order (for example "|12" will be sorted before "|13". Or you can use a month/date notation for the page number (11.03 being the 3rd of November; you want the month first so they line up and you want the leading 0 so that 03 is sorted before 11). Keep in mind how we regularize titles (Title case with a bit of a twist: the first word is capitalized; all later words are capitalized except for "and", "or", "the", "a", "an", "for", "of", "in", "on", "by", "at", "from", "with", and "to"; hyphenated words have the first letter after the hyphen capitalized) and author names (spaces between initials, dots after initials and so on (see this for more details). How you capitalize in the fanzine itself is irrelevant - you need to regularize when adding here). Annie 18:06, 11 January 2019 (EST)

PS: And I linked the record to the award - see how you are connected now.

I think that we still need to discuss how to handle this specific case (being a non-fiction online only fanzine is a bit outside of the rules as was mentioned above but then nominations and all that...) so if you want to add content, go ahead and do it (for now). If the eligibility decision is reversed, well... :) Or you can leave it as is for a few days so we can try to restart the discussion. Up to you. :) Annie 18:06, 11 January 2019 (EST)

Thanks so much, again. I believe there are non-fiction Fanzines on ISFDB currently (for instance, FANAC). As for whether or not the Journey is fiction, there's definitely a fictitious element to it. After all, we don't *really* live in 1964. :) --Galacticjourney 18:59, 11 January 2019 (EST)

Non-fiction fanzines we have - they had been always in scope. Non-fiction web-only are a different matter - we did not accept anything web-only until a few months ago - if you do not have an ebook at least, you were not eligible at all. :) And if you look at the records of FANAC - they are here without publications - just an empty shell so we can link the awards. :) Annie 19:10, 11 January 2019 (EST)

Well, I think it's about time web-only Fanzines got included on the INTERNET Science Fiction Database. ;) Thank you, Annie! I hope you're going to Dublin. And/or Baycon, Loscon, etc. --Galacticjourney 20:15, 11 January 2019 (EST)

Speculative, not Science :) And it is a FICTION DB :) Non-fiction had always been secondary and supporting the Fiction. My current plans are to make it to Dublin but work can get a bit unpredictable sometimes so who knows :) Annie 20:48, 11 January 2019 (EST)

Advanced Search and dynamic/JavaScript-driven pages

FR 1142 "Lift the 3 search values per Advanced Search limit" says:

Instead of displaying three input lines per Advanced Search type, we should display just one line followed by a '+' button. It should be similar to the way '+' buttons work elsewhere and allow entering as many search values as needed.

I was about to start working on this FR when I realized something. If we were to implement the requested functionality, it would mean that advanced searches that use more than one selection criterion would require that the user's browser support JavaScript, i.e the ability to change ISFDB Web pages dynamically when the user clicks on buttons like "Add Author" or "+".

Up until now we have only required JavaScript when editing the data, but all of our display and search pages work with JavaScript disabled. You don't get to see some minor "quality of life" features, e.g. those dynamically changed drop-down lists on the Advanced Search page, but the core functionality is still available.

This was a conscious decision which we made back in the mid-2000s. The reasons were varied:

• Some really old browsers like Lynx do not support JavaScript.
• Many people browse the internet with JavaScript disabled by default (using NoScript or native browser controls) to minimize exposure to security issues.
• Some "spiders" that crawl the Web and index our Web pages can be thwarted by JavaScript, although major spiders have gotten better at it.

Some of these reasons are probably no longer valid, e.g. Lynx's market share is currently 0.00%. On the other hand, security has become more important over the years, which means that many users browse the Web with JavaScript disabled by default. I certainly do.

I also find JavaScript-based code to be markedly more browser-dependent; there have been times when I had to rewrite previously deployed JavaScript code because it didn't work under some older browsers. Finally, JavaScript is more time-consuming to implement, although it may have more to do with my lack of JavaScript experience.

So I guess we have two decisions to make. The first one is whether we want to change the current policy:

• JavaScript is only required for editing

to something like:

• JavaScript it only required for editing and running Advanced Searches

If we decide to keep the current policy, then we need to decide what to do about various requested Advanced Search enhancements. For example, FR 1143 reads:

Create Advanced Search options for all ISFDB record types, i.e.:

• Advanced Series Search
• Advanced Publisher Search
• Advanced Publication Series Search
• Advanced Award Type Search
• Advanced Award Category Search
• Advanced Award Search

Some of these search types would be more useful than others, but it would be nice to add all of them simply to make things consistent across the board. They are not hard to implement, so the added effort wouldn't be significant. The only concern that I can think of at the moment is that it would make the Advanced Search page too busy. We could address this issue in a couple of different ways.

The easiest one would be to turn the Advanced Search page into a menu which would list all of the available search types. The user would then click the search type she is interested in and be taken to the Web page for that search type.

Another way to handle the issue would be to display a single drop-down list with all of the search types. Once the user chooses the search type, the software would display the appropriate input fields and drop-down values. It would be similar to how the software dynamically changes the drop-down values for different advanced search criteria. The second way would require a bit more work than the "menu" approach, but it should be doable.

If we decide to keep the current policy of "no JavaScript on search pages", then the second option is off the table and we are left with the first option -- see the bolded text above. Would that work? Ahasuerus 17:48, 13 January 2019 (EST)

P.S. In the meantime I am slowly rewriting the Advanced Search software to make it more maintainable. There should be no user-experienced changes for now. If you encounter anything unexpected, please let me know. Ahasuerus 20:33, 13 January 2019 (EST)

In the past, the Advanced Search used to interact badly with certain ad-blocking plugins, but I never bothered to mention this because, for one thing, it was no trouble to turn off the ad-blocker on the rare occasions that I needed to, and more importantly, you can't well take every single plugin in existence into account when you're coding. I think I will be able to figure out which future problems are due to plugins also and ignore them—but what sort of software interactions would you like to hear about? Any widely-used scripts that should be tested? --Vasha (cazadora de tildes) 21:05, 13 January 2019 (EST)

I expect that the most widely used third party tool that can affect Advanced Search behavior is NoScript. It allows or disallows JavaScript for individual sites. Chances are that the vast majority of our regular users who have NoScript installed have already marked isfdb.org as a trusted site, which enables the JavaScript used by Advanced Search to work. I use NoScript and haven't come across any issues, so we are presumably OK. Ahasuerus 09:15, 14 January 2019 (EST)

Fantascienza as an external ID

Can we add Fantascienza as an External ID and a template - most of our Italian titles are from there - and virtually all pre-2010 ones have entries there. They have permanent links (two of them actually: one via NILF (which stands for Numero Identificativo della Letteratura Fantastica): http://nilf.it/107292 for example and one via the catalog itself https://www.fantascienza.com/catalogo/volumi/NILF107292). I am working on resolving the Italian translators and variants so I may as well get some links added (because for some books, it requires some creative work to find the match). Annie 13:15, 14 January 2019 (EST)

Sure, we can do that. There are 46 references to "nilf.it" and 151 references to "fantascienza.com/catalogo" in publication notes, but I assume that a review of our Italian pubs should be able to add a lot more.

Since the 2 types of URLs redirect to the same set of Web pages, I guess it doesn't matter which URL pattern we decide to use for this External ID Type. We can always change it later. Ahasuerus 15:34, 14 January 2019 (EST)

We have 3376 publications that mention the word fantascienza - and 99.99% (if not all) of them can have an ID added to them from what I had found :) So yeah - there are a lot more out there. And NILF is probably a good name for the identifier/template - and either link would work (that's why I mentioned both) :)

I will move to a different language temporarily until we decide and add that one - so I do not need to walk through the same titles again later :) Annie 16:29, 14 January 2019 (EST)

FR 1254 "Create an External ID Type/template for Fantascienza" has been created. Let's wait a couple of days in case we are overlooking anything. Ahasuerus 18:06, 14 January 2019 (EST)

One more reason to use the nilf.it format - the template will be usable for referencing anything on the site - http://nilf.it/16781 leads to an author page for example - the NILF numbers are used for anything in the catalog - books, stories, authors, publishers and so on. So we will have a single template covering everything easily. :) Annie 11:59, 15 January 2019 (EST)

Sounds good! Ahasuerus 13:34, 15 January 2019 (EST)

(unindent) NILF has been added as a new External ID Type and as a template. Do we know if the ID is always numeric? If it is, I can add it to the cleanup report. Also, do we need a new cleanup report to look for raw Fantascienza URLs in publication Notes? Ahasuerus 12:04, 16 January 2019 (EST)

It is always numerical. Let's hold off the raw report for now? I can clean the ones we have off search - I want to see what other patterns will show up (considering that NILF is used for other things). Then we can build the report so that new links are caught. Annie 12:38, 16 January 2019 (EST)

Or if you build the report, go for "www.fantascienza.com/catalogo" and "nilf.it" thus catching all variations and we can precise later if needed. Up to you :) Annie 12:50, 16 January 2019 (EST)

Good point. I'll look for patterns in the database. You never know what kind of monsters are lurking under the surface! Ahasuerus 13:16, 16 January 2019 (EST)

I have created and deployed a couple of cleanup reports, one for "www.fantascienza.com/catalogo" and one for "nilf.it". The data -- a couple of hundred pubs between the two reports -- will become available tomorrow morning. Ahasuerus 16:26, 16 January 2019 (EST)

(unindent) Can we add "Fantascienza" to the "Numero Identificativo della Letteratura Fantastica" description in the drop down help (Numero Identificativo della Letteratura Fantastica/Fantascienza)? That will make it easier for people to understand/connect the dots for that one. Not urgent but next time you are changing something on the page :) Annie 13:00, 16 January 2019 (EST)

Sure, I can do that. It's database-driven, so I just need to change the value of one field. Ahasuerus 13:07, 16 January 2019 (EST)

Done. Ahasuerus

(unindent)External IDs table is updated for both. I saw NILF in the templates list, are you planning to add NooSFere? Annie 14:56, 16 January 2019 (EST)

I thought I ha dadded both, but apparently not. Fixed now -- thanks for the heads-up! Ahasuerus 15:08, 16 January 2019 (EST)

The "Non-numeric External IDs" report has been updated to look for non-numeric NILF IDs. Ahasuerus 17:07, 16 January 2019 (EST)

Label changes to finish a previous change

Last year we changed the wording around the art titles to acknowledge that they are variants and not translations so now we call them variants in the contents or even when they are a variant in a different language for a cover. But we still call them translations in the other titles table. Can we change these labels so they all match (Call them "variant in another language" for example... Annie 17:20, 14 January 2019 (EST)

Makes sense. I have created FR 1253. Ahasuerus 17:52, 14 January 2019 (EST)

NooSFere as an external ID

One more for the external ID/template list - NooSFere. The URL is https://www.noosfere.org/livres/niourf.asp?numlivre=2146570233 (the number can be positive or negative). We have some older links (like this one) but if you drop icarus from them, they still work (this is the same as this. And we have a few thousand cases where we use NooSFere as a source but without a link (so these can be added). I checked with Linguist (as I cannot find a statement that the URLs are stable) and we seem to be in the clear (and we have the links anyway so may as well move them). Annie 11:54, 15 January 2019 (EST)

Sounds reasonable. Unless there are objections, I will add two new External ID Types tomorrow. Ahasuerus 13:36, 15 January 2019 (EST)

NooSFere has been added as an External ID Type and as a template. I plan to update the cleanup reports later today. Ahasuerus 12:42, 16 January 2019 (EST)

A cleanup report has been created and deployed. The data will become available tomorrow morning. Ahasuerus 16:24, 16 January 2019 (EST)

I assume it looks both for the icarus/livres and the livres only links? Annie 16:31, 16 January 2019 (EST)

At this point it looks for everything that matches noosfere.org because there are a couple of broken links that a search limited to icarus/livres wouldn't find. There are a few other noosfere.org links in notes, but they shouldn't be publication notes.

At least that's the theory. We'll find out how closely it matches reality once we process the 1,579 pubs that the report will find tomorrow morning :) Ahasuerus 16:38, 16 January 2019 (EST)

The "Non-numeric External ID" report has been updated to look for non-numeric NooSFere IDs (one optional leading hyphen allowed.) Ahasuerus 17:08, 16 January 2019 (EST)

Ignore needed in report

We either need an ignore in Publications with direct Deutsche Nationalbibliothek links in Notes or we can template as {{DNB|gnd/107976218}} and then we will need an ignore option (or an extension to the check) in Publications with Invalid Non-numeric External IDs instead :) Thanks! Annie 17:15, 15 January 2019 (EST)

URLs like http://d-nb.info/gnd/107976218 are for author records, right? If so, then should we move this URL to the author level? We could then create a human-readable name for "dn-nb.info" URLs similar to the way we handle author pages hosted by WorldCat, the Library of Congress and BNF. Ahasuerus 18:18, 15 January 2019 (EST)

Yeah, the gnd ones are the author level records in DNB. The problem with this approach will be the cases where we need to link to one of those but we do not have the author in ISFDB (or they are not eligible yet because they are a translator only for "our" titles). Which is the problem here - we do not have Georg Altlechner in ISFDB under that specific pseudonym (the Axel Jeffers pseudonym is in the DB - so we can do that)... Let me think on how to rework the note so it works. But sooner or later, we may need to resolve the issue :)But for now, you are right, we do have a faster solution so I will fix it. Thanks for the idea! Annie 18:38, 15 January 2019 (EST)

Sure thing. The software has been updated to display "d-nb.info" URLs as "German National Library", e.g. see Adolf Sommerfeld. Ahasuerus 11:19, 16 January 2019 (EST)

Marion Zimmer Bradley's Atlantean cycle / novel

I do think that the way we have that cluster may be bettered by indexing the OMNIBUSes as NOVELs:

1) For the first 'OMNIBUS' edition, Locus observed that "This is the first hardcover edition as well as the first one-volume edition. The book was written as such but split in the U.S. for various profitable reasons.", so it seems to have been intended by MZB as one NOVEL, not two. (Also the five chapters are numbered 1-5 / I-V for the OMNIBUSes).

2) The OMNIBUSes don't seem to have the two separate novel titles listed, as can be seen here, here, here (a French edition), and here (a German edition).

Would somebody like to write some commentary on the matter? Thanks in adavance! Stonecreek 09:01, 16 January 2019 (EST)

Performance problems -- 2019-01-17

There is another bad bot sporadically sending hundreds of near-simultaneous queries to the ISFDB server. The pattern of abuse is different, but the end result (slow response time) is the same. Mitigation in progress. Ahasuerus 19:49, 17 January 2019 (EST)

Maybe we could implement a CAPTCHA for queries?--Rkihara 02:42, 18 January 2019 (EST)

I am sorry, I wasn't clear. When I wrote "queries", I meant all kinds of browser requests for random ISFDB URLs: title look-ups, author look-ups, author directory views, attempts to edit random pages and so on.

Unlike the last bot, which was at least trying to do something meaningful (download ISFDB data using Advanced Search), this one was all over the place. It's not clear whether it was a malicious attempt to overwhelm the server or whether the bot operator was particularly inexperienced/incompetent, but the effect was the same. Ahasuerus 11:00, 18 January 2019 (EST)

Art detectives needed

Does anyone recognize any of these covers:

• cover1
• cover2

? Or if you do not recognize the image, any ideas of who the artist may be so I know where to start digging? I managed to match 3 of teh covers in the same series, these 2 are getting on my nerves :) And as I will have a lot of those going forward, if you are interested in helping track down covers, let me know and I will ping you personally when I cannot figure out any of them. Thanks in advance! Annie 17:41, 18 January 2019 (EST)

The second one is the same as the first Avon printing of Zelazny’s Trumps of Doom. I’d link it, but I’m posting from my phone. --Ron ~ Rtrace^Talk 18:28, 18 January 2019 (EST)

I knew I had seen it before! No worries - it needs a note change and all that anyway. All connected. Thanks! Annie 18:55, 18 January 2019 (EST)

Cover1 I found created by Fred Gambino. --Zapp 13:26, 27 January 2019 (EST)

Double novel covers

It came to a discussion on ISFDB:Help_desk. But it should further to be discussed here. Since the books of the series Armchair Fiction Double Novel usually have a second cover artwork on the back, I wanted to know how those ISFDB records could get an unitary and standardized appearance. There is no common sense for this. Are the second cover artworks on the back valid for "Cover2" or "Interiorart"? Do the image titles need to get the particular single novel title or both of them the double title of the book? Looking on the pub titles of this series one can find all variations. So I'm confused. And I got rejected some contributions out of different reasons. --Zapp 13:19, 27 January 2019 (EST)

(copying some parts of the Help Desk discussion:)

I don't think entering cover art as INTERIORART is allowed under the current data entry rules. Help:Screen:NewPub says:

• Add Cover - If there is more than one COVERART record associated with this publication, use this button to create an additional record. This typically happens when dealing with "dos-a-dos" books.
• Interior artwork: Works of art published inside the publication are entered into the "Regular Titles" section of the data entry form and typed as INTERIORART. For more information, see INTERIORART. Note: Cover art credit is entered into its own separate section of the data entry form. (Ahasuerus 19:35, 25 January 2019 (EST))

Except that same help page also says "Artwork on the back cover of a publication is treated as interior art" (under Regular Titles, Title, Artwork). So it conflicts with itself. -- JLaTondre (talk) 20:14, 25 January 2019 (EST)

(after sleeping on it) Let me take a step back and provide more background information.

Back when version "2.0" of the ISFDB software was being designed (2003-2005), the "titles" table was set up to support 3 types of art records: COVERART, INTERIORART and BACKCOVERART. The last one is still defined at the database level -- see Schema:titles -- but it's not supported by the software. I don't know/recall why Al, who was the ISFDB developer at the time, chose not to implement BACKCOVERART. If he mentioned his reasons in 2006-2008, I must have forgotten them.

Since we could have only one COVERART title record per publication between 2006 and 2016, the assumption was that "COVERART" always referred to the publication's front cover. The only way to capture background covers was by using the INTERIORART title type. Once the software was changed to support multiple COVERART records, more options became available. It was no longer immediately clear whether COVERART records referred to front cover art or whether they applied to back cover art as well. Hence the current discussion.

I guess at this point we have the following options:

1. Use the COVERART title type for all works of art that appear on publication covers.
2. Use the INTERIORART title type for works of art that appear on back covers.
3. Implement BACKCOVERART as a separate title type. It would be handled just like the COVERART type: multiple BACKCOVERART titles would be supported per publication and multiple artists would be supported per title.

Although it would be possible to add BACKCOVERART to our menagerie, there are a few things to consider. First, it would be fairly time-consuming to do. Second, we would need to decide how to handle various unusual scenarios. Off the top of my head:

• "Wraparound" covers with the same image used on the front cover, back cover and possibly spine.
• Publications with multiple covers where it's not clear whether the book has two front covers (dos-a-dos) or a front cover and a back cover.

I suspect that these scenarios would make it difficult to decide what is COVERART vs. BACKCOVERART, so adding support for BACKCOVERART may create more problems than it would solve.

P.S. We may want to move this discussion to the Community Portal since it involves possible software changes as well as Rules and Standards changes. Ahasuerus 14:04, 26 January 2019 (EST)

Zapp's question above is specific to Armchair Fiction Double Novels. The Help Desk discussion was a bit more general.

• From a general perspective, I don't see a need for a separate BACKCOVERART type. In addition to the issues you mention, what constitutes a front vs back cover for a dos-a-dos? Simply recording multiple COVERART as we have been doing is sufficient. For any odd cases, use pub notes to document.
• From a Armchair Fiction Double Novels perspective, as a cover case, they are equivalent to dos-a-dos cover. Two separate covers, each illustrating one story, with a unique title and author credit. As such, they should be treated as dos-a-dos. That means: 1) each cover is entered as a separate COVERART record; and 2) each cover is titled to the specific work it's the cover for (the "TITLE / TITLE" of the overall pub is not present in the pubs themselves; they are individually titled).

-- JLaTondre (talk) 14:44, 27 January 2019 (EST)

Except that this will lead to more complications and arguments (see below) and that it was declared to me when I was a beginner here, that the only exception for a second cover entry was the dos-a-dos (back-to-back) binding specific to the Ace doubles. The Armchair Fiction publications don't have that binding; instead, front and back are clearly differentiable. And, as stated in the initial discussion, precedencing this will lead to many, many additionals arguments if art on the back of publication is to be regarded as worthy as cover art: there are numerous cases in which there is only a smaller fraction of the back or a vignette as artwork, accompanied by other information, like blurbs or price information (as with the Armchair pub.s). There's also the difficulty of art on the back of magazines, which should - in my opinion - in no way be regarded as cover art. So, the way it is right now stated in the help (to regard art on the back of publications as interior art), seems the most logical and the most easy path to follow. Stonecreek 02:33, 28 January 2019 (EST)

Double novels (whether bound dos-a-dos or not) are a unique case. They have individual covers for each story with one title and one author on each cover. The only difference from dos-a-dos is the binding. Treating Armchair Fiction Double Novels like we do dos-a-dos is not going to create any more difficulties, exceptions, etc. than dos-a-dos does. They both have the same clear criteria that can be easily written up in the help. That it would somehow open the gates to minor artwork is a red-hearing. -- JLaTondre (talk) 07:34, 28 January 2019 (EST)

No, it clearly is not. This are - in most cases - not double novels; instead these are for the most part ANTHOLOGIES, and many of them that still are indexed as OMNIBUSES have to be reexamined for their contents. The bulk of the texts don't fill the word count of 40,000. So, it is in fact a different case: dos-a-dos does mean the binding is back-to-back. The argument at the time was that there's no way to decide which is the front, and which is the back. That is not the problem we do encounter here. Stonecreek 09:11, 28 January 2019 (EST)

Content is irrelevant. Many dos-a-dos are anthologies & collections. For this question, the only thing that matters is the covers. This and this are significantly (and easily quantifiable) different than this and this. -- JLaTondre (talk) 09:45, 28 January 2019 (EST)

Covers - Definitions

Let me see if we can first agree on the basic definitions. (I will limit the following discussion to paper-bound books and magazines because audio- and e-publications can organize their data in many different ways. Although e-pubs mostly try to emulate traditional physical books, it's due to their readers' current expectations and may change in the future.)

• Paper-bound books and magazines have two physical covers. They are typically made of paper or cardboard with an optional dust jacket.
• There are two types of covers: front covers and back covers. Front covers are adjacent to the point where the text in the book begins and back covers are adjacent to the point where the text ends.
• Most books have one front cover and one back cover. Some books, which we call dos-a-dos, have two front covers because there are two points where text starts.
• Cover art is a single (sometimes composite) work of art which appears on a cover. It can occupy a part of or all of one cover. It can also appear on both covers, which is known as wraparound cover art.
• A single cover may have multiple pieces of cover art appear on it, although it's uncommon.

Are we all on the same page so far? Ahasuerus 12:51, 28 January 2019 (EST)

Yes, we are (well, at least I am :-) )
I recommend breaking this discussion up into several sub-discussions. As discussed, the dos-a-dos and Armchair cases are trivial, so let's agree that we want to update (upgrade?) the rules for these cases first. As pointed out by Stonecreek the original rule pertained to dos-a-dos only. This, however, does not imply we'll have to stick to that original rule indefinitely.
Once that one's settled - and the rules updated/clarified accordingly - we can have a go at the other cases that have come up. Whaddyallthink? MagicUnk 13:09, 28 January 2019 (EST)

Me too (on the same level, I mean). As pointed out, the matter is not one of content, but one of binding (JLaTondre brought up the term of double novels). I could live with a rule of exception for the Armchair titles, but I do still fear of a precedence for other publications that have art of any kind on their back. The most dangerous can of worms would be magazines.

On the whole, it has proven to be best for us to keep the number of exceptions as low as possible, and there really is no need to make one in this case. The Armchair Fiction publictions publish normal anthologies that only take advantage of the previous Ace Doubles, but still are quite different from them. Stonecreek 14:14, 28 January 2019 (EST)

Eh, I'm not a 100% clear on what you're saying here; are you implying that Armchair is covered by the dos-a-dos case, so no rules update required? (other than removing the statement that back cover is to be treated as INTERIORART- see above). But then you go on to say that Armchair is quite different from dos-a-dos; which I interpret as the rules need an update for Armchair after all. MagicUnk 11:52, 29 January 2019 (EST)

(unindent) It sounds like there is no desire to add BACKCOVERART as a separate title type, so we just need to decide when to use COVERART and when to use INTERIORART. Ahasuerus 17:27, 30 January 2019 (EST)

I agree that we don't need an additional type for back covers. I'm fine with ordinary back covers being treated as INTERIORART. I'm also fine with dos-a-dos books having two front covers. I do have a couple of other monkey wrenches to throw into the discussion. I have used a second COVERART title to represent the artwork on a slipcase when it differs from that of the cover. I have also used a second COVERART title to represent artwork from the dust-jacket when it differs from that printed on the cover itself. Is this different than how others understand these? --Ron ~ Rtrace^Talk 18:53, 30 January 2019 (EST)

I would also record a dustjacket and a slipcase as a separate COVERART in these cases with a note on why I have two covers added (and which one is which). Annie 19:06, 30 January 2019 (EST)

Agreed (and there's also the case when multiple earlier artworks are shown on the front of, for example, an OMNIBUS). But, since Armchair Fiction publications are not dos-a-dos, there's no need to make an exception: the back is clearly separable from the front; most pub.s show as a bonus an associated piece of art on their backs, and the publisher chooses wisely which one to show on the front. Considering the art on the back also as cover art only would IMHO lead to endless discussions if a given artwork is to be considered as full-grown cover art or not for other publications. Stonecreek 03:47, 31 January 2019 (EST)

Ah, but here I have to disagree. The Armchair examples given by Zapp are clearly cover art - front and back, so we should record both as cover art in the DB. MagicUnk 07:48, 31 January 2019 (EST)

We could variant the artwork on back even when it is called INTERIORART to the parent cover anyhow. I prefer Stonecreek's point of view since "cover" means usually the front cover not the back. --Zapp 14:04, 31 January 2019 (EST)

No, no, we can't do that. It's not because you -can- do something, that it's also a good idea to do. INTERIORART for COVERART is counterintuitive (not to say its plainly wrong). We have a perfect, easy-to-understand solution, and that is to treat cover art as cover art. What's not to like? [and by the way, I always read my books 'from cover to cover' :)] MagicUnk 14:27, 31 January 2019 (EST)

Sounds like any time there's more than one COVERART, there needs to be a note specifying which cover corresponds to which part of the book (slipcases, jackets, art-in-art, back cover). It addresses the immediate problem and we can monitor the books with more than one to see how it's used. A template to document in the note might make searching the notes easier in case we want to add comments to the publication - coverart relationship in the future. ../Doug H 20:23, 1 February 2019 (EST)

Yes, that to me is a simple solution. One thing though, what would such a template have to convey? Can you come up with a proposal? MagicUnk 14:36, 2 February 2019 (EST)

I'd rather not overthink it. I'd think a word or two three would suffice (e.g. Front, Back, DOS-A, Slipcase). The template would take this value and reiterate it with a wrapper of Cover( ... ). Give it a year or three and we could generate a report to see what people are finding to put in. Standards could be set based on some experience. ../Doug H 20:00, 2 February 2019 (EST)

That's actually a good idea. Could perhaps be a parameterized template or some such. Anyone else? Ahasuerus? Annie? MagicUnk 10:39, 3 February 2019 (EST)

Unfortunately, I am under the weather at the moment and I am not sure I understand the proposal correctly. A new template would be easy enough to add, but what kind of parameters do you have in mind? At the moment Notes templates (unlike Wiki templates) support only one parameter per template. Something like {{MULTICOVER|Front}} or {{MULTICOVER|Back}} is possible, but {{MULTICOVER|Front|12345}} is not. Ahasuerus 14:19, 3 February 2019 (EST)

Could a cover art location template such as {{CALOC | ''phrase''}} be used to reiterate the text phrase, where phrase may be one or more words? For example "Front", "Slip case" and "Art in art". The purpose would be simple to identify references to locations for cover art that could be easily parsed. ../Doug H 14:46, 3 February 2019 (EST)

I think Stonecreek oversimplyfies the present situation stating that 'only Ace doubles' are (to be) entered with two covers. I can think of at least two other publication series (e.g. Tor Doubles) where both artworks are also entered as cover art. In the case of Armchair the back cover artwork is also very much associated with the seond title of the anthology/omnibus/collection.--Dirk P Broer 19:13, 3 February 2019 (EST)

(unindent) I think making it really simple is best. Something like "If there are two covers (such as with dos-a-dos, Ace Doubles, Tor Doubles, and so on), create a second cover art entry by clicking Add Cover to add the second cover." Then for any interior art, something like "Any art appearing between the covers should be entered into the "Regular Titles" section of the data entry form with a Type of INTERIORART." ···日本穣 · 投稿 · Talk to Nihonjoe 23:42, 3 February 2019 (EST)

If my understanding is correct, the main concern with entering all works of art that appear on back covers as COVERART titles is that back cover art is frequently limited to a small picture. I can see how small pictures can be viewed as fundamentally different from regular front cover art.

However, it's been my experience that -- in most cases -- small pictures that appear on back covers are not credited and are rarely entered into the database. Does this match everyone's experience? Ahasuerus 00:24, 4 February 2019 (EST)

Sorry, but no! First, the main concern is not only one for 'minor art' (meaning size), but a general one, and especially extended to magazines:

A smaller sized art is quite often credited or creditable because the art is known.

The magazines are a special can of worms, I'd think, since the back is per definition interior space. And since anthologies / magazines of today are quite often transformed into the other type, we would have to have a third eye on the back of those. Stonecreek 02:29, 4 February 2019 (EST)

Rereading your comments above, I think that the last point may be the root cause of the disagreement. You believe that "the back is per definition interior space". On the other hand, the definitions that I proposed above -- see "Covers -- Definitions" -- suggest that back covers, just like front covers, are exterior rather than interior space. Would you like to expound on why you think that back covers are best classified as interior space? Ahasuerus 10:29, 4 February 2019 (EST)

Well, the rules for magazines state to give the whole page count, regardless of the pagination, so the back is part of the interior per definition. Stonecreek 11:05, 4 February 2019 (EST)

Template:PublicationFields:Pages tells us to count both front covers and back covers:

• For magazines, the rule is to use the actual page count - including the cover. For example, early issues of Fantastic Universe numbered the interior pages from 1 to 192, not counting the front or back covers. This would be entered into the ISFDB record as "196".

If we were to take it to mean that back covers are "interior", then it would also make front covers "interior", which wouldn't make sense. Ahasuerus 11:36, 4 February 2019 (EST)

I'm afraid I don't know the Tor Doubles, but there seems a need for dealing with them anyway: part of the lot is listed as 'pb', the other one as 'dos'. Which one would be correct? Stonecreek 02:29, 4 February 2019 (EST)

Some Tor Doubles were dos-a-dos and some weren't. To quote SFE3:

• Towards the end of their existence as a line, Tor Doubles began to appear with the two titles presented sequentially

I have most of them in my collection, but it would take a fair amount of time to dig them up and analyze their evolution. Ahasuerus 10:14, 4 February 2019 (EST)

And thinking back, the reason for the passage of treating backs of publications as interior space can't have been that we formerly had only one field for cover art, since the entry for the Ace Doubles was of the form 'Title 1 / Title 2' by 'Artist A' and 'Artist B'. The good reason for this really must have been the simplifying of decision. Stonecreek 02:29, 4 February 2019 (EST)

I don't remember all the details, but prior to 2016 the software handled COVERART records with multiple artists inconsistently. Depending on how the data was entered/edited, sometimes the software created a single COVERART title with multiple artists and sometimes it created multiple COVERART titles. (Once the software was upgraded, I created a cleanup report to look for potential problems.) In any event, when entering multi-cover publications like Ace Doubles, we just had to pray that things would come out OK. Ahasuerus 09:29, 4 February 2019 (EST)

While the discutants will (hopefully) be informed about the matter, the non-discutants and all the new editors will be not. This way we will have more discussions and corrections on a matter that really is not ISFDB's first concern. We decided at one time to include artwork, but we still aren't the ISADB: our concern is the speculative fiction, and there's a huge lot to with that! Stonecreek 02:29, 4 February 2019 (EST)

And there's some other problem I have encountered: one of the Armchair Fiction publications had the artwork on the back as cover, while the actual cover artwork was not credited, because it was not creditable (the original artwork was known, but uncredited). We would encounter this scenario on a regular basis, and this would pose another can of worms that has nothing to do with our core business! Stonecreek 11:05, 4 February 2019 (EST)

Sorry, but your last arguments are irrelevant to the discussion imo. How we choose to count pages for magazines has nothing to do with how to treat art that appears on the outside of the magazine (aka the covers), and, it is not because it isn't the core business of the isfdb to record coverart that we shouldn't discuss it further. It is also irrelevant that not everyone participates in this discussion, since if it would, not a single change or improvement (to the rules). Could ever be allowed.

Concerning your last remark, I 'm afraid I don't understand the issue. We do allow 'uncredited' art, don't we? MagicUnk 01:41, 5 February 2019 (EST)

You seem to have not read my remarks: the concern is not that we index artworks, but that it's not our core business to waste time on discussing and reviewing them, and that is what invariably would happen. And no, we don't index uncredited cover art, and not necessarily uncredited interior art (that depends on an editor's preference). Stonecreek 04:16, 5 February 2019 (EST)

I beg to differ. We do index artwork, so, it is our business to discuss - either what we're doing now, or when we have to go and lecture an editor he/she's not following the rules. And I do not share your fear we'll be opening endless discussions. In practice, it'll be glaringly obvious what cover art is and what is not. No need to come up with hypothetical worst-case scenarios; no-one is going to add artist information of 'vignettes' appearing on magazine back covers... Again, the way Nihonjoe phrased it above, makes the most sense to me and is understandable to editors.

Sorry, but as a relative newby you really don't seem to know what's possible, and what the amount of clean-up was and is going to be. Indexing artwork is mandatory: this train has left the station long ago, but altering the definition of cover art also would mean to review ALL items that have art on their back, a task that is of no value to our goal. Stonecreek 13:43, 5 February 2019 (EST)

Can you enlighten us and give examples of what review work you are thinking of? As far as I can see the work that needs to be done anyway is to review ace doubles & armchair pubs and clean up the mess that's already there, cfr. Zapp's examples. That would not be additional work, as it needs to be done irrespective: either remove all information related to the back cover for Armchair pubs that you suggest should not be there, or regularize and make sure that there are two COVERART records (for as far as they are credited) and consistently and uniformly applied as per the rules clarification suggested by Nihonjoe below. Either way, work to be done... MagicUnk 14:06, 5 February 2019 (EST)

Quite easy: each piece has to be reviewed if it would be to considered as a worthy 'cover art' or not. Or are you suggesting just a 'lex Armchair Fiction'? Stonecreek 14:22, 5 February 2019 (EST)

Eh, sorry, I have to admit I don't understand what you mean exactly? The vast majority of pubs in the DB are standard novels and the like and don't have problems at all. We're talking about specific cases such as, indeed, the Armchair fiction series, no? Admittedly, reviewing and updating with both cover art records (when available) will take some time but surely not unsurmountable? And doesn't have to be done immediately, right?

And can we actually get an idea of how much records we are actually talking about? Surely less than the current cleanup round going on to get the Bleilers and Reginalds in their won ExternalID field I'd guess? MagicUnk 15:56, 5 February 2019 (EST)

As experience has shown many times it's best to keep the number of exceptions as small as possible. The problems we'd encounter with them have been adressed above, and any unnecessary exception will invariably lead to questions of the kind: "when it's here allowed, why not there?". Stonecreek 02:41, 6 February 2019 (EST)

You're misrepresenting the proposal, I'm afraid. It is actually intended to eliminate exceptions, not create more of them. MagicUnk 06:46, 6 February 2019 (EST)

Concerning uncredited art; that's a valid point you're making - looks like there's an undefined rule, leaving choice to the editor in question whether to add uncredited cover art or not... MagicUnk 13:23, 5 February 2019 (EST)

I thought of a modification, just to make things clear: "If there are two covers (such as with dos-a-dos, Ace Doubles, Tor Doubles, and so on), a second cover art entry can be created by clicking Add Cover to add the second cover. If the back cover is an advertisement (which often happens with magazines, and occasionally with older books), that is not considered cover art and should not be documented." Thoughts? ···日本穣 · 投稿 · Talk to Nihonjoe 13:38, 5 February 2019 (EST)

I'm all for it. Makes perfect sense to me. MagicUnk 14:06, 5 February 2019 (EST)

Your proposal adresses none of the problems that are already there and the ones we're going to encounter. Please make some responsible suggestions how to deal with those! Stonecreek 02:36, 6 February 2019 (EST)

What Nihonjoe proposes is actually to solve the problems that are already there by making it clear what to do when there are two covers with cover art on them. Clearly the mess we're seeing now is caused by inadequacy of the current rules, doesn't it? To be honest, I have yet to read convincing arguments, preferrably supported by factual evidence, that when we update the rules to allow front- and back coverart to be treated equally we inevitably will run into numerous problems. They just aren't there. The magazine case is moot, as it is clear that the 'vignettes'/ advertisements aren't going to be considered as cover art.

Nothing - under the circumstances that there are dos-a-dos publications - can be more easy than the current rule of art-on-the-front-of-a-publication --> cover art; art anywhere else in the publication ---> interior art. If you still think there are no problems coming up then you have to reread the paragraphs above.

Also, something else to keep in mind is that the majority of editors and moderators that have chimed in on this discussion are expressing preference to varying degrees (or are impartial) for updating the rules to unequivocally allow front-and back coverart records both entered as COVERART (and not INTERIORART). After all, why else did Ahasuerus update the software in 2016 to accept multiple COVERART records if not to allow exactly that? MagicUnk 06:46, 6 February 2019 (EST)

Can't you read? This step was to ensure that different pieces of cover art on the front can have their respective own entries. There is no other solution for this problem as easy and elegant as the existing one! Allowing your proposal will lead to unnecessary problems that are already solved (at least in principle) and endless fiddling. The artwork is represented at ISFBD and that's all that is of business to us! Stonecreek 09:39, 6 February 2019 (EST)

I quote: between 2006 and 2016, the assumption was that "COVERART" always referred to the publication's front cover.. 1) it is an assumption, and 2) this definition was only valid between 2006 and 2016. MagicUnk 11:02, 6 February 2019 (EST)

Yes, and the assumption is still correct in 2019, so what? It says something that you don't try to argue with the simplicity of the current rule: in the other cases one would have to define what a genuine cover is made of or what it has to fulfill; what's to do when there's creditable art only on the back? (it would be displayed - totally misleadingly - as the one and only cover art), and so on, and so on. You'd have to think of all dangers and try to figure a way out of them, which would all be futile work. Stonecreek 11:18, 6 February 2019 (EST)

Well, that's where we disagree, obviously. As to your case how do address if there's only backcover art (does that actually exist?); the solution has been suggested somewhere in the text above: use the notes to disambiguate, as we're doing everywhere else we need additional clarification.

Now, all that said, and as we're the only two that keep discussing, I'll propose to put the discussion to an end (for now) since it's obvious we'll not be reaching agreement anytime soon... MagicUnk 11:32, 6 February 2019 (EST)

(unident) Sure it does exist, and shuffling notes around would change nothing for the obvious gap between the displayed cover art (the one that's not creditable) and the text right near it, which would be for a totally different piece. And there are more cases that eventually will pop up: mutltiple artworks on the cover and/or multiple artworks on the back came to mind without stretching my mind. The thing is that - although some things seem to be quite easy & intuitively right - they get rather complicated to define and difficult to maintain in the complex ISFDB environment. Stonecreek 13:43, 6 February 2019 (EST)

Summarizing the Arguments for and against

I have been trying to wrap my brain around all the arguments and counter-arguments, but I don't seem to be making much progress. If I am feeling OK tomorrow, I will try to query the database. It may help us determine what types of cases we are dealing with and what the scope of the problem is. Things like the number of INTERIORART titles whose page number is "bc" and so on. Ahasuerus 00:06, 7 February 2019 (EST)

I'll try to summarize (I'm sure Stoncreek will chime in if I've been incomplete). The question is: do we record cover art that appears on the front and/or back of a publication of any type? Is it allowed per the current rules to do so or not? If it is, how?

Arguments against, voiced by Stonecreek:

• It is not our core business to index backcoverart, except for dos-a-dos
• Additional cover art records are currently created as INTERIORART, let's stick to current practice or we'll have to update a whole bunch of pub records
• If multiple coverart records would be allowed, it would lead to all kinds of questions & discussions. For example:
  • A single cover art picture is uploaded, either showing only one cover or two, and two records exist for front and back; which record refers to which part of the picture? You don't know unless you add notes to clarify.
  • A cover art picture is uploaded showing front-and backcover, and only artist for frontcover is credited (the other one is unknown), so only one coverart record; again to what part of the picture does that record refer to?
  • A back cover (of a magazine) shows small art that is recognizable, the so called 'vignettes', who's to prevent editors from entering a coverart record for each 'vignette'?
  • Allowing back cover art in general would require a massive update of currently existing records, eg the ones that have backcoverart as interiorart --> This is where the question of how many records are we talking about is coming from.

Arguments for, voiced by MagicUnk and other editors:

• The rules allow it. Nowhere in the rules it says that entering front-and back coverart as COVERART records is forbidden, there is no restriction on pub types mentioned (that I could find), dos-a-dos is mere mentioned as an example, not an exception, where it is relevant to have two records - but there are serveral other cases - eg see Armchair above.
• Allowing multiple coverart records per pub makes the rules simple and easy to understand (actually the rules already explicitly allow for this - we only have to make them clearer and more explicit that the multi-coverart case is also covered
• It is very unlikely 'vignettes' will be considered as coverart, and moderators will catch them quickly; and we can add clarification that it this kind of art is not to be entered)
• Disambiguation has to be done in the notes irrespective - cfr. cases against. Either you add notes to describe front-and back cover, or you use multiple coverart records and notes as needed.
• There is no definition that restricts COVERART to mean front cover art only

Applicable rules

These are the ones I've found, with relevant excerpts

• Cover Art
  • If the cover art is credited to more than one artist, use this button to create a second artist field.
  • If there is more than one COVERART record associated with this publication, use this button to create an additional record. This typically happens when dealing with "dos-a-dos" books.
• Contents always included: interior artwork
  • Note: Cover art credit is entered into its own separate section of the data entry form. If there is a significant reproduction of the cover art inside the publication, it can also be entered as a separate INTERIORART record. That you COULD interpret as "enter backcoverart as interiorart" - but that's really stretching the rules imo., and it really means that for a significant reproduction of the cover you enter an additional INTERIORART record (next to the coverart one).
• INTERIORART
  • This section has no explicit rules about cover art, except perhaps (but I doubt it) for Rules for including artwork. If artwork illustrates a particular story, it should be included. If it does not, but is a significant piece of artwork, or is signed by or credited to a well known sf artist, then it should be included.

From the preceding discussion it has become clear that further clarification of the rules re. COVERART definition and data entry is in order since there's clearly two different interpretations given to current ruleset. And BTW, I've reread the rules & current discussion carefully, but of course I could have missed something. Let me know if I have. MagicUnk 06:44, 7 February 2019 (EST)

WRONG! The help pages clearly state that it is forbidden to enter art on the back as coverart: I cite what JLaTondre has found and posted somewhere at the beginning of this thread: Except that same help page also says "Artwork on the back cover of a publication is treated as interior art" (under Regular Titles, Title, Artwork).

Don't shout please. That's rude.

Sorry, I didn't want to be rude; it was just that I couldn't believe you had overlooked that prominently featured rule. Stonecreek 11:58, 8 February 2019 (EST)

Somehow I missed that piece of text that you pointed out. It must be removed as it is inconsistent with all the other helptext. MagicUnk 15:14, 7 February 2019 (EST)

We are now in a situation where it is quite easy to deal with the title types COVERART & INTERIORART, and this is precisely because of this rule. It seems most likely that it was written down because someone thought of the difficulties we would get into otherwise. We have encountered publications where the COVERART is not creditable. Allowing the art on the back also as COVERART will lead to the absurd situation that the title of the then 'cover artwork' has nothing to do with the displayed cover image (including the artist). Also we will encounter the situation where there are multiple artworks displayed on the front and/or the back, which will only lead to more confusion. Stonecreek 09:37, 7 February 2019 (EST)

I propose a couple things:

1. Change the rule mentioned by Stonecreek so that coverart is never treated as interior art. It's counterintuitive and will only continue to confuse people in the future.
2. Add some way to mark a COVERART entry as the back cover. This could be a checkbox next to it that makes it display as "Back Cover Art", or adding a BACKCOVERART type that does the same thing.

This will alleviate the "absurd situation" Stonecreek mentions. As for the situation where multiple artworks are used on a front cover, those are fairly rare, and can already be handled by adding additional artists for whichever cover. So, if you have a front cover that has art by both Michael Whelan and Bob Eggleton, you simply have both artists listed on the single entry since we treat each cover as a single work. ···日本穣 · 投稿 · Talk to Nihonjoe 14:07, 7 February 2019 (EST)

Hmmm. Makes me think a checkbox wouldn't be needed as disambiguation by appending either '(front)' or '(back)' would work equally well. MagicUnk 15:28, 7 February 2019 (EST)

We'd only need "(back)" because the default assumption is that it's the front cover. ···日本穣 · 投稿 · Talk to Nihonjoe 16:19, 7 February 2019 (EST)

(unindent) So how about this:

• For COVERART: "If there are two covers (such as with dos-a-dos, Ace Doubles, Tor Doubles, works with back cover art by a different artist than the front cover, and so on), a second cover art entry should be created by clicking Add Cover to add the second cover. If the back cover art is a wraparound work (meaning the front cover picture wraps around to the back cover), do not create a separate back cover entry. If a front or back cover has art from multiple artists or the art was collaboratively created by more than one artist, click Add Artist to add an additional artist for the cover. If needed, add a note to the Pub. Note field to clarify which entry is the back cover. If the back cover is an advertisement (which often happens with magazines, and occasionally with older books), that is not considered cover art and should not be documented."
• For INTERIORART: "Any art appearing between the covers should be entered into the "Regular Titles" section of the data entry form with a Type of INTERIORART."

I tried to address the concerns expressed by Stonecreek. Hopefully, I did so. Thoughts? ···日本穣 · 投稿 · Talk to Nihonjoe 17:18, 7 February 2019 (EST)

With the addition of not allowing uncredited COVERART, it nicely summarizes the proposal. Thanks. MagicUnk 01:18, 8 February 2019 (EST)

I'd think it is really vice versa: COVERART does mean the art on the front: it is the one displayed in bookstores of any kind (and it's also here at ISFBD in very most cases); the art on the back always is a gimmick on the plus side, just the way it is with INTERIORART.

And I'm afraid that your proposal does in no way deal with the problem of 'minor' art (small size, art somewhat hidden under the cover blurb or price/EAN information (like here, etc.). Stonecreek 04:09, 8 February 2019 (EST)

Why would it be a problem to have two COVERART records that refer to the same art? In these cases we simply variant the back to the front. Or we explicitly add "...do not create a COVERART record for the back cover if it is a variant of the front cover art" if we really don't want these cases recorded. MagicUnk 06:46, 8 February 2019 (EST)

That wasn't an example for a doubled piece of coverart (though that would be pretty useless), it should emphasize an example for 'minor' art.

I think I understand. Let's review what we mean by COVERART. I believe that when you refer to 'minor' art appearing on the cover you focus on the individual pieces of art that together make up the cover art, right?

Whereas you can also look at COVERART from the whole, and not the individual parts. Now, for the sake of argument, we can record a composite cover (ie a cover composed of individual pieces) in two ways:

1) have a single art record with multiple artists, ie referring to the whole and listing contributing artists, or

2) have multiple art records with a single artist, ie each record referring to the individual parts.

Do you refer to the second case when you point out the issues with multiple art or other types of composite art covers? MagicUnk 12:26, 8 February 2019 (EST)

The proposal also still doesn't explicitly distinguish the publications the rule is thought for (all with art on the back, also magazines, even pamphlets?) Stonecreek 11:58, 8 February 2019 (EST)

And the rule Artwork on the back cover of a publication is treated as interior art is not 'inconsistent with all the other helptext' as you stated above. What gave you this idea? Stonecreek 00:01, 9 February 2019 (EST)

Well, Cover Art says it is allowed to enter multiple coverart records (and merely refers to dos-a-dos as an example). It is clear that there are other cases where multiple cover art records are relevant, including the case that started all this, namely Armchair. So, back cover art as we're currently discussing is imo included and can be treated as a second coverart record as per this rule. Then, indeed, the rules go on to say that backcoverart is to be treated as interiorart. That, to me, seems an inconsistency in the rules. (PS. could you provide an answer to my question above concerning 'minor' art whether I've understood what you're saying?) MagicUnk 08:45, 9 February 2019 (EST)

No! The rule just clarifies that art on the back of a publication is treated as interior art (like with Armchair Doubles), but when you have two fronts (like with Ace Doubles or in the case of dustjackets & corpus of a publication both bearing art) then it is allowed to add a second COVERART title). That's really the most easy & elegant solution. I really'd like to see another relevant site that treats or announces the art on the back of a publication as COVERART. (Even the rules you found mention This typically happens when dealing with "dos-a-dos" books., because it is the most common & relevant case.) Stonecreek 23:55, 9 February 2019 (EST)

I'm beginning to suspect that we are more or less saying the same (but not quite). Let's first discuss definitions - see below. MagicUnk 06:46, 10 February 2019 (EST)

What Is Currently in the Database

Thanks for trying to summarize the arguments! I have run a few queries and here is the current breakdown of titles that appear on the back cover of a publication and are entered as "bc":

• 1170 INTERIORART
• 79 ESSAY
• 45 POEM
• 23 SHORTFICTION

There are no COVERART titles. It doesn't address the issue of Help inconsistencies, but at least now we know what the current scope of the problem is. Ahasuerus 14:10, 7 February 2019 (EST)

There will be more records that'll need attention though, such as Armchair records that could be revisited to improve data quality, which can be done gradually over time. Note the 'could' as I don't want to imply a 'must'.

What must be done though is to remove the helptext inconsistencies, either way. MagicUnk 15:14, 7 February 2019 (EST)

An Alternative Software-Based Approach

Now that I have read all the arguments and reviewed the data, it occurs to me that there may a different way to approach this issue. Let's take a step back and consider why the New/Edit/Add/Clone Publication pages have a separate "Cover Art" section. The reason is that we capture different data elements when creating COVERART titles vs. regular types of titles, including INTERIORART titles.

First, the following fields -- "Title", "Date", "Title Type" and "Length" -- are not displayed when creating a NEW publication or editing a publication without a COVERART title. A new COVERART title record automagically uses the title and the date of the publication. It's just a data entry shortcut because the resulting COVERART records end up with the same fields as INTERIORART title records: Title, Artist(s), Date and Title Type.

The second difference, however, is more fundamental. The "Cover Artist" section doesn't have a "Page" field because the original assumption was that COVERART titles would always appear on the front cover and because we didn't consider other scenarios like separate dust jacket or slipcase art. The result is that, because we don't capture page numbers for COVERART titles, there is no easy to way to distinguish between the art that appears on:

• front cover
• dust jacket
• slipcase
• back cover
• wraparound cover

Hence the need to use Notes and the various proposals above to add support for "back cover" checkmarks, a new BACKCOVERART title type, etc.

On the other hand, INTERIORART is a regular title type which is entered in the "Regular Title" section, so it can have page numbers associated with it. That's what makes INTERIORART so attractive when deciding how to enter art records -- it lets you easily capture the art work's location, including "bc" for "back cover".

Now then. If we were to add a "Page" field to the "Cover Art" section, it would address a whole bunch of issues. If left empty, there would be no changes to the way the "Cover" line appears on Publication pages. For dos-a-dos publications, we could add a code like "afc" which would then be displayed as "alternate front cover" on the "Cover" line. Ditto "dj" for "dust jacket", "slip" for "slipcase", "wrap" for "wraparound cover" and "bc" for "back cover". It would take some work, but nothing insurmountable.

So, what do you think? Ahasuerus 17:31, 7 February 2019 (EST)

I have an objection to treating back cover illustrations as COVERART. The current rules disallow COVERART when the artist cannot be identified. Changing this has been discussed before and the consensus was to not enter uncredited COVERART. I don't know how many of the 1170 INTERIORART bc records are uncredited, but by our rules, this conversion would necessitate their deletion. This could include other artwork that appears on the back covers of magazines that is reflected with the actual page number rather than "bc". Aside from the expansion of COVERART to include back covers, I support this software approach. --Ron ~ Rtrace^Talk 19:27, 7 February 2019 (EST)

I guess you mean to say you don't support back cover illustrations as uncredited INTERIORART? But you do support back COVERART if they are credited? If so, that can/should be added to Nihonjoe's proposed wording, as he mentions below. MagicUnk 01:11, 8 February 2019 (EST)

No, uncredited INTERIORART which occurs on the back cover, is legitimate today. I don't want to decrease the scope of what can be entered and I especially don't want to delete data that has already been entered. --Ron ~ Rtrace^Talk 02:31, 8 February 2019 (EST)

Right... why wouldn't we remove obsolete and/or inconsistent data? On one hand we argue that we don't want to record coverart that's uncredited. But then we go on to say that it's fine to record coverart that's uncredited as long as its an interiorart record?

Because, the illustrations on back covers which are entered as INTERIORART are not currently obsolete nor inconsistent. They would be made so only As a side affect of this proposal, wiping out decades of data entry. Is there an argument that these titles that have been legitimately added now should be prohibited based solely on location? --Ron ~ Rtrace^Talk 07:42, 8 February 2019 (EST)

Now, there's another solution to the coverart-as-interiorart-credited-to-uncredited, and that is to allow uncredited coverart records and then move the interiorart uncredited ones to their proper place in the DB. MagicUnk 07:00, 8 February 2019 (EST)

I would support allowing uncredited COVERART to be included. However, it has been litigated before and the consensus was not to do so.--Ron ~ Rtrace^Talk 07:42, 8 February 2019 (EST)

Well, situations change and I wouldn't mind reverting that earlier decision and start including uncredited coverart. It then of course begs the question what were the criteria of the editors to include or not include uncredited coverart as INTERIORART? We could add all uncredited coverart, but I don't think that'd be a good idea, so what were/are the particular criteria editors use(d) to decide whether to do so or not? Now, for me, what's important is that we can come up with a rock-solid rule that is simple, elegant and has as few exceptions as possible (preferably none at all). MagicUnk 08:36, 8 February 2019 (EST)

I don't support adding back cover art (or front cover art) if we can't identify the artist, and none of my suggestions included that. We don't generally enter INTERIORART if we don't know who the artist is, either. I do, however, find it confusing and unintuitive to enter art found on the cover as INTERIORART. Perhaps if we changed INTERIORART to ILLUSTRATION instead? ···日本穣 · 投稿 · Talk to Nihonjoe 20:37, 7 February 2019 (EST)

We have 11,833 INTERIORART titles with the author as uncredited. It's always been my understanding that this is permissible. --Ron ~ Rtrace^Talk 21:00, 7 February 2019 (EST)

Sure. It is permissible because it conveys there is contents on a particular page that is art, not text. I would not generally enter uncredited interiorart myself, but I would add a clarifying note. MagicUnk 01:31, 8 February 2019 (EST)

Regarding what to enter for the type of INTERIORART, perhaps add a drop-down (like the one for story length) with the options you mention:

• front cover
• dust jacket
• slipcase
• back cover
• wraparound cover

That way, a new editor (or one just unfamiliar with thing) wouldn't have to try to figure out what abbreviation to use. ···日本穣 · 投稿 · Talk to Nihonjoe 20:40, 7 February 2019 (EST)

I propose to first determine what we want and only then to discuss best way to implement. That said, treating COVERART data entry mechanisms much the same way as INTERIORART does has its merits, provided they do not stánd in the way of rules implementation as suggested by Nihonjoe (including the rule to not allow uncredited coverart). By the way, not allowing uncredited COVERART makes perfect sense to me since it wouldn't convey any additional useful iinformation in my opinion. MagicUnk 01:11, 8 February 2019 (EST)

One question regarding the idea of not allowing uncredited COVERART - would this mean I couldn't upload a cover image if I didn't know who had produced it? ../Doug H 11:03, 9 February 2019 (EST)

I don't think so. COVERART records are titles. Cover scan URLs are fields in publication records. Multiple editions of the same book can share the same COVERART title (i.e. art) yet have slightly different cover scans due to different prices, blurbs and so on. At this time we can and do upload cover scans regardless of whether we know who the artist is. Ahasuerus 14:33, 9 February 2019 (EST)

Requirements?

Apologies for starting another section, but I couldn't figure out where to put some random thoughts in the various sub-threads above. --MartyD 09:00, 9 February 2019 (EST)

I don't have any strong feelings about this, but I think we're missing a key consideration, which in my software product development world would be "user requirements". The arguments above seem to be focusing on how best to describe the publication and its full complement of artwork but do not seem to be considering how that information will be used. If the ISFDB is meant as a public bibliographic resource, then we should identify how the non-ISFDB bibliographic world thinks about publication artwork. For example, to my layman's eye (I am definitely not a bibliographic expert), it seems to me there is a somewhat universal concept of "cover" and "cover artist" that, when used in non-ISFDB places, is restricted in scope to the front/"main" cover (which might be the dust jacket). So to support non-ISFDB bibliographers using the ISFDB as a resource, I see these requirements:

• Provide a way to identify the artist(s) for the front cover
• Provide a way to find front covers (or publications having front covers) by an artist
• Distinguish an artist's front cover works from any other works by the artist
• Identify, and distinguish among, multiple possibilities for "front" cover (e.g., dos-a-dos, dust jacket vs. physical front cover beneath the jacket)

Perhaps there are some more. (One thing that occurs to me to ask: Is there any standard/common bibliographic need for doing anything in the list above with a "back cover" context? I sort of doubt it, but I am definitely not an expert.) So how do we do that, and how do we do it in a way that is most natural for ISFDB users to work with?

I think we're a little hung up on the literal meanings of "COVER" in COVERart and "INTERIOR" in INTERIORart. COVER might mean front/main or might mean the entire physical wrapping (front and back). INTERIOR sure sounds like it means "inside", which in the vernacular would be "between the covers", in which case it is counter-intuitive to apply that term to artwork on the exterior. I.e., if our terms were FRONTCOVERART and OTHERART, would we be having such an extensive debate? Yes, there are still some tricky nuances, but perhaps by considering each situation in terms of: "Ok, it's not what an ordinary person would think of as the front cover, so how should we document what it is and where it is?" we might be able to come up with something, and we could probably have some clear treatments for, and explanations of, the few ambiguous situations that remain. --MartyD 09:00, 9 February 2019 (EST)

You're making a very good point there. Along the way I have come to realize that we may not be talking about the same thing - indeed, what's the definition of COVERART? And what definition is the ISFDB using?

Is it art located on

1. the front cover, and only the front cover? If it is, we then need to answer questions like: 'Can a pub have two front covers, and consequently no back cover'? (see dos-a-dos (two front?) vs. Armchair (front - and back?). What with dust jacket? etc etc., or,
2. the exterior of the pub, irrespective of whether it's the front, back, slipcase, etc...

And does it mean

1. to reference the whole only, ie meaning a composite cover treated as a single piece of art that can have multiple artists credited? or,
2. we treat the individual pieces as individually distinct pieces of cover art each with their own artist credit?

Defining cover art as meaning front cover only seems to be too narrow to me, and consequently requires explanation of exceptions such as dos-a-dos, dust jacket,... Defining cover art as meaning appearing on the front and/or back, and treated as a single piece (even when composed of multiple art) makes the most sense to me - and I suspect that that is also what our customers would expect as well. It would still require clarification re dos-a-dos, dust jacket... Of course, if we take these as requirements (or another set), then the detailed specifications and subsequent implementation still needs to be fleshed out... MagicUnk 11:23, 9 February 2019 (EST)

An interesting angle. Another way to put it is that we currently segregate art titles into two categories:

• "Privileged" art titles which appear at the top of the Publication page. At this time they are called COVERART titles and are displayed on the "Cover" line.
• All other art titles which appear in the Contents section of the page. At this time we call them INTERIORART titles and they are displayed slightly differently compared to non-art titles.

So I guess the first question to answer is whether the rest of the bibliographic world draws the line between different types of art pieces in the (substantially) same way even though the terminology may be somewhat different. Ahasuerus 14:26, 9 February 2019 (EST)

Definitions - Redux

• Cover: Definition from Wikipedia: A book cover is any protective covering used to bind together the pages of a book. Beyond the familiar distinction between hardcovers and paperbacks, there are further alternatives and additions, such as dust jackets, ring-binding, and older forms such as the nineteenth-century "paper-boards" and the traditional types of hand-binding. The term "Bookcover" is often used for a book cover image in library management software.
• Cover art: From Wikipedia: Cover art is a type of artwork presented as an illustration or photograph on the outside of a published product such as a book (often on a dust jacket), magazine, newspaper (tabloid), comic book, video game (box art), DVD, CD, videotape, or music album (album art).
• Front cover: No definition found online, but workable definition can be: The cover that is closest to the start of the text when you begin reading a publication.
• Back cover: The cover closest to the end of the text when reading a publication

From these definitions it is inferred that

• A publication can have multiple covers at the same time (front, back, dustjacket,...).
• Any art on the covers of a publication is cover art.
• A publication can have two front covers, and no back cover - a workable definition to capture dos-a-dos type books could be if you have to flip a pub upside down to be able to read all of its contents, that pub has two front covers (might not be all that accurate, but you get the gist, I hope :))

These are the definitions that our customers most likely will have in their mind when looking for specific information related to cover art. So, can we come up with user requirements and rules that are in line with the definitions as listed above? MagicUnk 07:42, 10 February 2019 (EST)

That there's no differentiation to be found can be interpreted either way, it's just as reasonable that by 'cover' the term 'front cover' is meant implicitly. Our closest associates for bibliographic purposes would be SFE3 (and its offspring, the Fantasy Encyclopedia) and, also, Wikipedia. The latter only seldomly credits or depicts covers, the first regularly names artists for first editions, but when they do, all of them speak only of front covers. Also comic dbs name the artist of the art on the front cover. So, where are the examples of major bibliographies that would do otherwise? Stonecreek 10:13, 10 February 2019 (EST)

"Percent of Publications by Format by Year" updated

The statistical report "Percent of Publications by Format by Year" has been updated. It now displays two graphs, one for magazines/fanzines and one for books. The new layout will become available tomorrow morning. Ahasuerus 17:25, 30 January 2019 (EST)

I don't see Green-hc (have only red, orange, black, yellow)? I'm using chrome. MagicUnk 07:51, 31 January 2019 (EST)

And I don't see Blue-digest on the magazine graph either (an issue with first graph not displayed?) MagicUnk 07:53, 31 January 2019 (EST)

Could you please do a full page reload (Control-F5 in most browsers)? I added new CSS colors and it's possible that your browser is still using the previous version of the CSS file. Ahasuerus 13:58, 31 January 2019 (EST)

Unfortunately, ctrl-F5 didn't do the trick. Andeh, it's not only Chrome, but IE11 and Edge as well. MagicUnk 14:24, 31 January 2019 (EST)

Logout, look at it while not logged in and login again (or clean your browser data I guess - the out/in was faster) - that's what got the green to show up for me. No amount of Ctrl+F5 did anything :) Annie 14:32, 31 January 2019 (EST)

Thanks for experimenting and sorry about the aggravation. Unfortunately, I find it hard to test color-coded graphs because I am colorblind. Perhaps I should use different varieties of dotted lines in addition to colors to make life easier for me and the colorblind part of our user base. Ahasuerus 15:39, 31 January 2019 (EST)

The pink and yellow is hard to see against the white background. I don't see the blue in digest/magazines either. Did a reload in both Safari and Firefox.--Rkihara 16:40, 31 January 2019 (EST)

The Blue in magazines is indeed missing but we have two greens - one across the whole period (look at the spike in the 20s - webzines cannot even exist that early) and a second one starting in the late 90s and staying low there (they actually even cross a little after 2000). I think that the big green in magazines should be blue. Annie 16:48, 31 January 2019 (EST)

Thanks, I'll take a look. (It's really hard to test this stuff since I can't tell them apart :-( ) Ahasuerus 18:46, 31 January 2019 (EST)

(unindent) I believe I have found and zapped the bug. Could you please do a full page reload and let me know if it looks better? Ahasuerus 19:08, 31 January 2019 (EST)

And we have blue :) Looks good now. Annie 19:23, 31 January 2019 (EST)

Great, thanks! Ahasuerus 19:24, 31 January 2019 (EST)

Advanced Search revisited

With the latest tranche of Fixer-identified ISBNs out of the way, I am back to working on Advanced Search.

The first patch was installed a few minutes ago. It made a number of internal changes which should't be visible to the naked eye. If you encounter any problems, please post here. Ahasuerus 18:13, 1 February 2019 (EST)

Part 2 has been installed. The changes include:

• Each search types has a separate Web page now
• Asterisks are now fully supported as wild cards (except User search which requires an exact match)
• User name searches are now case-insensitive; the downside is that they take longer because of the way the Wiki software (which we have little control over) works

It was a fairly big patch. Depending on your browser, you may need to do a full page reload (Control-F5 in most browsers) for the drop-down lists to work correctly. If you run into any issues, please post here. Ahasuerus 20:12, 2 February 2019 (EST)

Thanks very much for that load of work! It took only a short time to adapt, but now it appears to me much more user-friendly! Stonecreek 12:02, 4 February 2019 (EST)

Excellent! :) Ahasuerus 12:24, 4 February 2019 (EST)

Awesome! ···日本穣 · 投稿 · Talk to Nihonjoe 16:44, 4 February 2019 (EST)

New Web page - Numeric External Identifier Ranges

As we discussed a couple of months ago, it would be nice to be able to see the gaps in our coverage of the Bleiler/Reginald IDs. I have created a new Web page, Numeric External Identifier Ranges, which display tables with currently entered ranges for the 3 supported Bleilers and the 2 Reginalds. A link has been added to the bottom of the Cleanup Reports menu.

I am not sure the current layout is optimal. Perhaps it would be better to display one External ID Type/secondary verification source per Web page. Let's give it a day or two and see what we all think. Ahasuerus 20:31, 4 February 2019 (EST)

Thanks for this. I've already discovered and corrected several errors. --Ron ~ Rtrace^Talk 21:15, 4 February 2019 (EST)

I actually like it as it is - it is not too much data for a single page. Most of the others won't have ranges anyway so we won't be adding more of those pages all the time. :) Annie 13:03, 5 February 2019 (EST)

Well, if it works, it works :-)

In the meantime I have improved the display a bit and made the data retrieval logic ignore alphanumeric IDs. The latter may make Reginald-3 ranges a bit more manageable. Even so, I suspect that Reginald-3 IDs will be difficult to represent as numeric ranges because of the way records are numbered. I guess time will tell. Ahasuerus 18:24, 7 February 2019 (EST)

Forgotten Worlds (Scotland)

I have a book in hand which has stories that are credited as being first published in "Forgotten Worlds (Scotland)". Issues 3 and 8. Anyone ever heard of it? We have nothing and Google searches find the Japanese game and the Forgotten Realm books. ../Doug H 17:41, 8 February 2019 (EST)

Might be this. Seems to be defunct. ···日本穣 · 投稿 · Talk to Nihonjoe 18:36, 8 February 2019 (EST)

In the meantime, go ahead and enter the stories. We can merge them later if we find info on the magazine. ···日本穣 · 投稿 · Talk to Nihonjoe 18:55, 8 February 2019 (EST)

Galactic Central appears to have a skeleton record for it. It's not in FictionMags nor Miller/Contento. --Ron ~ Rtrace^Talk 19:26, 8 February 2019 (EST)

I found some info and added these issues. I'm also checking with a few friends in the UK fandom scene who may know about it. ···日本穣 · 投稿 · Talk to Nihonjoe 19:39, 8 February 2019 (EST)

The issues and stories I was looking for were included in your assistance. Now I can link them up and get rid of the book. ../Doug H 19:48, 8 February 2019 (EST)

Okay, I found information on all nine issues and added it. ···日本穣 · 投稿 · Talk to Nihonjoe 20:22, 8 February 2019 (EST)

If there is someone in the UK who can visit the British Library, they have all nine issues in their collection. They are in "General Reference Collection ; ZK.9.a.10066". We are still need verification of the months in which they were published (I extrapolated based on information from one issue) and who edited each volume. ···日本穣 · 投稿 · Talk to Nihonjoe 15:28, 10 February 2019 (EST)

Multi-Series and Multi-PubSeries templates

Back in October we added two linking templates: "S" for regular/title series and "PubS" for publication series. We also discussed adding two more templates, one for multi-series titles ("MultiS") and one for multi-pubseries publications ("MultiPubS".) The discussion petered out because we couldn't decide what kinds of parameters the proposed templates should take.

Perhaps a tentative solution would involve the creation of 2 "Multi" templates without parameters. They would simply expand to something like:

• MultiS: "Note that this title belongs to multiple series. Because of software limitations, only one of them currently appears in the Series field."
• MultiPubS: "Note that this publication belongs to multiple publication series. Because of software limitations, only one of them currently appears in the Publication Series field."

If and when we add support for multiple series/pub. series, we should be able to search Notes for these templates and update the data. In the meantime we could use these templates in conjunction with "S" and "PubS", which support linking. Something like:

• {{MultiS}} The other series is {{S|A Modern Witch}} (volume 4).

which will expand to:

• Note that this title belongs to multiple series. Because of software limitations, only one of them currently appears in the Series field. The other series is A Modern Witch (volume 4).

What do you think? Ahasuerus 14:06, 9 February 2019 (EST)

Seems like a good way to handle them until the functionality is added to the database. ···日本穣 · 投稿 · Talk to Nihonjoe 12:16, 11 February 2019 (EST)

FR 1262 has been created and the functionality has been implemented -- see Servant of the Shard for an example. Ahasuerus 13:11, 16 February 2019 (EST)

New Advanced Search options

ISFDB Advanced Search has been updated to include ISFDB Advanced Publisher Search. The new functionality is similar to what is supported by the other record types.

Also, please note that the URL of the main Advanced Search page has been changed from http://www.isfdb.org/cgi-bin/search.cgi to http://www.isfdb.org/cgi-bin/adv_search_menu.cgi . If you have it bookmarked, you will want to re-point the bookmark to the new URL. Ahasuerus 15:49, 9 February 2019 (EST)

Advanced Publication Series Search has been added to the menagerie. Ahasuerus 17:13, 9 February 2019 (EST)

Advanced Series Search has been added. Ahasuerus 19:31, 9 February 2019 (EST)

Advanced Award Type Search has been added. Ahasuerus 19:51, 10 February 2019 (EST)

Advanced Award Category Search has been added. Ahasuerus 12:51, 11 February 2019 (EST)

Has one of the new Award search changes perhaps caused problems with the Award Directory link. I'm getting a Python script error using Firefox. I can provide more details if you are unable to recreate. Thanks. --Ron ~ Rtrace^Talk 18:59, 11 February 2019 (EST)

Oops! Let me see if I can fix it real quick... Ahasuerus 22:19, 11 February 2019 (EST)

OK, I think I got it. Sorry about that. Ahasuerus 22:46, 11 February 2019 (EST)

(unindent) Advanced Award Search has been implemented. You can't specify "Award Level" at this time, but I hope to add it shortly. Ahasuerus 20:36, 12 February 2019 (EST)

"Award Level" has been added. Nominations are a bit clunky, but there isn't much I can do about it because they are stored as "9"s in the database. Ahasuerus 15:49, 13 February 2019 (EST)

The Advanced Award Search results table has been modified to display award authors when available.

At this point the redesign of the Advanced Search is mostly finished. I plan to make a few quality of life enhancements and fix a bunch of bugs related to the way AND and OR work with Notes and third party URLs, but the main functionality is in place. How does it look? Anything you would like to see adjusted or changed? Ahasuerus 16:26, 14 February 2019 (EST)

(unindent) As per FR 1229, "Language of an Included Title (list)" and "Language of an Included Title (free form)" have been added to the Advanced Publication Search. Ahasuerus 14:08, 16 February 2019 (EST)

"Title Type" has been added to the Advanced Award Search page. Ahasuerus 18:55, 18 February 2019 (EST)

"Title (for title-based awards)" has been added to the Advanced Award Search page. Ahasuerus 18:55, 18 February 2019 (EST)

A. Williams / Arthur F. Williams

I've seen an enquiry on Facebook questioning our bibliography for the Astounding artist A. Williams, stating that it is possibly incorrectly shared with the British fanzine artist Arthur F. Williams of the same period and whose bio at Fancyclopedia makes no mention of providing art for Astounding. Is it possible they are two different artists, and can any editor shed any light? Thanks. PeteYoung 02:03, 15 February 2019 (EST)

I poked around earlier today, but couldn't find anything definitive. However, it looks like Arthur F. Williams was very active in fandom ca. 1940. Perhaps some digitized fanzine articles from that era may mention his Astounding work -- assuming it was him. Ahasuerus 20:19, 15 February 2019 (EST)

Award Category enhancement

The "category" column in the table that displays awards for an author/title/advanced search request has been enhanced. It used to link to the "Award Category" page which displays all awards for one category, which can get overwhelming for certain award categories, e.g. many Locus categories. It now links to a new Web page which displays all wins and nominations for one year.

For example, if you access John Brunner's The Squares of the City, the "Awards" table will tell you that it was a Hugo finalists in 1966. If you click the "Best Novel" link in the "Category" column, it will take you to Award Category: 1966 Best Novel (Hugo Award), which displays the 5 novels which were nominated for the Hugo award in the Best Novel category. The page is linked to the main Award Type page for Hugos as well as to the full award category page. Ahasuerus 18:45, 15 February 2019 (EST)

A different multiple COVERART question

Hopefully a simple one. This is based on some 19th Century volumes, so how it extends to today may not be relevant. If a book is published with different coloured covers, would it be reasonable to record a single publication with multiple COVERART entries? It might if you could associate a different image with each, but I don't think that's how it works. ../Doug H 09:29, 16 February 2019 (EST)

We have come across this scenario before. For example, the first edition of Patrick Rothfuss' 2007 name The Name of the Wind had two different dust jackets. One had the full version of Donato Giancola's painting while the other used a small portion of the painting as the cover image. We entered them as two different publication records.

One reason to create separate publication records for books with differently colored covers is that the color may have additional meaning. For example, to quote this record:

• Four-hundred and thirty-three copies have been prepared ... Six copies, bound in red leather, not offered for sale. Twenty-eight copies, bound in blue leather of which twenty-six have been lettered A through Z ... One-hundred copies, bound in blue cloth and numbered ...

Ahasuerus 09:58, 16 February 2019 (EST)

Awards revamped

The software that handles awards behind the scenes has been revamped. The changes were internal and shouldn't have affected what is displayed on any Web pages -- or at least that was the idea. Please let me know if you run into any bugs. Ahasuerus 20:36, 16 February 2019 (EST)

Publication series enhancements

"Publication Series" pages have been enhanced to display their contents faster. Long pub series pages like Perry Rhodan and Ace Double should take significantly less time to load. Ahasuerus 19:39, 23 February 2019 (EST)

Improving casual users' experience

(copied from my Talk page)

Most users would start with the home page. Once they get there, they have to wade through a lot of choices in small print that are useful only to editors. Enlarging the text in links, expanding the search capabilities and moving the links that are useful only to editors elsewhere or make them visible only to editors, would be a big help, IMHO, to casual users.-Rkihara 11:58, 26 February 2019 (EST)

It would be easy to restrict editing links to loged-in users. I should note, however, that displaying editing links for anonymous users was -- back at the dawn of time -- a conscious decision. The idea was that a casual user would see them, realize that he or she could help improve the data and click one of the links. The software would then tell the user that s/he needs to create an ISFDB account and things would develop from there.

Perhaps we overthought it. Another way to accomplish the same goal without displaying a bunch of irrelevant links would be to modify the message that anonymous users see at the top of each page. Currently it reads:

• You are not logged in. If you create a free account and sign in, you will be able to customize what is displayed.

If we were to add "and edit the data" to the end of the message, we should be able to remove editing links for anonymous users without losing functionality. Ahasuerus 13:03, 26 February 2019 (EST)

I think that would be a good change. ···日本穣 · 投稿 · Talk to Nihonjoe 15:20, 26 February 2019 (EST)

FR 1283, Do not display editing links unless the user is logged in, has been created. Ahasuerus 13:45, 24 May 2019 (EDT)

Performance issues -- 2019-02-27

As some of you may have noticed, the server was slow for about 10-15 minutes earlier this morning. The slowness was due to an apparent attack by a "botnet", i.e. a network of robots programmed to do malicious things. It's not clear whether they were trying to overwhelm the site with an excessive number of connections or whether they were probing it for security vulnerabilities, but they appear to be gone now. There was a somewhat similar incident a few days ago when a robot was probing the site looking for security vulnerabilities. Ahasuerus 12:01, 27 February 2019 (EST)

The botnet is back. It must have taken a quick break to munch on electronic cookies. Ahasuerus 12:04, 27 February 2019 (EST)

And it's back yet again. Ahasuerus 16:02, 3 March 2019 (EST)

Server downtime - 2019-03-07

The server will be brought down for maintenance at 2pm server time (North American Eastern Standard Time). I expect it to be back up by 2:15pm. The downtime will be used to improve performance and to add support for transliterated series names. Ahasuerus 13:16, 7 March 2019 (EST)

The server is back up; patch notes to follow shortly. Ahasuerus 14:18, 7 March 2019 (EST)

Transliterated series names

The software has been modified to support transliterated series names. The new multi-field is very similar to the other "transliterated" multi-fields that we have had for the last few years -- see Help:Screen:EditSeries for details.

A new cleanup report has been deployed. The data will become available overnight. I plan to create additional language-specific cleanup reports to look for non-Latin series with Latin names. For example, the Japanese series Sword Art Online: Progressive presumably needs to have its Japanese name added. Ahasuerus 14:35, 7 March 2019 (EST)

Language-specific cleanup reports have been deployed. The data will become available tomorrow morning. I expect that around 100 Cyrillic series and approximately 300-400 Japanese series will require attention. Ahasuerus 18:03, 7 March 2019 (EST)

Awesome! How do we want to handle the series that have two languages in the current title? I ask because the English title is not the same as a transliterated title, so it can't be put in as a transliteration. ···日本穣 · 投稿 · Talk to Nihonjoe 19:40, 7 March 2019 (EST)

I have been transliterating the non-Latin parts of series names and ignoring the Latin parts. Since we don't have a "variant" mechanism for series, some names are kludgy, but it's the best we can do for now. Ahasuerus 20:07, 7 March 2019 (EST)

Advanced Search Enhancements: The Next Generation

I am starting the next sequence of Advanced Search enhancements. The first patch was installed a few minutes ago. It fixed a Python bug that was causing certain Advanced Award searches to error out.

I expect that the next patch or two should have no impact on user experience. If I mess up and some parts of Advanced Search get broken, please post the affected URL(s) here. Ahasuerus 20:36, 9 March 2019 (EST)

Patch #2 has been installed. The URLs of record-specific Advanced Search selection criteria Web pages have changed. If you have them bookmarked, you will want to update the bookmarks. Ahasuerus 16:20, 10 March 2019 (EDT)

Patch #3 has been installed. The only user-visible change was the increase in the maximum number of supported selection criteria from 5 to 6. Ahasuerus 11:24, 13 March 2019 (EDT)

Patches #4 and 5 have been installed. Search results pages have been modified to display the selection criteria that were used to create the results. Ahasuerus 21:07, 15 March 2019 (EDT)

Date of Project Gutenberg ebook publication

For two Project Gutenberg ebooks we have publication records (linked here) with apparently spurious 1st-day-of-the-month publication dates. The official release/posting data are complex and I quote them here for reference.

• P363869 as 2005-05-01

Posting Date: September 19, 2014 [EBook #8180] Release Date: May, 2005 First Posted: June 26, 2003

• P416647 as 2006-02-01

Posting Date: November 23, 2011 [EBook #9956] Release Date: February, 2006 First Posted: November 4, 2003 Last Updated: November 14, 2015

Is it the official "Release Date" that we should enter as publication date? That is, in these two cases, simple correction to 2005-05-00 and 2006-02-00? --Pwendt|talk 14:43, 18 March 2019 (EDT)

See R.39 from their FAQ. For books with catalog id < 10k, they had a scheduled release date and an actual release date. They ended up being significantly ahead for awhile. So for example, in the first case, they scheduled to release in in May 2005, but actually released it June 26, 2003. -- JLaTondre (talk) 17:18, 18 March 2019 (EDT)

The Author and The Earl

1. If I understand correctly, the Directory Entry for an Author record should never begin with "The" (nor with "A" as an indefinite article, nor other cases?). And the name, or Author field value, should begin with "The Editor(s) of ..." only where that appears on the title page of a book, but for a magazine always "Editors of ...".

Thus "The Editors of the Saturday Evening Post" is probably correct here T1645902, as a variant of canonical author "Editors of the Saturday Evening Post".

That is, a variant of our Author name for publication records of non-genre magazine The Saturday Evening Post. Right?

Template link repaired (t, not p). --Pwendt

But the Author search report for "The Editor" [1] is mainly a list of mistakes. Right?

See #The Editor, below. --Pwendt|talk 14:04, 27 March 2019 (EDT)

2. Directory entries for (The) Author(s) of ... and (the) Editor(s) of ... should begin with "Author(s)" or "Editor(s)", evidently from the great preponderance entered thus. "Author" alone is the majority choice, but we also have many entries such as "Author of 'Our Neighborhood,' ".

The same should be true for (The) Publisher(s) of ..., I feel sure.

3. What about "(The) Earl of ...", and others of his ilk? --Pwendt|talk 18:11, 20 March 2019 (EDT)

SFERA award

Hi everybody! Is it possible to add SFera award. It was was established by SFera in 1981, Croatian SF comunity from Zagreb.

http://sfera.hr/sfera-for-foreigners/

http://sfera.hr/sfera-for-foreigners/the-sfera-award/

list of winners since 1981

http://sfera.hr/nagrada-sfera/popis-dobitnika/ Debolestis 07:18, 24 March 2019 (EDT)

It looks like a legitimate award. Most -- perhaps all -- categories should be eligible for inclusion. Ahasuerus 07:56, 24 March 2019 (EDT)

Great, I am not sure what is a procedure with this. Debolestis 08:14, 24 March 2019 (EDT)

If there are no objections in the next couple of days, I will create the new award type. Then we can create its award categories, enter the awards/nominations, etc. Ahasuerus 13:33, 24 March 2019 (EDT)

Sorry, I got distracted while working on Fixer. I will create the award type tomorrow. Ahasuerus 23:44, 28 March 2019 (EDT)

(unindent) The new award type and its categories have been created. Hopefully the current "display order" of the categories is reasonable. They seem to be fairly straightforward, but I would appreciate it if you could check that I didn't mess anything up. For example, I assumed that "Roman za djecu", "Dječji roman" and "Roman za mlade" were just successive names of the same category ("Juvenile Novel"), but it's possible that their scope may be different.

Other than that, I think we should be ready to start entering awards. The one that I have entered seems to look OK. Ahasuerus 11:05, 29 March 2019 (EDT)

This is great, thank you a lot! I'll start entering awards this week. Your translations are all good. Debolestis 18:18, 30 March 2019 (EDT)

One more thing, until 1991 it was awarded for Yugoslavia, from 1994 only for works published in Croatian. https://en.wikipedia.org/wiki/SFERA_Award#1981. Debolestis 18:18, 30 March 2019 (EDT)

Updated, thanks. Ahasuerus 19:27, 30 March 2019 (EDT)

Do I enter awards such as "comic book" or "illustration" when it was not given to a specific illustration? Do I use untitled award in this case? Debolestis 18:38, 30 March 2019 (EDT)

We have a number of "comic book" and "illustration" award categories on file -- see this list and this list. I would suggest following the same format and consulting Help:Screen:AddAward for the specifics. Ahasuerus 19:27, 30 March 2019 (EDT)

P.S. One category name that I wasn't sure about is "Povelja". The word is supposed to mean "charter", but awards in this category are apparently given to fanzines and articles. What would be a good way to translate it? Ahasuerus 19:37, 30 March 2019 (EDT)

"Povelja" means "charter", that would be good translation. "Protosfera" is given to best new young author. Debolestis 11:29, 31 March 2019 (EDT)

Updated -- thanks again. Ahasuerus 12:34, 31 March 2019 (EDT)

The Editor

Looking again at works we credit to The Editor(s), it occurs to me that it may be our policy to distinguish genre and non-genre magazines thus:

• for genre magazines where possible, credit particular editors (Author names) and publishers by name;
• for non-genre magazines, credit "Editors of [title]" even if the editor is a known person and optionally leave publisher field empty.

Three years' issue of The Pall Mall Magazine have been handled differently. The magazine is non-genre (although not all years' issue have been so tagged). The named 1899, 1908, and 1914 editors have no other works in the database and, I surmise or propose as necessary, they should not be named (except in Notes: series, annual Title records, or perhaps Author: Editors of The Pall Mall Magazine). --Pwendt|talk 16:27, 27 March 2019 (EDT)

This is allowed per these instructions: "If the actual editor is known with reasonable certainty, you may optionally enter the name(s) of the editor(s) as co-editors, but leave "Editors of PERIODICAL NAME" as one of the editors in any case." --Ron ~ Rtrace^Talk 21:18, 27 March 2019 (EDT)

Anyway, sometimes we do name the editor(s), as Hugo Gernsback alone for Wonder Stories, or Roger Dobson and Mark Valentine for Aklo (eg, ESSAYs linked below). In those cases, we may credit ESSAY contents of the magazine publications to that person or those people, regardless whether the work is uncredited or credited to The Editor(s) singular or plural. And we may do that directly, rather than by variant Author names such as "uncredited" and "The Editor ...". Should we always do that?

For instance, 1934 ESSAY "Our New Policy" by The Editors of Wonder Stories T2198828 should be credited to directly to Gernsback alone. And for clarity of the Gernsback bibliography, it should be in Series: Editorial (Wonder Stories) S26754. Right? (At the moment that series contains two titles as by "uncredited", all others as by Hugo Gernsback. There may be other Gernsback ESSAYs that belong in the series --perhaps everything credited to him as editor of Wonder Stories.)

For instance, 1988 ESSAY "The Arthur Machen Society Appeal" T2279843 should be credited directory to Dobson and Valentine jointly. And for clarity such ESSAYs should be segregated from others in their bibliographies, using an editorial Series. --Pwendt|talk 14:41, 27 March 2019 (EDT)

The idea behind treatment for non-genre magazines is to avoid creating Author records for people not associated with the genre. So the policy is to credit Editors of XXX always in such situations. If there's an author record for any of the editors, they may be added so that the magazines are listed in their bibliographies. --MartyD 21:58, 27 March 2019 (EDT)

Windy City Pulpcon

If anyone here is attending the Windy City Pulpcon on April 12-14. Let me know if you'd like to meet up.--Rkihara 13:18, 1 April 2019 (EDT)

As usual, I'll be there. It's practically in my back yard. I'd be happy to meet up if you're interested. Bob 17:43, 1 April 2019 (EDT)

Sure, I'd like that. I'll be registered at the Westin as Ron Kihara and will be arriving Thurs. evening. I usually drive down I-80 from the Bay Area, so it's possible the weather might prevent me from making it.--Rkihara 19:20, 1 April 2019 (EDT)

Great, Ron. I'll get there about 11 a.m. Friday when the dealer room opens. I'll hang around the entrance for a while, then wander through the ballroom. I'm Bob or Robert Lumpkin. If we somehow miss on Friday, I'll be there about 10 am on Saturday with my adult son. I'll hang around with dealers like Greg Ketter (Dream Haven Books), Lloyd Currey, Andy Richards (Cold Tonnage Books), Graham Holroyd, and Bill Cavalier. There are usually a bunch of friends from the Robert E. Howard community as well. Bob 20:06, 1 April 2019 (EDT)

I'll be going into the dealer room about the same time. I'll be easy to spot, I'm Japanese, wear glasses, bald, and I'm told I look like a Kung Fu villain. The dealers I usually spend time with are; Steve Hafner, Larry Hallock, Craig Poole, Dave Kurzman and Steve Spilger. I'm also in and out of the hospitality room.--Rkihara 22:39, 1 April 2019 (EDT)

Template that can be used in the wiki

I've created Template:Notice float right for anyone that wishes to use it. If you visit the template page, it gives an example of how to use it. You can also see it in use on my user page. Enjoy! ···日本穣 · 投稿 · Talk to Nihonjoe 16:15, 5 April 2019 (EDT)

Changes to ISFDB links

The way ISFDB links -- including links to author, title, series, etc records -- are displayed has been changed behind the scenes. The new linking format is supposed to be supported by all browsers going back to Internet Explorer 6, so the change should be completely transparent to all users.

Still, given the variety of browsers out there, you can never be 100% sure. If a link doesn't work as expected, please post the offending URL and your browser version here. Ahasuerus 16:11, 6 April 2019 (EDT)

Does this address the issue of Facebook adding the "&fbclid=" bit onto URLs? ···日本穣 · 投稿 · Talk to Nihonjoe 12:08, 8 April 2019 (EDT)

I am afraid not. It's a different change addressing a different issue. Ahasuerus 14:02, 8 April 2019 (EDT)

Some users report issues under certain versions of Safari. Investigating. Ahasuerus 11:31, 14 April 2019 (EDT)

Unfortunately, there wasn't anything that I could do to address the Safari bug(s). For now, I have reverted the software changes to enable Safari users to access our data. Ahasuerus 11:57, 16 April 2019 (EDT)

Out of curiosity, what is the technical change? Uzume 19:29, 18 September 2019 (EDT)

Canonical name Thomas Blot or William Simpson (I)

In relation to the 1891 pseudonymous NOVEL The Man from Mars, as by Thomas Blot T1013407:
This week I added the 1900 Third Edition, Revised and Enlarged, by William Simpson (I), and the 1893 stated second edition, identical to the first but also under the author's real name. --both of those as publications one NOVEL by William Simpson (I) T2536737, although 80 of 280 pages are new in the Third Edition, a 60-page Preface and a new chapter. Most information on how the 1900 differs from the earlier eds. is now in the 1900 publication record P710623.

That is: 1893 change in credited author name; 1900 change in contents --but not so great that we should treat it as a new NOVEL, in my opinion. Right? If so, then one NOVEL title should be variant of the author, depending on which should be the canonical author name.

Neither Blot nor Simpson (I) is likely to enjoy a growing bibliography here. SFE3 knows the writer as Blot, from the 1st edition. Library of Congress doesn't know him at all.

Thomas Blot or William Simpson (I)? That governs the (re)location of all Author data and much NOVEL data to the appropriate parent record, and whether the Novel by Simpson should be dated 1891 or 1893? --Pwendt|talk 16:53, 6 April 2019 (EDT)

(fix template links, rewrite and expand) --Pwendt|talk 17:52, 6 April 2019 (EDT)

I think "Thomas Blot" is the way to go. As you said, there is little chance of either list growing significantly. Uzume 19:22, 18 September 2019 (EDT)

Arthur C. Clarke Award - Categories

The Arthur C. Clarke Award is currently set up to have the following categories:

• Winner
• Runner-Up
• Shortlist
• Honorary Award

It would appear that the first three categories as they are currently defined in the database are actually different "award levels" of the Best Science Fiction Novel "award category". My guess is that the editor(s) who originally created these categories were trying to get around the fact that -- the way the software currently works -- an award category has to be either a "Poll" or "Not a Poll". As luck would have it, the Best Science Fiction Novel category doesn't fit this mold. In 1987-1996 there was one "winner", one "runner-up" and a number of nominations every year, but "runner-ups" were eliminated in 1997. Whoever originally entered the date presumably tried to get around this problem by creating additional categories.

It occurs to me that there may be another way to organize the data. We could merge the three categories into one, "Best Science Fiction Novel", and set it up as a "Poll". Winners would then be entered as "Award Level 1", runner-ups as "Award Level 2" and regular nominees as "Finalists" (a supported "special award level".)

What do you think? Ahasuerus 19:54, 9 April 2019 (EDT)

Sounds good. It also looks like there were 3rd place winners announced in 1989 and 1990, per the SFADB. The Award's site lists only winners and shortlist for all years. --Ron ~ Rtrace^Talk 06:58, 10 April 2019 (EDT)

Done! Ahasuerus 15:46, 14 April 2019 (EDT)

Wiki page elimination

Moments ago I made three submissions related to the ISFDB wiki page Author:Jonathan Scott. That page gives a medium-length quotation from the British DNB.

1. As the DNB quotation concerns Scott's edition of the Arabian Nights alone, [1] I copied all of it (typography unmodified) as a Note on Title 1340796.
2. I copied most of the first sentence (typography modified) as a Note on Author Jonathan Scott.
3. On the Wiki page I added a note, with signature and timestamp, explaining how the page is now redundant (entirely).

I have wondered about the procedure for wiki page elimination. I don't work on it systematically but I have added some Notes to the database that make some Author, Bio, and Publisher wiki pages redundant. Is it desirable to annotate the wiki pages as I have done here? Alternatively, is there another place to report that a page is now redundant?

The Author page is entirely redundant (pending approval of these submissions) in that all of its content is available in the database. Perhaps I should have included a link or some instruction in the Author Note --as I would have done for someone with a medium or long bibliography. For instance, minimally

• "In 1811 Scott published the work by which he is chiefly known, his edition of the Arabian Nights" in six volumes (the 1811 work listed below).

--Pwendt|talk 14:39, 10 April 2019 (EDT)

When content from a wiki page has been moved to the database, add the following to the top of the page: {{Page transferred}}. That will tag the page for deletion and it will (eventually) get deleted. You can also use {{Deletion candidate|your reason here}} (filling in your reason) for pages that are not a simple transfer case. I have deleted the author's page. -- JLaTondre (talk) 16:06, 10 April 2019 (EDT)

New Help page organized

I created Help:How to (wiki) (which previously only existed as a list of pages in a category). I organized the pages by what they discussed. ···日本穣 · 投稿 · Talk to Nihonjoe 15:01, 12 April 2019 (EDT)

SFE3 Reconciliation - New Report

The way our author URLs are reconciled with SFE3 has been changed. A new cleanup report has been created and deployed. The new functionality includes:

• All editors have access to the report (only moderators can "ignore" SFE3 URLs)
• The cleanup report is now updated nightly with the latest SFE3 data
• The software analyzes SFE3 article names and tries to guess what the corresponding ISFDB author should be

If everything goes smoothly, the new data will become available in a couple of hours when the nightly process runs. Ahasuerus 23:15, 13 April 2019 (EDT)

The nightly process ran successfully and added 4380 SFE3 URLs to the report. I have cleaned up the first 10 and the process seems to be working correctly. Things are looking up. Ahasuerus 11:33, 14 April 2019 (EDT)

Thanks!--Dirk P Broer 06:34, 2 May 2019 (EDT)

Glad it's useful :-) Ahasuerus 06:48, 2 May 2019 (EDT)

Inconsistent category names on Locus Poll Award in 2007, 2009 and 2010

I notice that there are a few instances of categories for the Locus Poll Award not being prefixed by "Best". This is visible at http://www.isfdb.org/cgi-bin/awardtype.cgi?28 , but a rough list of the offending categories/years is:

• Collection (2010)
• Fantasy Novel (2007, 2009, 2010)
• First Novel (2007, 2009)
• Nonfiction/Art Book (2010) - note that the "Best" category is "Best Nonfiction/Art", no "Book"
• Novelette (2010)
• Novella (2007, 2010)
• Short Story (2010)

Anyone know if there's a good reason for this? It certainly seems less than ideal from a UX point-of-view. I'm a newbie here, so I dunno how easy it would be to merge these back into the more common "Best..." categories - it looks like there might be enough records for some sort of SQL bulk UPDATE to be better than multiple updates via the UI?

(FWIW, Wikipedia has them all as "Best Whatever", SFADB just has "Whatever". Locus' site uses "Best Whatever", at least for 2010: https://locusmag.com/2010/06/2010-locus-awards-winners/)

In a similar vein - but it is at least consistent within the award - the Nebulas are just categorized as "Novel", "Novella" etc, whereas on the SFWA site they are "Best Novel", "Best Novella" etc. e.g. https://nebulas.sfwa.org/award-year/2018/

--ErsatzCulture 19:19, 14 April 2019 (EDT)

I can think of two possible explanations:

1. The person who created the additional categories overlooked the fact that related "Best" categories were already on file
2. The person who created the additional categories was working with what s/he considered an authoritative source -- perhaps Locus issues? -- and s/he wanted to be as accurate as possible

Normally, a minor name change doesn't result in the creation of a new category, so my inclination would be to merge them and to document any discrepancies that we may find in the Note field. Let me check submission history to see who created these categories first... Ahasuerus 20:07, 14 April 2019 (EDT)

I have scanned the submission table looking for clues. Here is what I think is happening with these categories.

Originally, award types and categories were handled differently. Award types used to be hard-coded using two-letter abbreviations -- see "award_ttype" on this Wiki page. Award categories used to be manually entered every time an award was entered -- see "award_atype" on the same Wiki page.

It was very messy and I ended up rewriting the whole thing in 2014 -- see Development/Archive/2014 for details. I then migrated previously entered award categories to the new table structure. I think I tried to merge near-duplicates where feasible, but it looks like I missed quite a few.

I think it should be safe to merge these categories. Roughly 200 awards will be affected, which shouldn't be too hard to do manually. Ahasuerus 21:22, 14 April 2019 (EDT)

OK, I think I have merged everything that could be merged without losing specificity. Please let me know if I missed anything. Ahasuerus 21:30, 17 April 2019 (EDT)

Fantasy and Science Fiction mentioned ISFDB

I forgot to mention it back on April 4. See this post. ···日本穣 · 投稿 · Talk to Nihonjoe 20:08, 16 April 2019 (EDT)

Nominating Biomassbob as Moderator

I would like to nominate Biomassbob (talk • contribs) for moderator. He has been an editor for seven years and has 44530 edits as of this date, covering the full range of the database. I have always found his edits to be detailed and complete. I believe that he is highly qualified to be a moderator and he has accepted the nomination.

Support

1. Support, as nominator.—Rkihara 10:41, 20 April 2019 (EDT)
2. Support. No concerns; I can't remember the last time I had an issue with any of his plentiful and varied submissions. p.s. Thanks for making me feel old, Ron. --MartyD 17:05, 20 April 2019 (EDT)
3. Support. About time :) Annie 17:55, 20 April 2019 (EDT)
4. Support. I have no concerns. ···日本穣 · 投稿 · Talk to Nihonjoe 19:38, 22 April 2019 (EDT)

Neutral

1. I haven't done a lot of moderating lately, so I'll abstain. Ahasuerus 19:16, 23 April 2019 (EDT)
2. I can't remember having moderated enough of his submissions to make a judgement, therefore I stay neutral. Jens Hitspacebar 13:21, 25 April 2019 (EDT)

Outcome

The nomination was successful; the moderator flag has been set on the account. Congratulations! Ahasuerus 11:29, 27 April 2019 (EDT)

Cropping Amazon images

Dirk Stoeker has identified a way to crop Amazon images:

• If the image contains unwanted elements (like large background or other books) you can add cropping before the final period (strip every other modification before): ._CR102,6,148,257 - The arguments are left, top, width, height (example: ecx.images-images-amazon.com/images/I/41SpdLJPcRL._CR102,6,148,257.jpg)

However, the way our software is currently configured, an Amazon image URL that contains "L._" generates a yellow warning ("Extra formatting in an Amazon URL") on post-submission pages. It also generates an exception on one of the nightly cleanup reports.

That said, it would be easy to modify the software to accept "L._CR" URLs, so the question is whether we want to allow them. On the one hand, "cropped" URLs can come in handy when dealing with unusual Amazon images. On the other hand, Amazon URLs are not always stable and this particular method has been apparently changed in the past. Here is what an overview of Amazon images says about it:

• This crops the image, trimming away everything but a selected rectangular area. The first two numbers indicate the upper left-hand corner of the selection area (the first is how many pixels from the left, the second is how many pixels down from the top). The next two numbers are the width and height of the selected area, in pixels.
• Note: this was a great way to trim away the excess white space left by other commands,,. which other commands used to create, although that was greatly reduced sometime around July 2007. You have to take the time to crop precisely for a given image. If you set the crop area to go beyond the edge of the image, the image will merely be resized with white space added.

Based on this history, I wonder if it may be better to create a cropped image on your computer and then upload it to ISFDB. What do you think? Ahasuerus 19:26, 20 April 2019 (EDT)

Amazon tends to change their formatting when it suits them and relying on it may end up with some weird side effects when they do in a few years. If the Amazon image is not good as is, then we either should live with it or find another one or fix it -- but relying on an undocumented and non-guaranteed external formatting is just asking for trouble. Annie 21:46, 20 April 2019 (EDT)

Downloading, cropping and uploading to ISFDB is not an valid approach, as there still can be copyright on the images (even if it is debatable if copyright can exist for a mere photo of a book). As far as history of Amazon shows they simply ignore old parameters when they change, so in this case the result would be an uncropped image again. --Stoecker 04:17, 21 April 2019 (EDT)

Copyright is irrelevant for this discussion. We use non-public domain cover images under fair use criteria. -- JLaTondre (talk) 08:32, 21 April 2019 (EDT)

Not for me. Even if the use of the motive is ok, modifying the photo taken by somebody else is a minefield I wouldn't want to step into, especially with the upcomming copyright changes in Europe. --Stoecker 11:58, 24 April 2019 (EDT)

Amazon covers are not always reliable, especially for older works. I'm not thrilled with the idea of locally uploading Amazon images, especially for works that were verified without a cover. If the cover is actually for the wrong edition, it creates a false record where it is harder to see that the image is not correct. Especially since uploaders are incredibly poor at properly changing the source field in the fair use template. The crop functionality has been around for sometime. If change were to occur, it would probably be that Amazon deprecates it, in which case, it would go back to showing the whole image so not much of a risk in my opinion. -- JLaTondre (talk) 08:32, 21 April 2019 (EDT)

While I don't care for relying on specifying the cropping in the URL, I don't see any harm in allowing it, and it does seem a little better to me in terms of preserving the provenance of the image than we might get if we recommend downloading and cropping and uploading. --MartyD 09:57, 21 April 2019 (EDT)

It looks like we have 3 editors in favor of allowing "L._CR" URLs, 1 opposed and 1 neutral (me). Would anyone else like to share his or her thoughts? Ahasuerus 11:06, 25 April 2019 (EDT)

If everyone else is in favor of it, I am not going to block the decision - so ignore the opposing editor. :) It is an easily reversible one if Amazon managed to make it a disaster with an update - because of the format, we can always strip those programmatically later. Annie 13:20, 25 April 2019 (EDT)

(unindent) OK, we have a consensus or as close to it as we are likely to get. I have changed the software to ignore "._CR" in Amazon URLs and updated Template:Image Host Sites with Dirk's instructions. In addition, the new cleanup logic is somewhat more robust than the old logic, so I expect it to find 30 new "problem" images tomorrow morning. Ahasuerus 19:59, 26 April 2019 (EDT)

Cleanup completed. Ahasuerus 14:14, 29 April 2019 (EDT)

Abridgements

There doesn't seem to be anything specific in the Help pages about abridgements, specifically what constitutes an abridgement, how to credit additional 'authors'/'editors' or note them if unspecified, varianting to the original title, effects of translation (e.g. abridgement of a translation vs. translated and abridged vs. abridged then translated) and distinguishing abridgements of an oft-abridged title. Are there rules or even consensus for these topics? If this turns into a discussion of actual rules, it should be moved to that forum, but for now I'm wondering about where we stand. ../Doug H 12:05, 29 April 2019 (EDT)

In my experience, it depends on the nature of the abridgement. If the differences are minor, they are documented in Notes. If they are major, a new title record is created.

If the person who performed the abridgement is known and a new title needs to be created, s/he is credited as a co-author -- see William Shakespeare's bibliography for examples. If the text was rewritten so extensively that it qualifies as a separate work (typically retellings for young children) then the original author is sometimes only credited in Notes. Ahasuerus 18:10, 29 April 2019 (EDT)

If we stick with crediting per the pub and it doesn't list the abridger on the title page, then it should be entered as by the author alone and varianted to as by the author and the abridger. We are inconsistent in doing that though. -- JLaTondre (talk) 18:45, 29 April 2019 (EDT)

If you variant to the title with the original author and abridger, then you can't variant that in turn to the original. Does that mean abridgements aren't supposed to be varianted to the original? Is there any way to keep track of the relationship? ../Doug H 19:37, 29 April 2019 (EDT)

Variants are the same work under a different title (or translations). They are not for variance in the work itself (other than translations). If an abridgment is different enough to be considered a new work, then it would not be varainted to the original work. If the changes are minor enough to not be considered a new work, then there would not be a separate title & it would be simply captured in the notes (as Ahasuerus referred to above). If it's different enough to have its own title record, then you would need to use the notes to link it back. The only exception would be if it's a minor abridgment and a title change (then a variant would be appropriate). -- JLaTondre (talk) 20:03, 29 April 2019 (EDT)

New External ID type for Science Fiction-Leihbuch-Datenbank?

The German SF database Science Fiction-Leihbuch-Datenbank has a fairly limited scope, but their data is solid and well organized. Their publication-specific pages can be accessed using "permalinks" like http://www.sf-leihbuch.de/index.cfm?bid=1691, which should make it easy to add the site as a new "External ID" type. Can anyone think of a reason not to add it? Ahasuerus 15:34, 3 May 2019 (EDT)

No reasons I can think of. --Ron ~ Rtrace^Talk 11:02, 5 May 2019 (EDT)

Me neither. Stonecreek 13:33, 5 May 2019 (EDT)

OK, FR 1274 has been created. Ahasuerus 14:12, 5 May 2019 (EDT)

A new External ID and a matching linking Notes template have been added -- see this record for an example. Ahasuerus 14:17, 6 May 2019 (EDT)

The Lord of the Rings / The Lord of the Rings (Boxed Set)

Hi, everybody! I just wonder why we have different titles for these: it seems to be somewhat illogical by our internal logic. After all, as for the titles involved in the respective OMNIBUSes, they should be identical; the form of publication shouldn't be reflected in the title, only in the publications. Am I missing something? Stonecreek 09:57, 5 May 2019 (EDT)

Agreed. It might relate to this discussion (at one time the single volume Lord of the Rings was apparently considered a novel). However, since they are both omnibuses, they should be merged. -- JLaTondre (talk) 10:10, 5 May 2019 (EDT)

Yes, I dimly remember that discussion. If no one objects, I'll do the merging: in any case, I'll wait until Tuesday before beginnin with it. Stonecreek 13:38, 5 May 2019 (EDT)

Can't find the original now that they're merged (without an long manual search), so a question - was the entry for the box itself, which had a separate ISBN and cover image than the books it contains? Should the other boxed sets be dealt with the same way? Two contrasting examples - James Blish Cities in Flight has a single publication for the set like Lord of the Rings, but A Game of Thrones: 5-Book Boxed Set will (hopefully) never see a single omnibus publication. ../Doug H 08:53, 8 May 2019 (EDT)

Re: GRRM. You never know, I think, though a printed one-volume edition will take a long time (and some muscles to carry home), there'll likely be an ebook someday.

Re: LotR. There were several boxes, some with one ISBN, some with separate ISBNs. I remember that there were several around the movie(s) release(s), apart from the one mentioned in the title note. Ideally, the page count should be split into three portions (XXX+YYY+ZZZ), though there was one (unverified) set that had only one number.

And yes, this was one example of the underlying principle, I should think. I have merged the Blish OMNIBUSes: thanks for pointing them out. Stonecreek 09:58, 8 May 2019 (EDT)

Lots more where those came from. Just pulled them from the list of 200+ titles with "boxed set" or "boxset". Didn't even try other variations. ../Doug H 19:29, 8 May 2019 (EDT)

To be merged, they would have to have the same contents and same title (minus the artificial box set disambiguation we add). If they had the same contents and different title, they would be varianted. There may be a few more cases out there if someone has the patience to go through them all, but I randomly poked through a bunch of the results from doing such a query and only found one. -- JLaTondre (talk) 21:58, 8 May 2019 (EDT)

So if there is no existing omnibus with a matching title when you enter the box set, the generated TITLE title would also include "box set", until such time as someone entered the matching omnibus without a "box set", at which point the box set should be merged with the second omnibus title? The inclusion of "box set" is acceptable as long as there isn't an existing non-boxed omnibus? In the case of Asimov's Foundation Trilogy, the box set title entry explicitly states it is not to be merged with the single volume omnibus. So it sounds like someone had a different opinion that a moderator agreed with. No verifications, so no names or dates. And a question for Stonecreek - what do you mean "by our internal logic"? And as a final observation - I thought box sets were different and were put in as OMNIBUS because that was the closest thing without coding a new type - which is the only reason I keep picking at this, I don't yet understand. ../Doug H 22:53, 8 May 2019 (EDT)

No, it seems that (box set) wouldn't belong to the title in general: it crept into our database just because vendors such as Amazon have that title, which only belongs to a specific publication, and is not to be found anywhere in or on the publication itself. Stonecreek 00:26, 9 May 2019 (EDT)

Crediting cover scans - proposed design changes

Here is how cover scan credits work at this time. When we link to a cover scan, we credit the organization operating the server that hosts the scan. The credit line reads "Cover art supplied by X", where [X] is the name of the organization (Amazon, Galactic Central, etc) linked to the organization's home page. Clicking the image takes you to the raw image hosted by the third party. Here is an example of how it currently works.

There are two partial exceptions to this rule. First, images hosted by the SFE3 Picture Gallery are handled differently. Due to SFE3 requirements, clicking an SFE3-hosted image takes you to the Gallery Web page that hosts it as opposed to the raw image. For example, if you click the cover of this publication, you will see its Gallery page, not the raw image. This was done in order to support SFE3 requirements -- see Template:Image Host Sites, which reads, in part, "all SFE3-hosted images must have a link to the associated "Gallery" page added after a "pipe" ("|") character", for an explanation of the technical details. (Ignore the fact that the linked publication page currently credits "Encyclopedia of Fantasy" instead of SFE3. It's a recently discovered bug which needs to be fixed. The bug has been fixed.)

The second partial exception is ISFDB-hosted images. As expected, clicking an ISFDB-hosted image takes you to the raw image. However, clicking the word "ISFDB" on the "Cover art supplied by ISFDB" line takes you to the ISFDB Wiki page for the image as opposed to the ISFDB home page. Here is an example of how it works.

This is inconsistent design because seemingly identical links behave differently depending on whether the cover scan is hosted by ISFDB, SFE3 or another party. To make it worse, recent discussions with the administrator of Science Fiction-Leihbuch-Datenbank have identified additional requirements for linking to their cover scans. They would like "Cover art supplied by" lines to link to their publication-specific pages the way they currently link to ISFDB Wiki pages for ISFDB-hosted scans. We could modify our software to accommodate their requirements, but it would further muddy the waters. Besides, who knows what kinds of additional linking requirements other sites may have in the future?

An e-mail discussion of this issue with the administrators of Science Fiction-Leihbuch-Datenbank and SFE3 ultimately resulted in the following proposal:

1. Change the software so that clicking a cover scan (including SFE3-hosted cover scans) should always take the user to the raw image file.

2. Modify the credit line to have two links instead of one. The first link will continue to be a link to the site's home page. The second, new, link will be to a publication-specific page (if defined.)

3. The new wording of the credit line will be "Cover art supplied by X on this Web page" where:

• "X" will be the name of the organization hyperlinked to its home page, and
• "this Web page" will be a link to a publication-specific page, specifically:
  • for SFE3-hosted images it will be the cover scan's Gallery page
  • for ISFDB-hosted images it will be the ISFDB Wiki page of the cover scan
  • for Science Fiction-Leihbuch-Datenbank-hosted images it will be the publication-specific page associated with the cover scan
  • all other cover scans won't have the words "on this Web page" displayed because there is no page to link to

For now we will keep the current way of entering links to publication-specific pages, which uses "|" to separate the cover scan URL from the URL of the publication-specific page. If the current proposal is accepted and implemented, I expect that we will revisit this issue at some point in the future and probably create a separate field for the data that is currently entered to the right of "|", but that's another discussion.

So, what do you think? Ahasuerus 13:56, 5 May 2019 (EDT)

Sounds overall very good to me! Stonecreek 13:45, 7 May 2019 (EDT)

To me as well. Jens Hitspacebar 14:50, 7 May 2019 (EDT)

Me three Annie 19:53, 7 May 2019 (EDT)

Sounds good to me. ···日本穣 · 投稿 · Talk to Nihonjoe 20:22, 7 May 2019 (EDT)

Thanks, folks. FR 1277 has been created. Ahasuerus 20:47, 7 May 2019 (EDT)

(unindent) The deed is done. (Hopefully without messing up anything else.)

BTW, there is another site that requires that we link to the underlying page -- Smashwords. We have a cleanup report that hunts invalid Smashwords links, but we need to add a yellow post-submission warning. I'll take care of it when I add Science Fiction-Leihbuch-Datenbank. Ahasuerus 17:53, 8 May 2019 (EDT)

Couldn't we get something similar for Amazon-linked cover art? I regularly get into a situation where I want to check out the actual page, so it'd be nice if I could click on a link and it'd bring me directly to the actual pub instead of to the frontpage where I then have to search for the pub I'm looking for...

Additionally, could it be done to have a link to any one of the Amazon sites, depending on which site is used to link to? For example, linking cover art from Amazon UK (or DE, or...) would link to Amazon.co.uk instead of to Amazon.com MagicUnk 14:00, 9 May 2019 (EDT)

Publications with ISBNs display links to the supported Amazon stores under "Other Sites" in the navigation bar on the left. Publications with ASINs link to the supported Amazon stores in the External ID section.

Having said that, we could add an explicit Amazon link to the credit line for Amazon-hosted images. Unfortunately, there is no way of telling which Amazon store the image is associated with because all Amazon images use the same URL structure (at least to the best of my knowledge.) I guess it means that we will have to link to Amazon.com. Ahasuerus 14:55, 9 May 2019 (EDT)

Ah yes, I could have found out myself if I'd paid attention. My apologies. Regards MagicUnk 17:56, 9 May 2019 (EDT)

Linking to SF-Leihbuch

SF-Leihbuch has been added to the list of third party Web sites that we are allowed to link to.

The software module responsible for post-submission warnings and cleanup reports has been rewritten to handle the three Web sites that require additional links for images -- SFE3, Smashwords and now SF-Leihbuch -- consistently. If you encounter any issues, please post your findings here. Ahasuerus 21:33, 9 May 2019 (EDT)

Crediting third party sites - Part 2

Now that the dust has settled and various queued up software tweaks have been implemented, I'd like to revisit the issue that I raised a week ago.

At this time there are three third party Web sites that require us to link to a particular Web page when displaying their images: SFE3, Smashwords and SF-Leihbuch. The way we do it is by entering the URL of the image, then the "pipe" character ("|"), then the URL of the third party Web page associated with the image -- see Template:Image_Host_Sites for site-specific details. In the future, we may be given linking permissions by other sites that have similar requirements.

I don't think the current design is a good long term solution. It's not consistent with how other data entry fields work, it can mess up Advanced Publication Searches and it makes the software more convoluted than it needs to be.

I propose that we create a new field for these links. Instead of entering something like "http://sf-encyclopedia.uk/gal/clute/FrazerS-Blow.jpg|http://sf-encyclopedia.uk/gallery.php?id=FrazerS-Blow.jpg" in the "Image URL" field, we would be entering "http://sf-encyclopedia.uk/gal/clute/FrazerS-Blow.jpg" in the "Image URL" field and "http://sf-encyclopedia.uk/gallery.php?id=FrazerS-Blow.jpg" in the new field. The new field would be optional, but we may be able to add pop-up validation to make it mandatory for the sites that have this requirement. Post-submission warnings and cleanup reports would be adjusted accordingly. Existing data would be migrated automatically.

I am not sure what a good name for the new field would be. Something like "Web page of the image"? Any reasons not to add it? Ahasuerus 18:39, 11 May 2019 (EDT)

Breaking it out into a new field seems reasonable, but it needs to have a name that is not confusing. Most users (especially new ones) will not be familiar with the unique requirements for this field. The name should be something that makes it clear it's only needed in certain cases (or another option, though more complicated, would be to use JavaScript to only make the field displayed when an URL for one of those sites it entered). I'm struggling to come with an good option, though. -- JLaTondre (talk) 10:50, 12 May 2019 (EDT)

Unfortunately, I can't think of a good name either. For now, I have created FR 1282, Create a new publication field for crediting 3rd party sites. Based on the limited feedback, I assume that it's relatively low priority. Ahasuerus 13:35, 24 May 2019 (EDT)

"SF Calendar" added

A new Web page, SF Calendar, has been added as per FR 998. It is accessible from the "Other Pages" section of the navigation bar on the left. After selecting a day, you are taken to the list of authors who were born and who died on the selected day.

Let's wait a few days to let everyone get used to the new functionality and address any page layout issues that may arise.

Once the layout is stable, we can revisit the discussion which prompted the creation of this FR in 2017: what do we want to appear at the top of the home page? Now that the birth/death data is available on a separate page, can we remove it from the front page? Pare it down? Replace it with a link to "SF Calendar"? Should we display other links at the top of the home page, e.g.:

• Author Directory
• Award Directory
• Publisher Directory
• Magazine Directory
• Statistics/Top Lists

? Ahasuerus 17:01, 13 May 2019 (EDT)

Enhancing the "Chronological Bibliography" page?

A new Feature Request, FR 1280, "Create a noncategorized-chronology page for authors?" has been created on SourceForge. It reads:

I frequently want to view all of a given author's works in a single chronological list, rather than a set of categorized chronological lists. Is there already a way to do that? (When I click "Chronological", I get categorized lists.)

Currently, "Chronological" author pages ignore series groupings. However, they display novels, collections, anthologies, magazines, short fiction, essays, etc in separate sections. "Alphabetical" pages use the same approach.

The proposal, as I understand it, would add a new Web page for author bibliographies. The new page would display all titles chronologically without regard for the title type. I can see how it could be useful in certain cases, e.g. if it's not immediately clear whether a short novel appears in the NOVEL section or the SHORT FICTION section. At the same time I suspect that combining reviews, chapbooks, EDITOR records, etc in one long list may be confusing. Perhaps displaying the title type next to each title would help? Ahasuerus 14:58, 20 May 2019 (EDT)

That's a nice feature. The title type should definitely be displayed. Moreover, information about language and variants are quite helpful. The display format could be a mixture of the way titles are displayed on the author page and in the "Contents" section on a publication page, which would also include title series and alternate name information. For example, the line for the German translation of Iain M. Banks' Consider Phlebas could look like this:

• "Bedenke Phlebas [German] (1989) • [Culture • 1] • novel (trans. of Consider Phlebas 1987) [as by Iain Banks]"

Also, the list could be organized into a separate section per year. Then the year could be omitted per line:

• "Bedenke Phlebas [German] • [Culture • 1] • novel (trans. of Consider Phlebas 1987) [as by Iain Banks]"

Jens Hitspacebar 15:43, 20 May 2019 (EDT)

I would love to have a "All fiction in order" section - but if you add the collections and anthologies and magazines and reviews and interviews, it will become way too long to be useful. Maybe if you can chose which types you see? I assume that variants will remain filterable in that new view as well? Annie 16:49, 20 May 2019 (EDT)

We could have multiple different Chronological" pages, e.g.:

• "Chronological by Title Type", which would be the same as the current "Chronological" page
• "Strict Chronological" -- or whatever we decide to call it -- which list everything chronologically
• "Fiction (Chronological)", which would be limited to novels, short fiction and poems
• Possibly more, at which point we will probably need an intermediate menu page for different options

That being said, it looks like Jens may be proposing an additional layer of layout changes that would apply to all "Chronological" pages. Is my understanding correct? Ahasuerus 09:14, 21 May 2019 (EDT)

No, my layout proposal was for the "noncategorized-chronology" feature request only, not for the already existing chronological summary author bibliography. But now that you mention it, making the layout per title more consistent across all chronological pages could be a good idea. Or maybe not, I'm not sure yet. The existing chronological author summary bibliography puts the kind of variant at the beginning:

• Translation: Bedenke Phlebas [German] (1989) [as by Iain Banks]

which, if used the same way on the new "Strict Chronological" page as well, could probably make the list hard to read if many lines would begin with Translation: or Variant:. On the other hand, changing the existing chronological summary author bibliography to also use "(trans. of ...)" instead of Translation: doesn't make sense there because it's obvious that it's a "translation of" due to the way parent and variants are formatted using the bullet lists.

Jens Hitspacebar 16:15, 21 May 2019 (EDT)

Notifying Primary Verifiers

Where do we stand on notifying primary verifiers of changes to their pubs? The "My Changed Primary Verifications" will show what field changed, but not what the change was (both columns show the new value). When that feature was implemented, the community was still saying that primary verfiers should be informed of what the change was. Since the "Note to Moderator" is displayed on the change page, that became the more frequently used method over posting to the verifiers' talk pages (as the moderator's note is easier and remains with the edit history, that makes sense). There is the added complication that not all changes (title level, import/remove) are shown in the changed verifications report. We used to be pretty stringent on the notification requirement. However, a recent conversation has indicated this is no longer the case. What are people's opinions on this? It doesn't help editors for moderators to be inconsistent so we should update practice and/or our documentation based on the community's current position. -- JLaTondre (talk) 17:37, 20 May 2019 (EDT)

Let me try to address the technical side. FR 1237 says:

• Enhance "My Recently Changed Primary Verifications" to capture Import/Export and Remove Title submissions associated with the user's primary-verified publications.

Once it's been implemented, the report will become more comprehensive and more useful. Also, Roadmap 2017 says:

• Create a history of changes to primary-verified publications by storing a snapshot of the way each verified pub looked like right before it was changed.

It will take more time to do, but it will make the report much more useful.

Both features are close to the top of my list, but I need to code and deploy the new Amazon interface first. Fixer takes so much of my time these days that making at least some of what I do publicly available (and thus freeing up more development time) is my top priority at the moment. It's kind of a Catch-22: working on other features means that I can't work on the Amazon interface, which means that I have to spend most of my time working on Fixer, which means that I don't have time to work on other features... Ahasuerus 09:06, 21 May 2019 (EDT)

Yes, I wasn't looking for a technical solution (though that will be nice when it arrives), but rather how we deal with it before the full technical solution comes. I appreciate the enhancements you have been giving us. -- JLaTondre (talk) 19:43, 21 May 2019 (EDT)

I am sorry, I didn't mean to come across as defensive, but I guess I did. Too many balls in the air and not enough productive hours, especially these days :=\ Ahasuerus 21:31, 21 May 2019 (EDT)

I do not notify when I am doing what I consider house-keeping actions (moving identifiers, adding transliterations, editing any of those elements, fixing typos in Notes, fixing a broken HTML or a formatting issue, updating a title or an author name to comply with the capitalization policy and so on) - but I always add a moderator note (and if I am replacing, I tend to add the old and new value in that note). Outside of that, I do notify - one by one if needed; as a group when I am working on a bigger project (swapping a canonical name for example). I also try to notify when making a big change on a title level that impacts a PV'd publication - with a note of what changed to what so we can backtrack if needed. Import/export also causes a notification in my book. For any big changes (changing an author name, fixing page numbers and so on), I try to discuss before I do the change. Annie 19:55, 21 May 2019 (EDT)

Annie's approach may be a viable compromise solution, at least for now: minor changes are listed in the Moderator Note field while significant changes require a note on the Talk page. Approving moderators get to decide what qualifies as a "significant" change and make sure that the Moderator Note includes the "before" version of the data for "minor" changes (since it would be lost otherwise.) Ahasuerus 21:05, 21 May 2019 (EDT)

Edit: On second thought, we probably need to create at least some standards re: what is considered a "minor" edit. Annie's list -- typos, transliterations, HTML fixes, etc -- would be a good start. Ahasuerus 05:44, 22 May 2019 (EDT)

One of the problems when approving someone else's update is that if they do not add the note, you cannot add it. In a lot of cases, the record has at least one more thing that needs fixing but now with most of the identifiers out of the way, we start running into no-issues publications. If I see something else needing fixing, I would fix it and add both edits to the moderator note - so the PV does not wonder what happened. Any chance of making the moderator note editable during the approval? That will also help with the other two issues we have - the fact that the moderator note is visible (a moderator will be able to clean up the private data if any) and the inability to add anywhere any notes when you approve something that at first glance looks weird. Add the ability to add the change if the original submitter had forgotten and we will have a much better system. :) Annie 22:40, 21 May 2019 (EDT)

Letting approving moderators change Moderator Notes would be possible although non-trivial. However, I am not sure it would be optimal. There would be potential for confusion re: who wrote what.

It may be better to create a new field and call it something like "Approval Note" or "Note by the approving moderator". We could then display the Moderator Note and the Approval Note side by side on submission history pages. We may also want to change the name of the "Moderator Note" field to "Submission Note", "Comment about the Submission" or something similar to avoid confusion with the new field. Ahasuerus 05:57, 22 May 2019 (EDT)

Sounds great! "Submission note" is fine, I'd think: it's short and right on the head of the nail. Stonecreek 08:35, 22 May 2019 (EDT)

That works as well :) Annie 17:19, 22 May 2019 (EDT)

I also like "Submission Note" and "Approval Note". ···日本穣 · 投稿 · Talk to Nihonjoe 17:31, 23 May 2019 (EDT)

OK, FR 1281, Add an "Approval Note" field; change "Moderator Note" to "Submission Note", has been created. Ahasuerus 11:12, 24 May 2019 (EDT)

I've gone ahead and approved the submissions I had on hold over this. I'll encourage the submitter to add moderator notes in the future, but given the low level of participation in this discussion, I don't feel it is appropriate to reject their edits when other moderators have been approving them. Thanks. -- JLaTondre (talk) 19:07, 22 May 2019 (EDT)

Sorry to be late to the party. I have not been paying attention to what others do. I don't worry about moving of identifiers or correcting of minor deviations from capitalization standards, but otherwise I would have done as you did: put on hold and ask that the PV(s) be notified. But I also prefer to avoid rejecting and losing work, so sometimes I will notify and/or check with the PVs myself and accept and remind the submittor to follow notification protocol the next time. It is a bit tough to be strict about notifications with so many inactive PVs and PVs with special notification instructions. --MartyD 22:30, 23 May 2019 (EDT)

Proposed Software Changes

Ahasuerus: As these "describe again what is obvious" type of requests is constantly causing me useless trouble: Would you accept a solution to this issue as patch? My solution would be to add the existing values at the time of accepting a request to the stored request, so that a historic view could show the accepted difference situation (That still can differ from the actual value at submission time in case of inbetween changes). --Stoecker 13:39, 23 May 2019 (EDT)

[P.S. later] I'd really like to have such a feature, because even moderators create request without any comment like this http://www.isfdb.org/cgi-bin/view_submission.cgi?4270746 and even moderators can be totally wrong. In this example the price was correct before with €15.00. It seems it is now €18.00, but I payed €15.00 as stated on the website at that time (plus €11.00 delivery) for that one Foster story (and the remaining useless contents :-) --Stoecker 13:55, 23 May 2019 (EDT)

We had a fairly extensive discussion of this issue in early 2017. Let me copy-paste the relevant parts of my comments:

Re: "a snapshot of all OLD values". Unfortunately, it would require a significant effort. Granted, it would be easy to do for fields like "ISBN" and "Price". However, consider publishers. The way the database works, we store publisher numbers (1, 2, 3, etc) in publication records. Then, when we display a publication, we retrieve the name of the publisher, including its transliterated name(s), from other parts of the database and display them.

This works well when displaying current data. However, suppose we were to save a verified publication record as it existed prior to submission approval. We would store publisher ID 12345 in the saved record. Then, a few months later, publishers 12345 and 12346 are merged, so publisher 12345 no longer exists in the database. When the original verifier goes back to check this pub's history, there is no publisher 12345 to display. The same thing can happen to publication series, authors and titles. Actually, it can get even more complicated with authors and titles if we want to preserve the pseudonymous/VT/series relationships as they existed at the time.

The ultimate way to address this issue would be to build a snapshot of the then-current version of each about-to-be-changed Publication Web page and store it in a separate database location prior to submission approval. We could then have a list of snapshots for every publication and display them on demand.

We'll need to add a warning about potentially broken links, but the textual part should be very close to what the data looked like as of the time of the edit.

[comments about the impact of this change on disk space omitted since testing has shown it to be manageable, especially if we compress the data]

[related comment:] We started work on a "history" system -- basically a log of changed data -- in 2007. However, we quickly ran into the problems outlined above and more. I spent many man-hours trying to get it to work in the early 2010s, but eventually had to give up because the underlying approach was flawed. Ahasuerus 18:42, 10 January 2017 (UTC)

As far as I can tell, the summary above is still our best bet.

I guess the first step would be to rewrite the software that displays Publication pages. Currently it displays the data as it retrieves it from the database, mostly one line at a time. We would need to change the code so that it would retrieve all of the data first, then build the Web page and then display it. Once it has been done, the next step will be to modify the submission approval process. For primary-verified pages, it will invoke the part of the Publication display code that retrieves and formats the data. It will then compress it and file it in a "history" table within the database. [Need to run, may post more later]. Ahasuerus 15:05, 23 May 2019 (EDT)

Having a complete version history like Wikipedia including the chance to revert edits and show history would be fine, but from my current perspective this would be impossible to do. My proposal only includes a small subset. At the time of a change request acceptance I'd simply add for changed fields the old and new information, so that the changes can be displayed with proper information. Maybe there could even be a cleanup process to drop that additional space wasting information after some time. --Stoecker 15:29, 23 May 2019 (EDT)

I agree that implementing Wiki-like "version history", including the ability to revert changes, would be time-consuming and not feasible at this time.

However, what I outlined above is more modest in scope. The filing software would simply capture the HTML version of each about-to-be-changed primary-verified publication record. It would then compress the HTML "blob" and store it in some table. Each "blob" would be linked to its publication ID and its submission ID. That way the HTML would be accessible from View Submission pages as well as from Publication pages. That shouldn't be too difficult to implement, I hope. Ahasuerus 17:00, 23 May 2019 (EDT)

P.S. I have added the above to FR 1238 "Create an Edit History page for publications". Ahasuerus 11:04, 24 May 2019 (EDT)

This proposed HTML method is nothing where there is a chance of successful cooperation between you and me. It is much to complex to get it to work. --Stoecker 04:34, 30 May 2019 (EDT)

I am not entirely sure I am parsing the 2 sentences above correctly. Are you saying that capturing the body (HTML) of primary-verified Publication pages is a complex process? It would appear to be fairly straightforward since the software already builds all the requisite data when a Publication page is displayed. We just need to capture and store it. Could you please clarify the nature of the complexity? Ahasuerus 10:09, 30 May 2019 (EDT)

Still my proposal is an option - a submission is stored as XML in the database. My proposal is to extend this XML at the moment of accepting the submission and store the old values for changed elements (and only these) in the XML. That way a diff can not only show new, but also old values of a change. That will help a lot to verify a change even if it does not cover all possibilities. This adaption needs only small modifications at submission acceptance and display of a change request and not a major new functionality. And it does also not conflict with any future plans. --Stoecker 04:34, 30 May 2019 (EDT)

Magazine issue navigation

How easy would it be to implement the "Previous ← Current → Next" issues navigation as a database feature, as shown in the Notes field here? Could it use the existing database table information used to create the issue grid, so any missing issues would be skipped? And maybe indicate that one or more issues were skipped (if it can easily see that)? Just curious, as it would benefit browsers of the site so they wouldn't have to visit the issue grid between each issue. ···日本穣 · 投稿 · Talk to Nihonjoe 21:05, 20 May 2019 (EDT)

How would you determine "next/previous" in the cases where we do not have months in the date field? Finding missing issues will be even more problematic - some magazines have 2 issues in some months, some are irregular... and not all magazines have nice numbers on all their issues. Not a bad idea but just wanted to make sure we cover all the logistics... :) Annie 22:19, 20 May 2019 (EDT)

That's why I asked. There's some sort of sorting happening in the issue grid, so that's why I suggested that as a possible method, if it can be used for that. ···日本穣 · 投稿 · Talk to Nihonjoe 02:28, 21 May 2019 (EDT)

Let's consider Science Fiction Quarterly (UK), which has 6 issues whose months of publication are known and 4 issues whose months of publication are unknown. January 1952 through August 1953 should be doable since all of the issues have months. However, what should we do about December 1953 and later issues? A human can figure out that the sequence should be:

• December 1953
• #6
• #7
• February 1955
• #9
• #10

but it would be difficult to replicate the logic programmatically. Ahasuerus 07:14, 21 May 2019 (EDT)

To add to that: The grid does not really do well if it needs to stick more than one issue in the same cell (the no month one for example for a specific year) - it does its best but... it can get a bit wonky and it does look weird for some magazines. Which is not a problem on a grid (you see them all, the slightly off order is not that problematic) but becomes a problem on a straight list. Annie 16:06, 21 May 2019 (EDT)

Perhaps a way to turn it on for magazines that have all the issues entered? Or some way to enable it on a per-issue basis? It could then be left turned off for any that don't line up correctly. Or, program it to display the out of order issues according to which was entered into the system first? Just tossing out ideas. :) ···日本穣 · 投稿 · Talk to Nihonjoe 17:27, 23 May 2019 (EDT)

We could use publication record numbers to determine the order in which issues were entered. However, there is no guarantee that magazine issues are necessarily entered in the order of their publication.

Having said that, it may be possible to piggy back on FR 1202, "Add fields for magazine issue numbers to pub records":

• Add the following new fields for magazine issue numbers to publication records: Volume, Issue, Whole Issue. On the data entry side, we could have a single line for capturing this information, e.g.:
• Volume: [field] Issue: [field] Whole Issue: [field]

For example, the proposed "Whole Issue" field could support the "|" format, which we currently use for page numbers. The first part of the value would be displayed to our users. The part after the "|" character would be a "sorting" value used to create "Previous" and "Next" links. Ahasuerus 10:45, 24 May 2019 (EDT)

Change ID

Good evening. I am sure this is covered somewhere but cannot for the life of me find the article. How can one go about changing their username on this site? Thanks in advance. Zybahn 23:22, 28 May 2019 (EDT)

I am afraid the ability to change user names is not supported by the ISFDB software. Ahasuerus 00:03, 29 May 2019 (EDT)

Guerra eterna (The Forever War) - Joe Haldeman - broken link

Hi, there's a broken cover picture link to the Mondourania website here, where they've probably reorganised their site. As I don't read Italian, can I ask that someone else rectify this. I think the link should be as below, but I may be wrong.

Covers page here: http://www.mondourania.com/collezione/collezione%201-20/uraniacollezione1-20.htm

Relevant cover here: http://www.mondourania.com/collezione/collezione%201-20/uraniacollezione10.htm

Thanks. BanjoKev 16:38, 30 May 2019 (EDT)

Fixed. The problem was an extra space in the string we had - so chances are that was never a correct image to start with. :) Annie 18:04, 30 May 2019 (EDT)

Thanks Annie! BanjoKev 19:06, 30 May 2019 (EDT)

Project Gutenberg publication records

Concerning publication records for Project Gutenberg ebooks:

1. Do we discourage formal verification? Only a small share of our records are Primary Verified, now 4 of about 50 for year 1998 and none at all 2017 to 2019.

2. Do we discourage use of the Cover field, in reference to the first image that is displayed at Gutenberg.org in HTML format (and probably some others)? Commonly the first image displayed, if any, is an image of the front book cover or front dust jacket. For instance, Ebook #25609 at Gutenberg (HTML format) and at ISFDB (COVERART credited). We now credit no cover artists for year 1998 and only four 2017 to 2019. (We link no cover images for those years.)

3. We have some old-fashioned records whose Notes contain an HTML link to the reported Ebook (not a different one) at Gutenberg.org. This one also provides an explicit list of available formats P291428. (That one is Primary Verified by User:DESiegel60, retired.) Do we recommend either retaining or deleting such notes?

--Pwendt|talk 16:54, 5 June 2019 (EDT)

As I have been the one keeping up with entering new PG publications, I will give you my rationale on how I've been entering them.

1. We don't discourage formal verification. I have not been verifying them as PG's are a special beast in my opinion. They are regularly updated with corrections, etc. and I don't see the point of creating different records for minor edits (we have had this debate about ebooks before & haven't really come to a decision on how to handle). There is also the issue that PG ebooks are available in a variety of formats. So what is actually meant by verification? One format checked? All formats checked? Anyone can use the PG link to view the ebooks so verification hasn't been important in my mind. I'm basing my entries on the HTML versions.
2. We don't discourage covers. However, I personally don't enter the image or the cover credit. My rationale is that PG ebooks are available in a variety of formats; some of which have "cover" images and some of which don't. But then again, I have been adding the interior art even though not all formats have that either. So I'm not necessarily consistent in my approach. ;-)
3. As we have the built in link, there is little point to the manual link. Likewise, PG formats are pretty standard (though possibly not all older ones have every format) so little point in the format note. There is no harm to them and as long as they are valid (ex. formats haven't changed since the note was written), there is no point in removing. In the case of the one you link to, I don't believe the Plucker format is available anymore so it should probably be changed to reflect that.

Those are just my opinions. You are free to verify, add covers, etc. as you see fit. -- JLaTondre (talk) 21:38, 5 June 2019 (EDT)

Star Wars (and other franchises)

I noticed someone recently took the series Star wars: Nonfiction out of the parent series Star Wars universe with comment "As this lists titles about the Star Wars franchise, it is not a part of the Star Wars universe". This may be true but imo all "Star Wars" should be consolidated under one parent series. Such a separation would also affect any franchise like Star Trek, Doctor Who, etc. And where to set boundaries: Is a character encyclopedia not fictionous enough to be part of a "Universe"?

As a user i'd like to find all of a "Universe" on one page, not divided to a few.

What do you think? WernerWelo 13:46, 6 June 2019 (EDT)

Hi, Werner! It was me who made that differentiation. The reasoning behind this is that if a title is truly in-universe it is fiction (not nonfiction). It's misleading to have a title like this in the same category as this one, where in fact fictional stories about spaceships are told. Stonecreek 14:03, 6 June 2019 (EDT)

What about Star Wars chronologies and 'nonfiction' ? ../Doug H 14:23, 6 June 2019 (EDT)

These are in-universe and as such part of the franchise. Stonecreek 14:43, 6 June 2019 (EDT)

I‘m all for differentation, however all of Star Wars should imo be combined under one parent series wether its fiction, fictionous nonfiction or real nonfiction. Has not to be the „Star Wars Universe“, maybe a parent like „Star Wars Franchise“ or similar could do it. Under such a series the differentation could be made in universe and others. Welo 14:53, 6 June 2019 (EDT)

I see no issues with having Star Wars: Nonfiction having the parent Star Wars universe because the nonfiction is about the Star Wars universe. Since it's all contained within a series that specifically states it's nonfiction, I doubt anyone will ever get confused. ···日本穣 · 投稿 · Talk to Nihonjoe 18:13, 6 June 2019 (EDT)

I have been struggling with this issue for years. Some non-fiction books are "in universe" and some are not. Some are written or co-written by the author(s) (like Roger Zelazny's Visual Guide to Castle Amber), some are "authorized" and some are completely unauthorized. I am yet to come up with a hard and fast rule to determine when a non-fiction series should be entered as a subseries of the main universe. Ahasuerus 00:12, 7 June 2019 (EDT)

I'd say distanced pieces like "Star Wars on Trial" or movie compendiums that rely more on the filmical aspects shouldn't be part of a universe. 'Encyclopedias' of characters, planets and space-ships should be imo. Stonecreek 02:59, 7 June 2019 (EDT)

The problem is, that makes it very subjective as to what is included. One person may think it clearly should be, and another doesn't. I think it's best to simply lump all non-fiction together in a non-fiction series, and then put it under the main umbrella series for the entire series. That makes it easy for everyone to remember, and we'll not have to debate it every time something new comes up. Straightforward solutions work best, IMHO. ···日本穣 · 投稿 · Talk to Nihonjoe 13:19, 7 June 2019 (EDT)

(unindent) Ironically I came here looking for a link about a book entitled The Land of Unreason by Fletcher Pratt and L Sprague de Camp. —The preceding unsigned comment was added by Chris humel (talk • contribs) . 18:31, 6 June 2019 (EDT)

How is it ironic? Also, the book you're looking for is here: 18960. ···日本穣 · 投稿 · Talk to Nihonjoe 22:24, 6 June 2019 (EDT)

Publication pages - tweaks behind the scenes

I have installed a minor patch to change the way Publication pages are built by the ISFDB software. I expect a number of additional patches to be installed in the near future. Once everything has been completed, these changes will facilitate the process of creating "history" Web pages for primary-verified publications.

These changes should not affect the "look and feel" of Publication pages. If you come across any issues, please post your findings here. Ahasuerus 16:09, 14 June 2019 (EDT)

There's a tiny glitch with "(View All Issues)(View Issue Grid)" in the "Publication:" line on a magazine's publication page. The pub's title and these two links are not separated by space any more. Example: Interzone, #263: "Interzone, #263 March-April 2016(View All Issues)(View Issue Grid)". If I'm not completely mistaken there used to be some space between these three elements, which made it easier to read, and it looks like this change has been introduced with one of the recent commits, probably rev. 435. Jens Hitspacebar 04:10, 15 June 2019 (EDT)

Good catch; fixed. The latest patch also fixed links for magazines issues whose EDITOR records are not in a series (not that we are supposed to have any.) Ahasuerus 10:30, 15 June 2019 (EDT)

"Transliterated values" fixes

There was a bug with the way "transliterated values" were being displayed on some pages. I have installed a patch to address the issue. If you see anything unusual, please let me know. Ahasuerus 15:45, 16 June 2019 (EDT)

Pseudonym display

Would it be too difficult to change the display of titles with multiple authors, and where one or more has a pseudonym, to reduce the space and make it easier to read. Right now, it's something like this:

Title by Author One and Author Two and Author Three and Author Four and Author Five and Author Six and Author Seven and Author Eight and Author Nine and Author Ten and Author Eleven and Author Twelve and Author Thirteen and Author Fourteen and Author Fifteen and Author Sixteen and Author Seventeen and Author Eighteen and Author Nineteen and Author Twenty [only as by Author One and Author Two and Author Three and Author Four and Author Five and Author Six and Author Seven and Alt Eight and Author Nine and Author Ten and Author Eleven and Author Twelve and Author Thirteen and Author Fourteen and Alt Fifteen and Author Sixteen and Author Seventeen and Author Eighteen and Author Nineteen and Author Twenty]

If it could be something like this, it would make viewing the information much simpler and much easier to read:

Title by Author One and Author Two and Author Three and Author Four and Author Five and Author Six and Author Seven and Author Eight (as by Alt Eight) and Author Nine and Author Ten and Author Eleven and Author Twelve and Author Thirteen and Author Fourteen and Author Fifteen and Author Sixteen and Author Seventeen (as by Alt Seventeen) and Author Eighteen and Author Nineteen and Author Twenty

Would this be reasonably easy to do? ···日本穣 · 投稿 · Talk to Nihonjoe 23:45, 18 June 2019 (EDT)

I agree that it would be desirable and I looked into it at one point.

What I found was that it would be possible to do if "variant-specific authors" and "alternative author names" -- which are entered and stored in the database separately -- matched in predictable ways. Granted, in many cases they do, e.g. Robert Heinlein is always an alternative name for Robert A. Heinlein, Maxine Reynolds is always an alternate name for Mack Reynolds, etc. However, we also have many collective pseudonyms and house names that overlap unpredictably. For example, consider The Whispering Gorilla series. The Whispering Gorilla omnibus of 3 novels is currently listed as follows:

• by David V. Reed and unknown and Don Wilcox [only as by Alexander Blade and David V. Reed and Don Wilcox]

"Alexander Blade" was a house name used by both "unknown" and by Don Wilcox, so one might think that this line could be rearranged as follows:

• by David V. Reed (as Alexander Blade) and unknown (as Alexander Blade) and Don Wilcox

However, if you examine the publication record, you will note that David V. Reed is credited explicitly, so the line above would be in error.

The only way to get it to work correctly for all records (that I could think of) would be to change the way we capture variant authors. Instead of simply entering the canonical names on the "Make Variant" Web page, we would have to specify which canonical name each alternate name corresponded to. It would require extensive changes to the way the software works and a subsequent review/revamp of thousands upon thousands of variants. Ahasuerus 11:19, 19 June 2019 (EDT)

Displaying the parent title's authors on the variant title's page

The recent discussion of Dates on COVERART variants has resulted in a software change. Variant title pages whose parent title's date is different from the variant's now display the parent's date on the "Variant Title of" line. For example, "The Skull" by Vincent van Gogh displays the following information:

• Author: Vincent van Gogh
• Date: 2011-00-00
• Variant Title of: Kop van een skelet met brandende sigaret (1886) (by Vincent van Gogh )

Note the bolded date.

While working on this issue, I noticed something that we may want to address. As you can see above, the "Variant Title of" line includes the authors associated with the parent title. The information is useful if they are different from the variant title's authors, but we display the parent's authors even when they are the same.

The "Contents" section of Publication pages only displays the parent title's authors if they are different from the authors of the variant title. It seems to make sense to change the Title page to do the same.

Are there scenarios that would be adversely affected by the proposed change? Ahasuerus 20:53, 25 June 2019 (EDT)

The change has been made as per FR 1287. In addition, the "Variant Title of" line has been changed to display the language of the parent title if it's different from the language of the VT. Ahasuerus 19:05, 29 June 2019 (EDT)

Advanced Author Search - Author Image URL

FR 1286, "Add Author Image URL to the Advanced Author Search", has been implemented. Ahasuerus 19:31, 29 June 2019 (EDT)