ISFDB:Moderator noticeboard/Archive 11

This is an archive page for the Moderator noticeboard. Please do not edit the contents. To start a new discussion, please click here.
This archive includes discussions from April - December 2011.

Archive Quick Links
Archives of old discussions from the Moderator noticeboard.

---

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

---

Expanded archive listing

Do we accept the copyright date as publication date?

We have a new editor, who has started using the copyright date (in this case Google's Catalog of Copyright Entries to pinpoint the publication date of a title's first edition (see here). I thought we agreed after this discussion not to use the copyright date. --Willem H. 11:03, 21 April 2011 (UTC)

Agreed, we don't use it for title or publication date. The Copyright date might be of some interest in Notes though. BLongley 14:10, 21 April 2011 (UTC)

The Copyright Catalog you've linked to only shows the date the book is registered with the copyright office. It's pretty clear those dates are not the publication date. We stick with more dependable secondary sources (like Reginald and Tuck) for these early books, even if they only give the year of publication. Months are nice, but who really needs to know the actual day unless you're a scholar? Mhhutchins 15:26, 21 April 2011 (UTC)

Sorry, accepting this one was my fault. I didn't realize that listing was giving a copyright registration date. I took the comment in the note at face value and only checked that the date matched.... --MartyD 01:56, 22 April 2011 (UTC)

Fixer and time travel

Now that June 2011 has been submitted, Fixer is gong back to what he was doing a couple of weeks ago: identifying "non-trivial" publications from the last 10 years one month at a time. All of 2009 as well as September-December 2008 have been submitted. There are roughly 100-130 "non-trivial" ISBNs per month in late 2006-late 2008, but then the volume drops off to a few dozen per month. At the rate we are going, we should be able to catch up by early summer. Ahasuerus 03:11, 24 April 2011 (UTC)

Additional cleanup scripts are now available

Additional cleanup scripts are now available, but it looks like one of them, "Find Stray Publication Authors", may generate false positives under certain circumstances, e.g. when an Anthology is embedded in an Omnibus. As you find these errors, please post them here so that we could fix the code. Thanks! Ahasuerus 03:14, 24 April 2011 (UTC)

I've already come across another bug. When a novel is in an anthology, a false positive is generated. They're marked "NOVEL IN ANTHOLOGY BUG" on the Stray Authors list. Mhhutchins 04:08, 24 April 2011 (UTC)

The script has been fixed -- or, to be more nearly precise, the problems identified above have been fixed. There may be other issues, but hopefully they will be rare. Ahasuerus 06:15, 25 April 2011 (UTC)

Sorry about the original problems with Stray Authors. The latest script results are here, ISFDB:Stray_Authors3 for those that want to treat it as a project. Once we've fixed those, hopefully this can be an occasional script run once a week or so for new problems. Having said that, fixing them is a bit of a chore at times: Unmerge/Search/Remerge is a pain for instance. So I would like feedback not only on how the script works, but how painful the fixing is - I can't promise to improve the fixing, but at worst I can help log the Feature Request(s). BLongley 20:35, 25 April 2011 (UTC)

Bill Drussaï

Is anyone familiar with this author and about the tests on his summary page. They're showing up on the stray review and interview records and I want to clear that list. Thanks. Mhhutchins 19:38, 25 April 2011 (UTC)

My fault, should have cleaned them up by now. See ISFDB:Help_desk/archives/archive_03#Entering_a_small_.22i.22_with_a_diaeresis for an explanation of why they were created. BLongley 20:09, 25 April 2011 (UTC)

Auto-generated Unmerge submissions?

Came across a couple of Vonnegut German editions that had been improperly entered under the 'Jr.' title. Unmerged them, then re-merged with the 'no-Jr.' title. No problem..... but, in each case a new submission was created [after the re-merge] for another unmerge. The first one I thought was just a fluke, so rejected it. When the second one showed up I accepted it to see what would happen. In both cases the records showed as I wanted them to, but that doesn't mean there isn't something I'm just not seeing, lurking, as it were. Could some of the new changes be causing this? And what is the correct procedure [if there is one]? --~ Bill, Bluesman 15:58, 27 April 2011 (UTC)

Actually the second time unmerged. It also put two records into the contents of the German edition. Managed to fix this [had to remove one content and after that a second title-remove submission appeared on its own] but don't know what's causing these submissions to appear. --~ Bill, Bluesman 16:11, 27 April 2011 (UTC)

And the script for these oddball submissions is different from the normal ones, different windows/format. --~ Bill, Bluesman 16:13, 27 April 2011 (UTC)

Which title or titles are this? I've just been working on some Vonnegut myself, so it might just be an editor clash if you were working on "Slapstick". Although having said that, I've also seen some Novels with both canonical title and variant in them, so there is quite possibly a bug. BLongley 16:31, 27 April 2011 (UTC)

First one was Timequake, second was A Man Without a Country. And now I tried to fix Hocus Pocus [there's a hungarian title put under the Jr.] and did the unmerge but now the Hungarian title won't even appear in an Advanced Search, so I can't re-merge. I think the 'ó's in the title are messing something up.--~ Bill, Bluesman 16:45, 27 April 2011 (UTC)

Fixed this one by changing the new Title record to the English, then merging. --~ Bill, Bluesman 17:03, 27 April 2011 (UTC)

(Before edit conflict) Bill Bluesman, after the unmerge did you correct the author credit of the newly created title record or did you merge them with the "non-Jr" records and reconcile the difference in author credit? The system retains the author credit of the original title record regardless of how the pub is credited, even though it matches exactly the title of the pub record in the title field of the new title record. I've asked Bill Longley if he could work on the software so that the newly created title record of an unmerged pub will match exactly the title and author fields of the pub. That should save us at least one extra submission when correcting pub record/title record author mismatches like this. Mhhutchins 16:53, 27 April 2011 (UTC)

The change to keep the Publication Authors rather than use Title authors has been submitted for testing. That should help with a lot of pseudonym errors. BLongley 15:58, 28 April 2011 (UTC)

I reconciled them during the merge. How does that account for extra submissions appearing in the queue that I didn't put there? --~ Bill, Bluesman 17:03, 27 April 2011 (UTC)

Now that I can't explain. Mhhutchins 17:09, 27 April 2011 (UTC)

If you see another of these extra submissions, hold it and let me know which it is and I'll see if I can figure it out. BLongley 16:24, 28 April 2011 (UTC)

Will do, though there were just the three, and none since... damned Gremlins.... --~ Bill, Bluesman 02:36, 30 April 2011 (UTC)

New site for linking images

I've received a copy of the response to the letter Dirk Broer sent to www.deboekenplank.nl for permission to link to their images. If Google Translate is correct, we now have that permission. Is there anyone who can change the software to credit the website for images drawn from their server? I'll update the permissions list. Mhhutchins 16:58, 27 April 2011 (UTC)

Change submitted for testing. BLongley 11:29, 28 April 2011 (UTC)

Tested and installed. Ahasuerus 03:37, 29 April 2011 (UTC)

Current Best Practices for Entering Translated Shortfiction Collections

I understand that Foreign Language support is still evolving at the moment, so I thought I'd ask about this. I have several collections of shortfiction that were originally published in English, then translated. In these cases the contents of the translated collections match the English originals, so I know I'm adding the collection pub records to the original English title record.

Having looked at examples where the contents have been cataloged, the title of the publication is represented in translation, but the title for each piece of shortfiction is listed in English, not with the translated title. I'm imagining that this is a reflection of the manner in which the DB is structured. In some cases I've seen the translated titles entered as part of the pub notes, but this is not always the case.

I've seen many mentions in the wiki that improved Foreign Language support is being worked on, but am not sure if anything is in the works to address this issue in particular or how soon those changes will be appearing, so I'm curious as to how I should proceed with things as they stand currently. Albinoflea 02:17, 1 May 2011 (UTC)

Hmmm... 3 days here without a response... is this the topic of which one does not speak? Albinoflea 21:23, 4 May 2011 (UTC)

I suspect it is not being actively worked on at the moment - I seem to be the only person actively submitting software updates, and Ahasuerus is mostly only coding improvements to my improvements. MY big problem is that I don't know enough about other languages to design any changes - and I try and keep my changes small. If editors with more language skills than me can suggest small software improvements, I can try them. Adding a "language" category against a title of any kind would be one small step. Displaying such would be another. Changing User Preferences to show only titles in selected languages would be another. Adding "Translator" is a bit bigger - but maybe that can be broken down into "Translators of Novels" first, and anthologies or collections where each content title can have a different translator can come later. Mass Update scripts are possible, but take a lot of persuasion to get past our tester - e.g. if we get a "language" field I could write a script that converts all titles starting "La", "Le ", "Les ", "Un " or "Une " to "French". If such a dangerous script doesn't pass, I can post such as a separate project that people can work on title-by-title. Any Software Designers with Non-English-Language skills volunteering? BLongley 22:37, 4 May 2011 (UTC)

To get back to the original question - if you add details in pub notes, they won't be lost. But I would understand if you waited for better software support. BLongley 22:37, 4 May 2011 (UTC)

As probably one of the main offenders in this field (for french titles), here's how I proceed :
1) If the collection is a strict (or quite close) translation of an english-language one, I just "add a publication" to the title, with the french title at the pub level. I import next the contents and verify them.
2) If it's an original collection, I start from scratch (so creating a new title in french). For the contents, as I'm more on the bibliographic side, I enter the original titles of the works (I've even done this with german or italian titles translated in french), there follows a long merging phase where each short story must be individually merged; sometimes easily sometimes with more difficulty in case of pseudonyms, missing or wrong data or when the original title given is itself a vt. I confess that I don't have the courage to record in the notes the french titles of each work, leaving this task for a time when foreign-language will be implemented. Hauck 07:33, 5 May 2011 (UTC)

If it's a close approximation to the original publication it would be better to "clone" the existing pub that has the closest contents. That way you don't have to re-enter the contents, and that saves you from merging the new title records with the ones already in the database. Then you could easily add or remove any titles that don't match the two pubs. Mhhutchins 18:37, 5 May 2011 (UTC)

Yes I've also proceeded in this way in precise cases (e. g. Cordwainer Smith). Hauck 18:43, 5 May 2011 (UTC)

Thanks for the input; I submitted my first pub today. I realized I already had the individual titles in a spreadsheet, so it was not much work to transform them into HTML and add them to the notes, so I figured I might as well. I took Mhhutchins's advice and cloned an existing pub to keep the merging to a minimum.

It does occur to me that rather than trying to write a script that would attempt to derive language from words in the title field, you could write a script that utilized the group identifier of the ISBN as a first pass, then let volunteers massage the resulting lists. That would grab a good portion of the last 40 years worth of non-magazine titles, which would be a good start. (ISSNs for magazines do not include location information, and don't appear to be consistently captured in the database anyhow.) Have you given much thought as to how the language information would be stored in the database and entered by the user?

I wish I had a bit more programming experience, but for the most part I just write CSS and administer CMS installations. (If the former ever would come in handy, please let me know.) If you're looking for help brainstorming implementation and (eventually) testing language support I'd be willing to set up a local install and lend a hand. Albinoflea 05:00, 9 May 2011 (UTC)

Please do - as you can see, there's a bit of a backlog on the testing front at present and the more eyes looking at any particular change the better. BLongley 15:05, 10 May 2011 (UTC)

OK. Can I assume that the installation instructions on the wiki are current/accurate? Albinoflea 00:20, 11 May 2011 (UTC)

I believe so, although I've only used the Windows version. BLongley 13:20, 11 May 2011 (UTC)

I've also just submitted some changes that allow you to enter a Language against a title. And see it (while editing it). And it should be preserved throughout other edits. It may sound like a small, useless change but it does affect quite a few modules, so if people with more language skills than me (and I don't mean Python/SQL) would review it then we might move a little bit closer to more useful language support. BLongley 01:46, 12 May 2011 (UTC)

OK, I'm working off a PC most of the time so I'll give the local install a shot this weekend.

The change you describe sounds like a necessary first step; we've got to start somewhere, after all. Thanks, Albinoflea 23:29, 12 May 2011 (UTC)

I based the change on what ISFDB:Proposed_Design_Changes#Enhanced_support_for_other_languages says - Ahasuerus has done the first step, this is the second. But that piece is so old (and the section below seems to suggest all sorts of other things) that we may need to revisit it. How's the local install going? BLongley 14:54, 15 May 2011 (UTC)

(Sorry, just noticed this.) I've got the files downloaded, but want to finish up a job for a client before I start, because I've already got Apache and MySQL installed (XAMPP) and I suspect I may need to reconfigure a few things. Hopefully I'll get to it this weekend; will keep you posted. Albinoflea 02:13, 18 May 2011 (UTC)

Deletion request

This seems to be non-existent. There is no trace of it in any standard place: amazon, abebooks, worldcat, blic, etc. Variety of Google searches also seem to give isfdb record only. I am posting the deletion request here rather than in the standard place because the pub looks quite legit otherwise with isbn, page count and even some notes. Thanks! P-Brane 11:36, 9 May 2011 (UTC).

For what it's worth, here's another source, albeit the author's website. Perhaps more reliable, the SFX website, which sponsored the contest. It appears the "book" appeared as a special issue of the magazine. Gollancz was a co-sponsor, which may account for it's name in the publisher field, although it probably shouldn't be. We should do more research before deleting the record. Mhhutchins 12:59, 9 May 2011 (UTC)

More proof of its existence: OCLC. I'll change the record to show the publisher as SFX based on this record. Mhhutchins 13:12, 9 May 2011 (UTC)

My questioning of the editor who originally submitted the record, but got no response. Sometimes all it takes is a little more research. Mhhutchins 13:17, 9 May 2011 (UTC)

Thanks! The source of OCLC record seem to be this one at Oxford libraries. It has additional information: "Free with SFX issue 148". Publisher is listed as SFX and Future Publishing. Oxford library also has in its catalogue two more issues of Pulp Idol (for 2007 and 2008). P-Brane 01:54, 10 May 2011 (UTC).

SFX did give away several books with the magazine over the years, not just "Pulp Idol". As they were only acquired that way they're quite unlikely to have ended up in libraries. I also have several other paperbacks without prices and marked "not for resale" that were given away with other magazines - they're valid entries for ISFDB, but are probably harder to trace than other editions unless you have the primary. And this reminds me, I have CD-ROM editions of SFX that I must get around to entering. I wonder if they still work on newer PCs? BLongley 15:13, 10 May 2011 (UTC)

So, what do we enter in the price field for pub in question? P-Brane 02:06, 11 May 2011 (UTC).

I enter free items as "$0", but I'm not sure if there's documentation regarding this issue. Of course, in the UK, I suppose it would be "£0". Mhhutchins 02:42, 11 May 2011 (UTC)

There could be an argument made for "$0.00" and "£0.00". And pre-decimal British pubs could be "-/-" (no shillings, no pence). At the moment, it's pretty much a cosmetic decision - I've tried to use our price data in SQL queries before, and it's a bit of a pain. But it would take a lot of work to enforce standard "Currency" and "Prices" - and many books, especially the ones I like, can have 7 or 8 prices on them. So there isn't even a Feature Request to improve such - I think anyone trying to take it on would go insane. (Or choosing to work on it would demonstrate that they already are.) Prices are useful for distinguishing printings, and maybe are some help in dating books, but really it's not our most important attribute of a publication. BLongley 15:09, 15 May 2011 (UTC)

Remember? Remember?

I have a collection of essays by Charles Beaumont called "Remember? Remember?". Many of the essays originally appeared in Playboy. Not all of the essays are genre-related, but some of them are either partially or wholly about science fiction or horror. How should this book be listed? As non-fiction? And should enter every essay, or only the essays that are genre-related? Do the partially genre-related essays get entered? (For example, there is an essay about the pulps.) Thanks! Sunnikay 02:57, 11 May 2011 (UTC)

Enter the book as NONFICTION (on the Home Page choose "Add New Nonfiction"). All spec-fic related essays should be entered. Record in the note field that only spec-fic related essays have been entered as contents. If it's borderline please create a record. The essay about the pulps should definitely be entered. Mhhutchins 03:08, 11 May 2011 (UTC)

Fixer - July 2011

There are about 40 2008 books left in Fixer's queue. Once they have been submitted and processed, I'll start July 2011. Ahasuerus 23:33, 12 May 2011 (UTC)

The rest of 2008 has been submitted. Ahasuerus 15:17, 13 May 2011 (UTC)

And are now (mostly) processed. Please forgive the lack of enthusiasm, I think most Moderators can think of a dozen better things to do. (And if they can't, I can probably provide a better mini-project!) BLongley 00:22, 14 May 2011 (UTC)

Thanks! And yes, January 2008 was a particularly poor month for some reason. On the flip side, it means that we already had the important ISBNs published that month, so there was nothing for Fixer to do but bottom-feed. Ahasuerus 00:24, 14 May 2011 (UTC)

Black Glass, Reader's Guide

I've submitted a content listing for Black Glass: Short Fictions that includes the "Reader's Guide" at the back of the book. The Reader's Guide consists of three parts: An interview with the author, a collection of "Reading Group Questions", and a longer than normal "About the Author". I've included each of those items as separate content items. Since I've included them, the content item for "A Reader's Guide" includes nothing of its own -- it's just a title for that collection of 3 other items. However, it does have its own title page. So while I'm not absolutely convinced that it deserves its own listing, it seems that including all 4 items is probably the right solution. But I'm willing to be corrected if need be. (I still have some cleaning up to do in the notes.) Chavey 11:45, 13 May 2011 (UTC)

Ditto for Sarah Canary Chavey 12:39, 13 May 2011 (UTC)

Either way is fine. It's up to you. The database has no way of handling a group title. I wish it did and it's been discussed before, but there's been no progress in that direction. Mhhutchins 13:33, 13 May 2011 (UTC)

I accepted the update to Black Glass. You'll need to add those extra pages to the record's page count. Mhhutchins 13:34, 13 May 2011 (UTC)

Done. Chavey 21:38, 14 May 2011 (UTC)

Moderator Notes and Fixer

Fixer has been upgraded to take advantage of Moderator Notes. Future submissions will list "browse nodes" and subject in the Mod Notes section, so moderators won't have to remove them from the Notes field any more. There will be cases when the approving moderator will want to remove something from the Notes text, e.g. "Edition statement: reprint", though. Ahasuerus 21:13, 14 May 2011 (UTC)

I've seen the first few, and like it. My only common edit for stuff "above the line" is usually for Large Print Editions, where edition and format state the same thing. Stuff "below the line" might be useful when determining inclusion, but is usually redundant after moderation, unless you want to add a publication tag like "Paranormal Romance - Avoid!". (Actually, that's not quite what the tag says, I just feel it probably ought to, if we don't want to add "Vampire Shagger", "Werewolf Shagger", "Shapeshifter Shagger" categories. ;-) ) BLongley 22:01, 14 May 2011 (UTC)

Fang bangers? Chavey 01:35, 15 May 2011 (UTC)

Fixer - July 2011, Part 2

Fixer is almost half way through the July catch, although it's hard to be sure because records are processed based on ISBNs. Publishers purchase big blocks of ISBNs, so sometimes Fixer runs into a few dozen comic books in a row or SF-flavored calendars. I try to eliminate them prior to submission, but you never know. (This also explains why paranormal romances travel in packs.)

Anyway, I have created a bunch of submissions and will process them tomorrow while I am away from the development server. Ahasuerus 05:46, 18 May 2011 (UTC)

I think I'll leave them to you. "Magic Tree House #42" does not appeal. :-( BLongley 16:59, 18 May 2011 (UTC)

US done, UK in progress. Ahasuerus 04:23, 22 May 2011 (UTC)

The rest of July has been submitted. Ahasuerus 04:09, 27 May 2011 (UTC)

The Search for Philip K. Dick

Hello, to whom it may concern: Actually, I am looking for the author of the note on this title, but if somebody else can take a look at the issue, this would be appreciated as well. Judging from the introduction I think that it'd be right to merge the three individual pubs, because it is essentially the same book, after all. All three comprise mostly (pp. 21-263 of the Tachyon pub) on the memories of Anne R. Dick which were written in 1984. There were revisions of the text in 1995 and 2009 but I tend to say that it is not enough to mark them as different titles. Also, there was new material added in 2009 but that seems to comprise only of new introductions and afterwords. The differences should be mentioned in the notes, though. Stonecreek 17:29, 19 May 2011 (UTC)

I'd make an educated guess that Marc Kupper is the person that added the note. Only Ahasuerus or Al von Ruff can check the submission table to be sure though. BLongley 18:32, 19 May 2011 (UTC)

We are pretty inconsistent about how to deal with "new editions" though. Sometimes we keep them as separate titles, sometimes we make later editions variants of the first, sometimes they all get lumped together under one title. I guess it's one of those woolly areas where the number/quantity/quality of differences might persuade an editor one way or another. Having read none of them, I don't really have a view in this case. BLongley 18:32, 19 May 2011 (UTC)

We have no method of associating titles other than through the variant process. And variants should only be created based on a change in title or author, NOT text. I agree with the person who wrote the note, but possibly for a different reason. The Dick titles should not be merged, but made into variants only because of the title change, NOT because of the revised text. Mhhutchins 19:36, 19 May 2011 (UTC)

I slightly disagree, a BIG change in contents or text is worth a separate title in my opinion. HOW big that needs to be is one of those fuzzy areas. How do people want to adjust Paul F. Ford for instance? BLongley 21:55, 19 May 2011 (UTC)

Then we don't disagree (I think). My point was that variations in text are not the basis for either a merge or a variant. There should be a new title record when a significant portion of a book is different from another one with the same title and credited to the same author. Unfortunately, the system doesn't have the ability to associate these titles and that's what bugs me. That leaves merging as the only option. In the case that was presented to us (the Dick title), the variants should be created because of the difference in title. The variation in text is not a factor here. Those Paul Ford titles should be varianted because of the change of title, again, not because of a change in text. Now if the subtitles are entered here aren't actually stated in the book itself, they're wrong and should be corrected with proper notes indicating the editions. Mhhutchins 05:07, 20 May 2011 (UTC)

Another thought, just after I posted the above - we could use Title or Publication Series to differentiate editions? BLongley 22:01, 19 May 2011 (UTC)

A stopgap measure. Nevertheless, I've seen other uses of series when chewing gum, wire and spit weren't available. Mhhutchins 05:07, 20 May 2011 (UTC)

We're a creative lot, aren't we? :-) Still, we're not much closer to printing number support than a few years ago, let alone "edition" support, so stopgaps will have to do for a while longer yet. I can only code so fast. BLongley 15:36, 20 May 2011 (UTC)

Yes, now I see that it would have been right to submit the various editions as variant titles and I will do so tomorrow. Thanks. Stonecreek 10:39, 20 May 2011 (UTC)

Sorry on the late visit to this thread. Yes, I wrote the original title note note plus started Author:Anne R. Dick#The Search for Philip K. Dick to document my research at the time on what editions may exist for this work. The reason I had had left them as separate titles is that this is the only book-length work by Anne R. Dick. Had the three titles been scattered throughout a large bibliography I would have bundled them with a variant title or series.

I believe the VT that was added will cause confusion. Specifically, it gives the appearance that the 1995 and 2010 editions are identical or at least leaves someone wondering what the relationships are for the "Revised with New Material, 2009" edition vs. the other two. My local library has copies of both the 1995 and 2010 editions and so back in November-2010 I added to the list of projects a note to check out both copies and to compare them. That note is now mid way up or down the pile. --Marc Kupper|talk 10:19, 24 May 2011 (UTC)

Data from library catalogs

I have been slowly building up a collection of publicly available catalog records. Some of them come from complete dumps of library catalogs, others come from querying online catalogs about SF subjects.

Over the last few years I have accumulated millions of mostly overlapping records (about 40 million, give or take) and I have taught Fixer how to parse them. The next logical step would be to convert the parsed data into submissions, add "Data from XYZ library" to Notes and we should be in business. Or so I thought when I started this project.

It turns out that unlike Amazon.com, which, for all it faults, has a database with fields like "Title", "Author", "ISBN", etc, library catalogs are, well, different. Most of them use some flavor of the MARC standard, which also supports fields, e.g. field 245 is "Title Statement". However, this field has a complex definition and supports numerous sub-fields which sometimes interact in non-trivial ways. When you add the fact that different catalogers use different rules for populating these sub-fields, you end up with a rather messy picture.

As programmers have discovered, sometimes it's easier to analyze the existing data and look for patterns of data use (and abuse) than to come up with a priori rules. Unfortunately, the data that I have accumulated comes from hundreds of catalogs, each with its own quirks, so if I were to go that route, it would take years of uninterrupted work to come up with something usable.

At this point I am thinking that the easiest way to handle the situation is to create submissions based on minimal MARC information and add any complexities to Moderator Notes. It will require more work on the approval side, but at least it will get us started. Would that be OK? Ahasuerus 04:03, 20 May 2011 (UTC)

So long as you don't swamp us with all 40 million at once, I think it would be a nice change for Mods. BLongley 15:58, 20 May 2011 (UTC)

Most of these records are not SF-related, so one of the challenges will be finding the ones that are. There are multiple ways to do this, e.g. you can use OCLC's and/or Amazon's ISBN cross-reference services which tell you which ISBNs are related. In addition, I can scan catalog records for subject heading like "Science fiction, French" or "Fantasy -- Dictionaries". It's not trivial, but perfectly doable, I just need some time to get this logic in shape. Ahasuerus 19:50, 20 May 2011 (UTC)

If the source isn't directly linkable, then a Worldcat or even an Amazon link in the Mod Notes would be helpful. BLongley 23:55, 20 May 2011 (UTC)

It might be a nice idea to make the submissions from a different bot account though? It seems many mods ignore Fixer submissions almost automatically, they won't even bother to read the "Data from" note. BLongley 15:58, 20 May 2011 (UTC)

OK, User:Bibliovore has been created. He will be dormant for a bit while I am teaching him the tricks of the trade. Ahasuerus 19:50, 20 May 2011 (UTC)

I'm sure he will be welcomed, although given his parentage, I don't think it's worth using the welcome template on his talk page. BLongley 23:57, 20 May 2011 (UTC)

Cover art record creation process

I'm not sure why, but I thought this had been fixed. When entering a new pub with one cover art by two artists, the system creates two different records. I have to go back to the pub (not the cover art record), delete one of the artists, resubmit, accept, then go back to the cover art record, update it by adding the second artist, submit and accept. Now if I remember this when I'm adding the pub, I'll only give one artist, and then update the cover art record. I believe the system works this way because of the two covers on an Ace Double, but that's so rare these days. It's more likely that there would be one cover by two artists than one book with two covers. There should only be one record created giving both artist credits. Can this be corrected? Mhhutchins 00:07, 21 May 2011 (UTC)

It's logged as a bug, Bug 2943196, assigned to me. My first try at fixing it didn't work out so well, and I haven't yet gotten back to it. I will see what I can do. Editing the title is the only workaround. --MartyD 00:23, 21 May 2011 (UTC)

[before an edit conflict] How many editors know that? There could be hundreds of books that have two cover art records when there should only be one. It took me some time to straighten out the records for Leo and Diane Dillon, and every now and then I go back to fix more that slipped through. I'm not sure that it's really a bug (if we're defining a bug as a programming error). It seems the system was designed that way to handle two covers. If I were entering an Ace Double, it would not be considered a bug. Mhhutchins 00:43, 21 May 2011 (UTC)

I'm not sure we've sorted out the Ace Double covers yet anyway. We should probably have the cover artist credited for the cover he/she did, rather than have both for the combined title. (Unless they did both sides, maybe.) BLongley 00:48, 21 May 2011 (UTC)

I do that when I come across one that I'm editing anyway. I haven't gone out of my way to fix them, but I'm sure 90% of the records give the titles for both sides. At least there are two records for each book, so it should be fairly simple to fix. If anyone has need of a project to work on... Mhhutchins 00:51, 21 May 2011 (UTC)

I don't think we've considered it as high priority - we are the "The Internet Speculative Fiction Database" after all, the art is a little secondary. Still, we aim to please. And if it's Marty looking at it I can concentrate on things I am capable of. ;-) BLongley 00:37, 21 May 2011 (UTC)

There is a real bug because editing a name in the pub screen will split a shared record. There are other issues as well. See this list. The discussion here seemed to arrive at a consensus that having a way to specify explicitly additional cover artwork and having the "Artist" field(s) contribute to a single collaboration record would be acceptable. It seems to me the frequency of shared credit is much higher than the frequency of separate works of art, but I'm no expert. --MartyD 02:12, 21 May 2011 (UTC)

Corrections to the James White Award

The years that we have listed for the James White Award do not agree with the years listed on the Award's Home Page History Listing. (The standard "off by one" year error on year of publication vs. year of award.) I will be submitting a substantial number of corrections to that to bring our awards into agreement with the (current) official years. One that I cannot correct is a 2002 listing without a link to Vita Brevis, Ars Longa, by Julian West. I will create a new award listing for 2003 for that story, which we now have in the database, but I can't remove the previous listing. Alvonruff or BLongley will (presumably) have to do that. Chavey 14:21, 21 May 2011 (UTC)

Fixed, but I'm not sure why you couldn't do it. I just went to the Listing by Year and used "Remove an Award". BLongley 15:31, 21 May 2011 (UTC)

I guess it's cause I hadn't noticed I could do that! Now I have a few more things to try, which make my recent posting on the Community Portal obsolete. Chavey 15:56, 21 May 2011 (UTC)

I guess this is one of the areas where you can just update help screens rather than ask for software improvements or Mod help. ;-) BLongley 17:16, 21 May 2011 (UTC)

Sidewise award submission for "Panacea"

This short story is not a bibliographic title. It was posted online at the old SciFi.com, but that site deleted all of their stories. A copy is kept at an archive of SciFi.com, and I have included a link to that online source of the story under the "Movie URL". I don't know if that's a good way to submit such a link, but since that award item can't get linked to an ISFDB title record, linking it to that site seems like a reasonable option. Please let me know if this makes it easier or harder to deal with the submission. Chavey 23:59, 21 May 2011 (UTC)

The link disappeared. Not sure why, but this is all new to me as well. Mhhutchins 02:08, 22 May 2011 (UTC)

Figured it out. The movie URL field links directly to IMDB.com regardless of what you place in the field, even another website's URL. So we know that little experiment doesn't work. Mhhutchins 02:14, 22 May 2011 (UTC)

I just finished processing "Five Guys Named Moe", which has a similar publishing history. First I created a new Novel record for it, entered the title, author and data, then approved it. I then deleted the publication record, which left a "hanging" Title record. I then changed the Title type to "SHORTFICTION" and added its Sidewise Award nomination. Finally, I entered the URL of the Web page where the story is archived in the "Web page" field. Does it sounds like a reasonable way to process this type of award nominated records? Ahasuerus 01:43, 22 May 2011 (UTC)

I've been adding, slowly, records for the stories published on SciFiction. It's current up through the first few months of 2003. More will follow and I'll properly link "Panacea" to the award when I've created a record for it. Mhhutchins 02:08, 22 May 2011 (UTC)

I went ahead and created a record for the issue of SciFiction in which "Panacea" appeared. It's linked now to the db title record. Mhhutchins 02:08, 22 May 2011 (UTC)

Looks good! Ahasuerus 02:20, 22 May 2011 (UTC)

And a record for Sci Fiction, February 23, 2004 has been created for "Five Guys Named Moe". Wouldn't want any orphaned/stray/stub title records now, would we? (I'm being sarcastic. If you've read my comments lately about stub title records, you know I'm fine with them.) Mhhutchins 02:36, 22 May 2011 (UTC)

I figured you'd be knocking on its door soon enough :-) Ahasuerus 02:57, 22 May 2011 (UTC)

Interviewers from Locus1

I'm entering the contents for the magazine "Paradox: The Magazine of Historical and Speculative Fiction" with data primarily from Locus1. All of the interviews that they have are listed by Locus1 as being done by "Misc.", so that's what I've been entering. But that doesn't seem right. If not, what should I be entering as the interviewer? "unknown"? I have to go back and do a little cleanup on most of the issues (I neglected to list the short fiction "length" field, which I'll do after current submissions are approved), so if there's something better to use for the interviewer, I can change that then. Chavey 06:52, 22 May 2011 (UTC)

I've answered my own question. The moderator approving the submission says I shouldn't trust that "Misc.", and I agree; and the Author Template] says I should used "unknown" in this case. Chavey 13:16, 22 May 2011 (UTC)

Sorry, I didn't see this before handling the original submissions. Use "uncredited", not "unknown". --MartyD 15:01, 22 May 2011 (UTC)

The relevant bit of the help here seems to be "If you are working from a secondary source which does not specify the author, but does not explicitly state that no author is credited in the publication, use "unknown" rather than "uncredited". " BLongley 15:43, 22 May 2011 (UTC)

Never knew that. But in this case there is a contents list of sorts on the Paradox Magazine site, and the interviews are listed, but uncredited, there. At least the three or so I spot-checked. --MartyD 16:51, 22 May 2011 (UTC)

Yes, but it's not clear whether those are uncredited within the actual magazine, so it seems safer to leave it as "uncredited" until/unless someone physically verifies them. A friend of mine has published in the magazine, so I'll try to check with her as to whether she has some actual issues of it. Then I could at least do a transient verification. Chavey 17:53, 22 May 2011 (UTC)

I don't believe that "unknown" (as an ISFDB credit) means the author is not known to me or a secondary source, but that the author is unknown to the publisher or editor. If an editor chooses to publish a story and he doesn't know who the author is then our record should be credited to "unknown". Or if a story is published using a pseudonym for which we don't know who the original author is then we create a parent record as "unknown". (That's how many records for house pseudonyms are credited already in the database). IMHO, the documentation that Bill has quoted is wrong, at least as far as our de facto standards go. If I were entering data from a secondary source which doesn't give author credit, I would enter it as "uncredited" and give the secondary source in the notes and/or do a secondary verification. When an editor with the primary source comes along, he can always update the record, if the author is actually credited. Mhhutchins 20:06, 22 May 2011 (UTC)

The Help page that Bill quoted accurately reflects the original intent, which covered undisclosed pseudonyms as well as the "secondary sources" case. Here is the error message that editors see if they try to enter an author-less pub:

• Every publication must have an author/editor. For uncredited publications, use 'uncredited'. When entering data from a secondary source and the author/editor is not known, use 'unknown'. If the author is explicitly credited as 'Anonymous', use 'Anonymous'.

The idea is that "uncredited" is a positive statement and we are sure that there is no credit given in the publication. On the other hand, "unknown" reflects our lack of information about the way the author is credited. Of course, it's possible that the de facto standard has evolved, although I haven't noticed it. Ahasuerus 20:29, 22 May 2011 (UTC)

I agree - "uncredited" to me is a statement that we know there is no credit given. I'm not in favour of creating parent records of "unknown" as it just hides the fact that we really would prefer to know. Nobody ever clears up "unknown"s author page. BLongley 20:33, 22 May 2011 (UTC)

It's admittedly iffy, but if we don't create "unknown" VTs when we don't know who is hiding behind a house name, then the main page appears to suggest that it's the name of a real person, e.g. see the way Alexander Blade and Alex Archer look at the moment. Ahasuerus 22:20, 22 May 2011 (UTC)

Those don't look too confusing to me. When it's just one known author and the rest are unknown then it might be confusing. BLongley 22:30, 22 May 2011 (UTC)

The easiest way to make it clear that an "author" is a partially undisclosed house name like Rob Kidd may be to make "Used As Alternate Name By:" stand out at the top of the page. The way it appears at the moment, it can be easily skipped by a casual user. Ahasuerus 03:38, 31 May 2011 (UTC)

Sidewise award, An Alternate History of Chinese Science Fiction

I have submitted a Sidewise award, without a link to the item. This is a non-bibliographic item, published online at blogspot.com. Chavey 22:51, 23 May 2011 (UTC)

Fixer - 2007

November and December have been submitted. Ahasuerus 03:27, 31 May 2011 (UTC)

October 2007 has been submitted -- and from the new development server too! The pubs were not particularly exciting, but on the bright side, our new robot, Bibliovore, is almost done parsing 7,030,372 publicly available catalog records from the Library of Congress. Here is what he has found:

• 25,010 ISBNs that we already have on file;
• 7,825 ISBNs that we don't have, but that Worldcat indicates are linked to the ISBNs that we do have;
• 12,867 additional ISBNs that have subject lines of SFnal interest
• 287 ISBNs that have titles of SFnal interest
• 6,133 pre-ISBN books with SFnal subject line (unfortunately it will be hard to tell whether we already have them on file)
• 290 pre-ISBN title of SFnal interest

It will be another week or three before Bibliovore can start submitting these items, but they should be more exciting (and hopefully more reliable) than what Fixer usually finds in old Amazon records. Ahasuerus 03:49, 6 June 2011 (UTC)

September 2007 has been submitted. Once it gets processed, August 2011 will be up next. Ahasuerus 00:34, 14 June 2011 (UTC)

"Scott Grimando." (sic)

There is a bibliography page for the artist Scott Grimando. (with a period in his name). This page has no bibliographic entries, and even an advanced search gives me no hints as to why that page is there. Presumably, there's some type of spurious record that refers to this name, instead of linking to Scott Grimando (no period), but I can't find that record. I'm hoping one of you can, and can fix it up. Chavey 01:10, 1 June 2011 (UTC)

There is a known bug with cover art deletion which leaves title-less author records behind. We don't know what exactly causes it, but it's fairly common. Ahasuerus 02:30, 1 June 2011 (UTC)

I was able to replicate the issue of orphan artists by cloning a random Scott Grimando publication to be by "Scott Grimando." and then doing a delete-pub on the new record. This created and then orphaned the author record 158118. I've merged both of the orphan "Scott Grimando." records back with the artist's main record. --Marc Kupper|talk 01:14, 12 June 2011 (UTC)

Duplicate publication records

[Error] There are two publication records for the May 1986 edition of C. J. Cherry's "Visible Light", one with less detailed notes and one with more detailed notes. We don't support a "Merge Publication" process, and the Help pages tell me to delete one of them. But each of these records has verifications, so deleting either one loses some of those verifications. I hope that a moderator can figure out how to do an appropriate merge. Chavey 04:58, 2 June 2011 (UTC)

My mistake. One edition is the U.S. printing, the other is the Canadian printing. This isn't obvious, since each edition lists the U.S. price in the "Price" field, and all data outside the notes are identical. Chavey 05:07, 2 June 2011 (UTC)

Yeah, it would be nice if we could indicate the price in Canadian dollars...but unfortunately, in some cases, a book could be published by a US publisher, but printed in Canada and sold in both countries with only one price. In this case, there are prices for each country in the Canadian printing. I would suggest reversing the price: place "C$4.50" in the price field and the US price in the note field. Verifier would have to do it though. I'll point him to this discussion. Mhhutchins 19:18, 2 June 2011 (UTC)

We've had this discussion [before]. If only DAW hadn't been so contrary. I have done these DAWs both ways [the dual pricing came into effect in Mar.'84 and DAW ceased printing in Canada in early '88[?]] so the number of possible cases like this might be 250-300. Unfortunately, for me, DAW was publishing mostly fantasy in those years so I can't narrow the latter date down any further. HELP leans toward country of origin for the pre-eminent price, though in these cases that's just for the cover. Still, it's where the prices are. I would [almost] rather see DAW (Canada) or NEL/DAW [no space as one is not an imprint] as publisher, as NEL Canada actually published/printed/distributed them. Maybe that's worth exploring in an over-all sense, would render this discussion moot. That said, I don't see a compelling reason to make forced differentiations in the bibliographic page when it can be done in the notes. Each record is only a click away from its' apparent 'duplicate'. My 2¢, Canadian of course! [but hey, with today's exchange that's worth 2.04¢ US!!] ;-) --~ Bill, Bluesman 20:30, 2 June 2011 (UTC)

We should probably improve "Publication Diff" to show the notes as well. BLongley 20:38, 2 June 2011 (UTC)

Thinking about it some more, would it be desirable to make the price field a "repeating" field? Ahasuerus 17:01, 10 June 2011 (UTC)

I'm fine with the price field a "repeating" field. I assume it's so someone can record those publications with multiple prices such as: $3.50 U.S.A, C$4.50 Canada.

I was not aware that DAW had Canadian printings of the dual priced books. Per what Mhhutchins wrote, I'd also say the Canadian publication's record should be changed to report "C$5.40". Both of them state "(Canada) $4.50 • U.S. $3.50". The U.S.A. publication would be $3.50 and DAW is allowed to sell them duty free in Canada for C$4.50 while the Canadian edition is sold for C$4.50 and DAW is allowed to sell them duty free in the USA for $3.50.

Four people have verified the USA edition and apparently are fine with the note "(the Canadian edition apparently has a different DAW number)." I'm wondering what that note means. The number looks the same in the Canadian edition.

For a while I entered them using the publisher name "DAW (Canada)" but some editors objected to that and so now I price them using C$. I've also gotten into the habit of always noting the country of printing when verifying records to reduce the chance of a future "gotcha" like what happened with DAW. --Marc Kupper|talk

Imagining Mars Color Plates

I just submitted detailed contents for Imagining Mars and there's a bit of weirdness now that it has been approved.

The book has a series of color plates inserted between consecutively numbered pages. When documenting the page numbers, I used this format from the documentation:

<text> -- descriptive of the location in some other way

However, the "page numbers" do not display on the Publication Listing (Full Listing View), and stranger still, the titles are not linked.

A) Did I misunderstand the documentation? B) Is the documentation out of date? C) Are the angle brackets breaking the update process? (Unescaped angle brackets wreaking havoc on XML?) D) Some combination of the above? Albinoflea 15:59, 10 June 2011 (UTC)

I believe the answer is "C". I don't think you're supposed to use the carats, which are usually there to indicate where the text would go. And honestly, I don't understand the documentation. Personally, I don't place page numbers on unpaginated plates, and just record the two pages that they're between in the record's note field. Mhhutchins 17:44, 10 June 2011 (UTC)

I removed the carats and they appear to be displaying correctly. Was there a Plate 3? Mhhutchins 17:46, 10 June 2011 (UTC)

There was a Plate 3, but it was just a photograph of the front page of a newspaper, so I didn't include it, nor did I include the other photographs; however, re-reading the documentation I come up with this apparent contradiction:

Artwork. Cover art is always included. Interior art is usually included; see INTERIORART. Photography. Treat as artwork. from EditPub#What_to_Include

If an article is illustrated with diagrams, or with photographs, these do not need to be included; they are not "artwork" in the sense that we are indexing. from EditPub#EntryType

I can always go back and add them in... Albinoflea 01:30, 11 June 2011 (UTC)

Local Install

Two issues with the local install: one I think I've resolved, but want to make sure before I update the Windows Setup documentation, the other I need some advice about.

1) When editing localdefs.py the current documentation doesn't list WIKILOC, so my install was throwing an error. I set this to www.isfdb.org/wiki (in four locations, because I had already unpacked) and got my install to run, but want to verify that this is correct before I alter the documentation. I figure just because it seems to work doesn't mean it's the right way to do it.

That's fine. At least, that's what I use. Keep in mind that it's pointing to the real Wiki. --MartyD 00:07, 14 June 2011 (UTC)

I use the same. If anybody has a local Wikimedia installation working then we could do with advice on how to do that - although I think the Live one is so out of date you can't download the files to exactly replicate it locally anyway. :-( BLongley 01:02, 15 June 2011 (UTC)

2) When viewing an Author my install chokes about halfway through. The error it is throwing is as follows:

<type 'exceptions.AttributeError'>	Python 2.5.4: C:\usr\bin\python.exe

A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.
 
C:\Apache\cgi-bin\ea.cgi in ()
   95                 bib.printASeries()
   96                 bib.displayWorks('ANTHOLOGY')
   97                 bib.loadCBSeriesData()
   98                 bib.printCBSeries()
   99                 bib.displayWorks('CHAPTERBOOK')
bib = <biblio.Bibliography instance at 0x01794A08>, bib.loadCBSeriesData undefined

<type 'exceptions.AttributeError'>: Bibliography instance has no attribute 'loadCBSeriesData'
      args = ("Bibliography instance has no attribute 'loadCBSeriesData'",)
      message = "Bibliography instance has no attribute 'loadCBSeriesData'"

Not sure what I'm missing here... any thoughts? Albinoflea 21:50, 13 June 2011 (UTC)

This is because Ahasuerus changed biblio/biblio.py and common/SQLparsing.py and in the process eliminated some changes I had made in the prior version of each, on which other changes that are checked in are dependent. See the "Chapterbook series" row in Development#Outstanding_changes. He has marked the files he changed with "needs merge". You can get the earlier version of those and edit back in the changes, or you could get the earlier version of each of the other files. I've been avoiding changing anything while he finishes up what he's doing. --MartyD 00:07, 14 June 2011 (UTC)

Sorry about that. There are still some changes in the queue that need to be tweaked and installed first, but we are getting there. Ahasuerus 05:35, 14 June 2011 (UTC)

If we were more experienced with CVS we should be able to tag the current live release of software and let people download that, rather than have to put up with the unstable versions that I create. We're always open to ideas though, so if you can explain how to do such, or even explain "branching" a bit better I'm sure we'll listen. Unfortunately my past experience was for about six months six years ago, so I never developed any expertise in that area. BLongley 01:02, 15 June 2011 (UTC)

Restored the two files that MartyD mentioned, then rebuilt, but now I'm getting other errors... not sure if I need to get earlier versions of the other 3 files that were included in the Chapterbook Series update as well... I'll hold back for now. Albinoflea 04:45, 15 June 2011 (UTC)

Edits that won't clear/Title Record edits that don't 'take'

Two separate issues, one just from yesterday/today; one that's been happening to me all too frequently.

1. Yesterday accepted three Award edits from Chavey [8730, 8732, 8735] but none will clear from the queue. Tried the first one today with the same result.
2. For several weeks have been having trouble with edits [that I do] of Title Records. No matter what the change, the edit will not go through. At first it was intermittent, now it's nearly every one. If I let the window/tab just sit there and 'spin' eventually I get a message "Internal Server Error". The odd time I'll mess up the month [80 instead of 08] and will immediately get the error message about the format but then the corrected edit still won't go through. I've tried everything: back out of the window, re-load, try to submit again. Sometimes it takes over a dozen tries before the edit will go through. At first it was just an annoyance, now it's [pick your own epithet] awful! Have any of the current changes tweaked something? And this happens only to edits of Title Records [though the image upload has slowed dramatically of late]. I run a MAC, latest Safari browser, but I don't think that should matter [though it has in the past, some things on the DB didn't display properly, but nothing recently].

I'm at a loss, on both issues. --~ Bill, Bluesman 22:20, 14 June 2011 (UTC)

I can explain the first - there are bugs in Award Editing that mean Award Categories with ampersands in get XML Parse Errors, and ones with apostrophes silently error out. I've submitted a fix for those, we just need to give Ahasuerus some time to test and implement it. (If you see an "XML Parse" Error just hardreject it for now as those will never work even after the fix.) BLongley 00:38, 15 June 2011 (UTC)

As at 21 Nov 2011, I've developed a workaround for the apostrophe issue (I've been adding some Aurealis awards):

Step 1 - Add the award without an apostrophe (eg Childrens), then moderate

Step 2 - Edit the award & put in 2 apostropes (eg Children''s)

Hope this helps ...--clarkmci / j_clark 07:43, 21 November 2011 (UTC) (Later: I did try the html & # 39 ; (no spaces), but this didn't seem to work for Add Award Info either (silent reject too). --clarkmci / j_clark 10:12, 21 November 2011 (UTC)

As to the second - I'm at a loss there. I've noticed more timeouts in recent days but put those down to the latest anti-spam measures (and on balance think it's still a good trade-off). Nothing special about title edits that I can think of though. :-/ BLongley 00:38, 15 June 2011 (UTC)

Seventeen minutes, four "Time-Outs" [Internal Server Error], to get one minor edit [not to a title record this time] through..... I don't see this as a good trade-off. --~ Bill, Bluesman 22:29, 15 June 2011 (UTC)

If it's been for weeks then it's not the anti-spam measures. And if it's now affecting non-Wiki, and Non-Title-edit stuff, then maybe we need to pin down exactly when it happens and let Ahasuerus go moan at the hosting provider. BLongley 03:23, 16 June 2011 (UTC)

(after edit conflict with Bill) The way the spammers were defeated earlier this week shouldn't have had much to do with performance, especially considering that you have been seeing these errors for weeks. The interesting thing is that I haven't seen timeouts on the ISFDB side in months and no one else is reporting the same problem, so something about the way you go about editing is likely to be different. Perhaps you edit very late at night/early in the morning, around 2am-3am Eastern (1am-2am server time)? We share the underlying physical server with other systems and some of them appear to put a significant load on the system around 2am ET, which causes ISFDB to slow down. Then we do our backups at 3am ET, which means that the system is unavailable for 10 minutes or, if other systems are doing something major at the same time, possibly longer. Ahasuerus 03:28, 16 June 2011 (UTC)

I'm not sure about the "way you go about editing" above? Other than open a record, make a change and submit, how many "ways" are there? Seriously, if there are other options that will make the process work as it should, I'm all ears!! And with no hair left to pull out they're wide open! :-)) --~ Bill, Bluesman 14:47, 16 June 2011 (UTC)

I'm frequently editing between 2 and 3 am EST and have run into the backup, at which time everything grinds to a halt. I never knew what it was but I assumed some sort of indexing or something was transpiring on the server side since it was so regular. (It is also usually my cue to go to bed...) Something should be posted to the blog regarding these regularly scheduled down times. Albinoflea 04:53, 16 June 2011 (UTC)

If I'm home I'm on here from around 8AM Eastern to maybe Midnight Eastern [not continuous, that's just the range], rarely later [old farts need their beauty rest!]. The speed of the server has been decreasing steadily, pretty much from the time a lot of the newer features became live. The Title edits were the first 'hung' edits noticed, because it's so frustrating. And sometimes [think maybe three times in the last week, once yesterday] the edit will hang until the error message, and yet going to the queue instead of trying to re-send and the edit is there! I tried to leave a note on an editor's page yesterday, got the error message a couple of times before the edit went through, only to find six copies of the note on the Talk Page. Then it was another ten minutes to delete five because the deletions would hang. I can't pick a time of day when it's better or worse. --~ Bill, Bluesman 14:37, 16 June 2011 (UTC)

If the problem were with the server (e.g. some hardware component going bad, too much extra load, etc), all editors would see the same behavior, but it would appear that you are seeing the worst of it. It's possible that some of your edits are unlike other edits, e.g. you may be populating some obscure field(s) that the software has a problem with and that no one has been able to recreate consistently so far. Or perhaps you tend to edit pubs with a lot of titles in it, which can make the software unhappy. It's hard to tell without analyzing your submissions in depth, which is possible (they are all saved) but time consuming.

95%+ of the Title edits simply change the date, either adding a month or zeroing the day portion if the pub can't support it. I see from below that Wiki edits are out of your control, but so far, other than that one weird incident mentioned above, those 'hangs' are less prevalent [still frustrating]. I have let some Title edits 'hang' until the error message but have opened another tab, done different edits while that one was going nowhere, and they went through just fine. The strange thing is, there doesn't seem to be anything new added to the fields [that I can see, anyway] of a Title record. Yesterday morning the edits were going through fine and then started to 'hang' around 10AM [noon Eastern], and remained intermittent all afternoon. There were no titles with odd punctuation, or author's names that had any strange characters [first thing I started to look for when one would hang]. There is nothing different about the way I do these edits, one from another, in the ones that 'hang' from the ones that don't. --~ Bill, Bluesman 15:00, 17 June 2011 (UTC)

Also, keep in mind that the Wiki is a totally different application written by a different team. We have little control over it and we don't understand what it does internally. We all know that Wiki edits can take a long time to be processed, but we don't understand why. I know that the software performs a lot of checks when you click "Save page", but I don't know what most of the checks are and whether some of them are unnecessary and could be disabled to improve performance.

The ISFDB application is a different story. When we identify a problem with the software we can usually do something about it. For example, a few months ago I got tired of waiting for the Moderator queue page to load and looked under the hood. To my surprise, I discovered that the table which stores submissions was misconfigured: every time you refreshed the page, ALL submissions had to be scanned. With hundreds of thousands of submissions on file -- and growing all the time -- you can see how it was a problem. As soon as I realized what was going on, I added an index so that only new submissions would be scanned and performance improved noticeably.

Wish we had the same level of control over the Wiki, but we don't. Just figuring out what it does and why would take dozens of hours of my time. Ahasuerus 04:30, 17 June 2011 (UTC)

That's ... strange. Has anyone else run into this problem with Title edits? Ahasuerus 04:30, 15 June 2011 (UTC)

I've only seen problems with Wiki edits. Non-Wiki edits have been behaving normally for me. This 6:00-7:30am ET. --MartyD 00:22, 16 June 2011 (UTC)

I've only seen it on the Wiki side, not on DB edits. And the Wiki seems to be OK at the moment, although I really don't want to have to adjust my editing time to these hours regularly... BLongley 03:15, 16 June 2011 (UTC)

Can I be helped?

I have a story of sorts (titled You Can Be Helped) written as a letter to 'Fellow sufferers of the malady that afflicts us' in SF Eye #13, but I don't know if to credit it to 'uncredited' or to 'Anonymous', since it is in fact uncredited but introduced with the short paragraph: 'Anonymous posting discovered on Internet - today's version of yesterday's clandestine xeroxed posters and fliers'. Does this introduction allow for a work by 'Anonymous'? Stonecreek 18:26, 16 June 2011 (UTC)

Seems like one of the few examples where an uncredited piece should be credited to "anonymous". If the editor knew who wrote it, but chose not to credit its author (a book review, for example), I would have pushed for "uncredited". But it appears the editor did not know the author's name and would have probably credited it to the author if he did. Without the introduction and not knowing its source, it would have probably been entered as "uncredited". Mhhutchins 21:07, 16 June 2011 (UTC)

Hm, I have more the impression that the editor, Stephen P. Brown, did know exactly who wrote this piece (in fact he may have been the author himself). The introduction seems more to be part of the story (it is not credited, too). It is difficult to decide. Stonecreek 15:01, 17 June 2011 (UTC)

I'll leave it up to you to decide which way to go then. It's 6 or half-a-dozen... This reminds me: I wish we had a type for FA ("facetious article") like Contento uses in the Locus database. Mhhutchins 16:15, 17 June 2011 (UTC)

Fixer - August 2011

Just a note that we are more than half way through August (thanks to certain moderators who have been diligently working on the queue :-) ) Ahasuerus 04:27, 21 June 2011 (UTC)

US done, UK in progress. Ahasuerus 01:51, 22 June 2011 (UTC)

August submitted, approve away! Ahasuerus 00:10, 24 June 2011 (UTC)

Proposed Make Variant Submissions

As Dirk seems to be playing Cover Pelmanism, do you think we should improve the Mod screen for COVERART merges to show thumbnails of each title, where we have them? It should be possible, but for popular art it may be a bit of a burden on Mod bandwidth. (Although I suspect most Mods are on Broadband now.) BLongley 16:53, 25 June 2011 (UTC)

I would suggest thumbnails only if the submissions are for merges and variants of "COVER:" title records. Mhhutchins 13:19, 26 June 2011 (UTC)

OK, I'm thinking something like this:

I'm not sure what size they should be - I could make them smaller and clickable-through to larger images if people want to check DAW Issue numbers or something? And I could put a limit on the numbers of images shown for each title. BLongley 19:08, 26 June 2011 (UTC)

I like it. For me, bandwidth's not an issue. --MartyD 23:15, 26 June 2011 (UTC)

OK, that's one Plus point. I'll try and dummy up something similar for Variants (and I see I started with Variants and got side-tracked to Merges) and if that goes well I'll overload Ahasuerus again. ;-) (Actually, all this should be far simpler than the Award work I've left him to check.) BLongley 00:08, 27 June 2011 (UTC)

Yes, it looks useful. I doubt it will overwhelm the 18MBps connection used by my test server, but it shouldn't be much of a problem even on a 56Kbps line. Ahasuerus 01:02, 27 June 2011 (UTC)

The proposal is fine with me though I believe it's quite rare that a publisher will ask the same artist to do a second painting for the same story where we'd want to reject merge. If a moderator has very limited bandwidth they can chose to not process coverart merges. Using thumbnails will conserve screen real-estate but not bandwidth as we are sending the full-size image file but telling the browser to scale it down to 200px high or whatever thumbnail size we decide to use. --Marc Kupper|talk 03:27, 27 June 2011 (UTC)

It's probably more useful for identifying miscredited art, e.g. I'm suspicious of this:

And yes, the scaling down won't affect the bandwidth but limiting the number of images shown would. I don't know what the limit should be though. BLongley 15:30, 27 June 2011 (UTC)

I've dodged the issue by making any limit Ahasuerus' problem. ;-) BLongley 17:24, 27 June 2011 (UTC)

(Unindent) Also committed a change for COVERART variants like this:

I think this might start a discussion on when to use "(excerpt)" for Artwork. ;-) BLongley 18:05, 28 June 2011 (UTC)

Duplicate publication records [2]

Not perfect by any means, but for the few people that DO "Diff Publications", is this an improvement?

I suspect people will want a load more publication details Diffed, but most of the other differences are visible on the page you see before you attempt this. BLongley 23:20, 26 June 2011 (UTC)

Actually, I probably posted this in the wrong place as ALL Editors should benefit from this, if it's considered good. It just adds Notes and Cover Image to what we already have. There's probably a possibility of making the Diff, Edit and Mod screens more useful in this way - although the recent "Moderator notes" improvements might make it a lower priority than some other stuff. BLongley 23:37, 26 June 2011 (UTC)

If someone could explain to me how to use the "Diff Publications" function, I'd gladly use it. Unless I'm doing something stupid, sometimes it works and sometimes it doesn't. What I've learned is that it only checks contents. Why doesn't it check data fields? Mhhutchins 02:31, 27 June 2011 (UTC)

You'd have to ask Al what the intention was. It's particularly useful when checking whether two collections or anthologies have exactly the same contents, but absolutely useless when the difference is so minor it's only in the notes or the cover. I suspect that the thinking was that most of the other differences will be spottable just from the list of publications. BLongley 15:42, 27 June 2011 (UTC)

I've committed the above change, if people want further changes just add them to FR 3337923 "Improve Diff Publications". BLongley 17:41, 27 June 2011 (UTC)

Walt & Leigh Richmond: Siva! as VT of The Lost Millenium

Some submissions have room for notes to moderators, but not submitting a title as a VT of another, so I revert to the older form of sending a note to the moderator who checks this. The book Siva! by Walt & Leigh Richmond is currently listed as a standalone book (previously unverified). I just got a copy, and the copyright page says "Originally published as The Lost Millenium". The introduction, by Leigh Richmond (a somewhat recent widow when this book was re-published) discusses the previous version only slightly:
Several years after the publication of "Siva!" under the title "The Lost Millenium", Erich von Daniken did an excellent job of outlining a great many of the anomalies of history at which I had been looking ... and on which "Siva!" is based.
Although it is possibly that she revised the work somewhat based on von Daniken's book, she makes no claim to that effect. She continues discussing the development of the story, but again makes no distinction between the two titles. Chavey 21:22, 4 July 2011 (UTC)

I've added software changes to allow Mod Notes for Make/Remove Variant too. BLongley 18:14, 5 July 2011 (UTC)

Cool! "Ask and it shall be given" (Matthew 7:7) Chavey 02:18, 6 July 2011 (UTC)

"Maybe." (Bill 7:7) ;-) Seriously though - we are trying to get people to report things on Sourceforge as opposed to random requests on various parts of the Wiki - see Development#Software_Development_Process - but if I spot something on the Wiki that looks like a Bug Report or Feature Request that I can work on I might transfer it for you. But while we have lots of Editors, they outnumber the Moderators, who outnumber the Coders (although you don't have to be a Moderator to be a Coder), who outnumber the Testers (maybe "Tester", singular?), and we only have one "Implementer" so it's not necessarily going to be a quick fix. But if you don't ask, you probably won't get. BLongley 01:02, 8 July 2011 (UTC)

I've submitted 7 feature requests and bug reports [One by "nobody" when I forgot to log in. Feature Request to SourceForge: Notice when somebody does that, and ask them if they want to log in :-) ]. But I hadn't really thought of this as a feature request. I'll try to submit some more FRs and Bugs so you have more options when you want to ignore "X, the Y Fairy" books :-) Chavey 14:17, 8 July 2011 (UTC)

Cover art merges

Thanks to Bill's unrelenting attack on bugs and FRs, you can now see all cover art images affected by Cover Art merges. Ahasuerus 06:27, 12 July 2011 (UTC)

I think that makes a nice round 50 "improvements" I've done now, unless I've lost count. Some of them might even be useful! BLongley 22:39, 12 July 2011 (UTC)

They are. It's just that we all get used to good things and take them for granted :) Ahasuerus 05:59, 13 July 2011 (UTC)

That's a feature I like, too. Thank you, Bill! Stonecreek 14:03, 13 July 2011 (UTC)

Cover Art VTs

In related news, you can now see all affected cover images when one a COVERART title is about to be turned into a VT of another COVERART title. Ahasuerus 05:59, 13 July 2011 (UTC)

Help with a Voyager / HarperCollins dating issue

Would someone familiar with Voyager / HarperCollins lend me a hand with this discussion around this held submission? If you're not familiar but have an opinion based on experience, that's of course quite welcome, too. Thanks. --MartyD 17:45, 16 July 2011 (UTC)

It has to be later than 1996 - voyager-books.com wasn't registered till two years later. And that "14" is indeed a number-line, albeit a very short one. BLongley 18:37, 16 July 2011 (UTC)

Add Variant improvements

After approving an "Add Variant Title" submission, you can now go to the canonical title or to the newly added variant title record. Ahasuerus 09:27, 17 July 2011 (UTC)

Amazon changes and Fixer - October 2011

Effective October 26, Amazon will be changing (and greatly restricting) the way you can search their database. This will make Fixer's current logic pretty much useless for our purposes. I will have to spend some time working on alternatives, but in the meantime I hope to squeeze as much data out of the current interface as possible. At the very least, I hope to have all SF books published in 2000-2011 saved off for future processing before the deadline. Always something... Ahasuerus 13:02, 27 July 2011 (UTC)

Fixer has finally finished extracting all (potentially) eligible records for 1990-2011 from Amazon. Even after suspending a ton of comics and other ineligible books, we have 115,000 ISBNs in the main queue. It will take a long time to process, but at least we have the data on file. I will set it aside for now and ask Fixer to get the October 2011 data next. Ahasuerus 15:22, 1 September 2011 (UTC)

It boggles the mind to think that we have missed on the average more than 5,000 books per year for the past twenty years. What really sends the brain reeling is the idea of moderating those submissions. Mhhutchins 02:05, 2 September 2011 (UTC)

I checked the figures - you've averaged over 4000 Approvals a month, so you'll only take 29 months to clear them all. :-) I only do about 3000 a month so you don't want to leave them all to me or it'll take 10 months longer. We may need to get the blackjack out again... BLongley 19:09, 2 September 2011 (UTC)

Fixer - September 2011

US submitted, UK to follow in a day or two. Ahasuerus 03:25, 1 August 2011 (UTC)

September 2011 submitted. Trying to come up with workarounds to keep Fixer in business after October... Ahasuerus 19:41, 7 August 2011 (UTC)

Something goes horribly wrong when editing persons

. --Dirk P Broer 23:37, 3 August 2011 (UTC)

I think some of our improvements may have come back to haunt us. Long long ago... I don't think we had the color highlighting, and the 'proposed' column looked different. Now to understand what will be... add up the green for lack of a better description. The green cells will be saved into the updated record. The red cells will be discarded. Kevin 23:51, 3 August 2011 (UTC)

Dirk, I'm not sure what you're speaking about. Can you point out what's wrong? From the screen capture, it looks like you added the birth and death dates, and the wikipedia and website URLS, indicated by the green fields under "Proposed". Thanks. Mhhutchins 04:14, 4 August 2011 (UTC)

Probably should just let Dirk answer for himself, but I'm guessing he had the the same reaction as the first time I saw that screen. The layout implies that the edit is deleting the canonical and lastname fields (the proposed column has them as blanks). The color coding is not the most intuitive approach. To make it more user friendly, the unmodified fields could be listed in the Proposed column also and then just color the fields that have changed. --JLaTondre 20:51, 4 August 2011 (UTC)

Couldn't express it better than above. --Dirk P Broer 00:00, 5 August 2011 (UTC)

Same here. Hauck 14:23, 5 August 2011 (UTC)

Well, we could certainly do it, but it would mean either a complete change to the way moderator pages work or a discrepancy between editor-viewable and moderator-viewable pages. Considering that it's a one time "surprise", it may not be very high on the list of priorities. Ahasuerus 00:54, 8 August 2011 (UTC)

Toyman

I have indeed a 1976 copy of Toyman (Arrow) at hand, like the one at ebay here. --Dirk P Broer 00:12, 8 August 2011 (UTC)

"Make Pseudonym" improvements

The moderator approval page for "Make Pseudonym" now links to the canonical name and the proposed pseudonym. Ahasuerus 01:54, 8 August 2011 (UTC)

New Worlds

Does anyone know why many of the issues of New Worlds have issue number and month but no year? I just approved a submission that among other things removed the 1962 from that December issue, but the editor was only making it consistent with all of the other entries for that year. And looking at other years, I see the practice is pretty widespread. I'd be inclined to "fix" them, but I figured someone must have approved them as they are.... Thanks. --MartyD 10:55, 14 August 2011 (UTC)

I'm afraid a part of this is of my doing (under torture, I can even confess that I wasn't very constant in this). In fact, I found that "New Worlds Science Fiction, #79 January 1959" was probably a bit redundant (datawise as I'm a member of the generation who was taught that every bit gained counted) with the following "1959-01-00" and has also the effect of cluttering the grid (it's still the case with "Jan" on top of the column and "January" in the cell) and the listings with lines like this : "New Worlds Science Fiction, #79 January 1959, (Jan 1959, etc." (like here). Left to my own devices, I would even only have kept the issue number a bit like this (barring the lenghty discussions about the standardization No., Nr, Number, No etc...). If you wish to "fix" (standardize) this in any way you like, feel free to proceed, you can even consider that I'm notified as the main primary verifier. Hauck 12:12, 14 August 2011 (UTC)

Non-English translations of shortfiction

Are we in a position now to accept non-English translations of shortfiction, with titles entered in the translated language, which will be made into variants of the original English title records? This submission prompts the question. Mhhutchins 16:50, 17 August 2011 (UTC)

We need to implement two more rounds of changes to fully support non-English shortfiction. I believe some editors have been submitting them since they figured it would soon become the new standard, but it's not officially supported yet. I will be traveling next week, but after that I should have some free time and will hopefully clear the backlog of submitted software changes. Ahasuerus 01:59, 18 August 2011 (UTC)

I'd say at LEAST two rounds. "Working Languages" needs following up with some display changes. A "default language" for editors might not need a followup, but I know some editors won't fully use languages till translator support is implemented. And although I have some ideas about such, those again will need some display improvements. And a major discussion as variants will no longer be based on text or title or language changes alone, but on WHO translated them. It's all a bit complicated really, so hopefully people can bear with us a little longer - this stuff isn't easy, you know! BLongley 23:20, 18 August 2011 (UTC)

Author Birth & Death dates

In my "spare time", I'm currently going through a list of the birth and death dates for authors as listed in Locus1, and adding them where missing to the ISFDB. So you will be seeing a bunch of these in the next few weeks. (39 so far, alphabetically through "Anderson"). In all cases, the documentation for these is coming from Locus1, and I will add that notation to the "Bibliographic Comments" for the authors, albeit somewhat later so I can do it in fewer groups. Just thought I'd let you know what I was up to when you see these submissions :-) Chavey 13:02, 25 August 2011 (UTC)

Can we get a field 'Note to the moderator' for author edits for instances as described above? I'd like to name my sources as well when editing authors. --Dirk P Broer 13:40, 25 August 2011 (UTC)

I second the request. Chavey 17:20, 25 August 2011 (UTC)

I think Bill Longley is working on that feature. Until then, consider this a note to the moderator (on this, the moderator's noticeboard). You've just saved having to writing hundreds of notes on each of the submissions. Chavey, I think you should record the source on the "Biography" page, designed for biographical comments. The "Bibliographic Comments" was created for, well, bibliographic comments. Thanks. Mhhutchins 14:01, 25 August 2011 (UTC)

Will do. Chavey 17:20, 25 August 2011 (UTC)

Yeah, FR 3360845 "Author Editor should have Note to Moderator" is on my to-do list. I'm waiting for Ahauserus to catch up on "Working Language" first before I submit anything more in that area though. And to some eyes, including mine, such sources might be more desirable in the DataBase itself rather than left behind on the Wiki - so maybe "Author Notes" would be another useful feature? BLongley 23:21, 25 August 2011 (UTC)

It would be great if, for each piece of data about an author, there was a field that said "where did this data come from?" But that's one heck of a lot of work, both the programming end and the data entry end. As Mike H. points out, it would be best to have this data embedded in the "Biographical Notes" page, and having a way to do it conveniently when adding/updating the information in an Author Edit (i.e. without the multiple steps it takes now) would be nice, and would encourage folks to record this data. That, however, would take some serious design modifications together with substantial coding. But wouldn't it be nice? Chavey 23:44, 25 August 2011 (UTC)

Reviews under pseudonym

Am I alone in thinking this screen a bit confusing? Now who wrote the book? Sybil Lawson, being a pseudonym for eh.. Robert Reginald? But no, Peter Harding is here the alias of Robert Reginald. Is it an idea to have it phrased as Variant Title of: The Possession of Tamara (by Sybil Lawson) Reviewer: Robert Reginald or even better perhaps Variant Title of: The Possession of Tamara (reviewed by Robert Reginald) --Dirk P Broer 20:50, 25 August 2011 (UTC)

I agree. The display could be confusing and changing the wording would help. I suggest:

Variant Title of the Review by: Robert Reginald of The Possession of Tamara (by Sybil Lawson).

The first underline links to the reviewer, the second links to the variant review record, and the third links to the reviewed author. Mhhutchins 21:11, 25 August 2011 (UTC)

I suspect it's going to get even trickier with pseudonymous reviewers of variant titles by pseudonymous authors, but hey, put it on the list and someone will be brave enough to try an improvement. ;-) BLongley 23:34, 25 August 2011 (UTC)

Dian Crayne/Dian Girard

Note to Moderator re "Remove Pseudonym" submission: Dian Girard is a pseudonym for Dian Crayne (full name: Dian Girard Crayne). Currently, the canonical name for her is set to as Dian Crayne. However, we have 7 short stories credited under Dian Girard, and only 1 non-genre book under Dian Crayne (which I suspect should be deleted as "under the threshold"). Hence one of my submissions tonight, to remove that pseudonym, is the first step in changing the Canonical Name from Crayne to Girard. Chavey 05:44, 29 August 2011 (UTC)

Thanks for the note. I accepted the removal. --MartyD 10:30, 29 August 2011 (UTC)

User needs help logging in

Can someone who is familiar with logging in and associated problems please help Lindzay? Thanks. Mhhutchins 18:31, 2 September 2011 (UTC)

Message left. It's usually a Capitalisation problem, or people not realising that logins to the Wiki and to the Database are separate. BLongley 18:49, 2 September 2011 (UTC)

Fixer: October 2011

We are more than half way there. There are 100 US titles left in the queue and another 200+ UK titles. The actual number of submissions will be lower because I always reject a fair number of comics, game books, etc that slip through the software filters. Ahasuerus 22:30, 5 September 2011 (UTC)

US submitted. Ahasuerus 00:32, 6 September 2011 (UTC)

UK submitted as well. It's back to sucking in e-books without ISBNs for Fixer, who has found 69,024 "SF-positive" ASINs published in the last 24 months. Of course, not all of them are eligible for our purposes. Ahasuerus 06:27, 8 September 2011 (UTC)

Fixer: Capturing e-books

In light of the impending change to the way Amazon serves data, I have been trying to capture their records for all Kindle/e-books, which will become effectively inaccessible in late October. It proved to be a more difficult task than anticipated, but I think I got the last software problem resolved earlier today and the data is now streaming in.

Once the dust settles and Fixer transitions to the new way Amazon does business, we'll have to figure out what we want to do with the catch. If you sign on to Amazon.com and check "Books › Science Fiction & Fantasy › Science Fiction › Last 90 days", you will find 320 hardcovers, 1,123 paperbacks, a hundred audio books and 3,136 (sic) Kindle books. The numbers are very similar under "Fantasy". That's a lot of e-books to process... Ahasuerus 22:48, 5 September 2011 (UTC)

I think Fixer should hold on to those a while longer - we're really not prepared for them. I've moderated a handful of such submissions (mostly from Dgeiser13 I think) and even added one myself as the author had been quite communicative in my "Speaker-to-LiveJournal" role. But we can't get a definite price for Kindle books, nor an official publisher, or page count (although "file-size" might be an option if we update the software). It's a can of worms I don't think we should open just yet. BLongley 03:16, 6 September 2011 (UTC)

I agree with Bill. As one of the few persons who actually handles Fixer submissions, or at least it seems that way at times, I'm not ready to handle ebooks. And I don't want to navigate through a queue trying to avoid ebook submissions. If any moderator objects to that, I've got one suggestion: start working on Fixer submissions. Mhhutchins 03:31, 6 September 2011 (UTC)

I also agree with a pause. I cautiously try to "do" some Fixer submissions (refraining from my first urge to reject most of them on basis of insufficient data on the book content) and I'm not that satisfied with e-books and the bibliographical problems they entail. Hauck 05:25, 6 September 2011 (UTC)

Fair enough. I will have better numbers when the download finishes, but it's going to be huge, well into the six figure range. I may be able to pare it down some by linking certain e-books to their dead tree counterparts and then auto-suspending vanity publishers and such, but even so it will be a mountain of data. Still, it's better to have this data sitting in one of Fixer's queues rather than to lose access to it, possibly forever. Ahasuerus 17:31, 6 September 2011 (UTC)

Any idea how many of them have ISBNs? Those might be worth getting into the database.... --MartyD 00:38, 7 September 2011 (UTC)

I think starting with ebooks with ISBNs would be a good move, when we're ready. Those are easiest to double-check against sources. And I'd definitely prefer the "Add Publication to Existing Title" ones rather than adding a new Title. BLongley 01:04, 7 September 2011 (UTC)

But I'm still concerned about Publishers: when I check a Kindle Book, I can normally find it on Amazon US and UK, but with different "Sold By" details and no publisher details. And Prices: those vary over time and by where you try and buy it from. And Dates: some sources give first print dates and not first ebook dates. And Binding - sometimes we get one ebook with several specific formats. I suspect we'll have to make some changes to "Bibliographical Warnings" and lots of clarification of help before we could cope with a mass input of ebooks. BLongley 01:04, 7 September 2011 (UTC)

Has anyone noticed that Amazon is replacing their "Look-Inside"s with just the Kindle version, which gives so little info that it's worthless, as appealing as a come-on from a Vegas street walker. It's nothing more than those "previews" (read: ads) that Baen and other publishers print in the backs of their paperbacks. Mhhutchins 01:37, 7 September 2011 (UTC)

Yes, in checking some recent submissions I've noticed that. If Amazon are going to become less useful and more Kindle-promoting I'd be happy to drop them as a Fixer source and go back to library data. And I say that as a Kindle-Owner and Amazon customer - but as a bibliographer I don't like the changes. BLongley 02:42, 7 September 2011 (UTC)

My understanding of the ebook has shifted over the last several years..... I now believe that ebooks are going to change our definition of 'publication'. We will still have Titles, Authors, Series, etc... but publications will become more fluid. Many of the key identifiers of a 'publication' as we now define it are evaporating... Price, binding/format, and date being the most obvious ones... but also the Author is going to become the publisher, the store might define the format, and the date of purchase might define the price... all other things remaining the same. Kevin 02:28, 7 September 2011 (UTC)

Yes, It'll left us with a very few "sure" fields (mainly the three you cited). Perhaps it's an heresy, but shouldn't we envision a brand new category "EBOOK" at the same level as "NOVEL" or "NONFICTION" (intellectually a bit like the infamous CHAPTERBOOK )? Hauck 04:33, 7 September 2011 (UTC)

I think it's less about a new type of container, and more about blurring the lines between the different publications... ebooks have almost reached the point where they accompany most / all fiction titles. A published title (Book, or story) will eventually or automatically have an accompanying electronic form. It's becoming both redundant, and important at the same time. Kevin 11:37, 7 September 2011 (UTC)

Yes, I can believe it. And I'm willing to help us change to match new practices - price is going to be too variable to record, and page-count may well be replaced with download sizes. Formats may proliferate too much for me to want to moderate them. Distinguishing ebook formats with different IDs is going to be tricky if they move away from ISBNs. I can see myself becoming an old fogey that still likes physical books even if I have to read them with spectacles (which Star Trek movie had William Shatner admitting that?) but at the moment I can still understand both sides. And I'm now even more likely to be given an ebook than a paper one - I'm not soliciting bribes, but the occasional "thank-you" reward tends to be more electronic than paper these days. BLongley 03:01, 7 September 2011 (UTC)

(unindent) Re: ISBNs. Fixer has been submitting e-books with ISBNs for some time now. The changes that I had to make to Fixer were just for e-books with no ISBNs. That was the main reason why this round of changes was non-trivial -- "ISBN" is the primary key in most of Fixer's tables. Which, in retrospect, wasn't a very smart thing to do, but... Ahasuerus 21:31, 7 September 2011 (UTC)

Fixer: November 2011

Fixer is in the process of submitting November 2011, but there is no hurry, so the batches will be smallish.

In other news, Fixer now has 85,522 books with no ISBNs sitting in his internal queue. Most of them are Kindles and other types of e-books, but there are also paper books with no ISBN or at least no ISBN that Amazon is aware of. In many cases it appears to be a deficiency in Amazon's database, so we'll have to take a closer look at some point. Again, no hurry.

Now that we have grabbed everything that we needed from Amazon, I am trying to change the data capturing software so that it would still work going forward. It's a little tricky since Amazon has added numerous constraints, but it should be doable. Ahasuerus 18:43, 11 September 2011 (UTC)

US submitted. Ahasuerus 21:09, 17 September 2011 (UTC)

Unfortunately, the Amazon changes are more devious than I first thought. They are purposefully obfuscating many of their Web pages and rolling back the API to make it harder to get certain types of data. After reviewing other people's workarounds and checking alternative sources of information, I think I have a way to get what we need, but it's another headache. Ahasuerus 07:59, 18 September 2011 (UTC)

UK submitted. Still fighting Amazon. Ahasuerus 01:19, 21 September 2011 (UTC)

Multi-way borderline submission

I accepted Die Konstitution des Wunderbaren after seeing it sit for a while. I realize it's at least doubly-borderline: German with no apparent English translation and with film as a center. Let me know if that should have been "out". --MartyD 03:07, 17 September 2011 (UTC)

English translation of the title: "The constitution of the Miraculous: To a poetics of the science fiction film". I did a search of non-fiction books about films. Excluding those about film and literature, but including those about film, film & TV, film & comics, etc., I found 181 books about Science Fiction, Fantasy, and Horror films in the ISFDB. So that boat sailed a long time ago. We obviously are including such books, so that's not exactly a "borderline". I'll let others decide whether the "German only" is enough to exclude it. Our rules say yes, but we've been including an awful lot of non-English stuff lately, not all of which has ever been translated into English. Chavey 05:43, 17 September 2011 (UTC)

I confess to have added lots of such stuff myself, but, to my defence, the books that I entered are about written SF. More broadly, I think that the desirable future for the ISFDB is to became a multi-language reference tool. Hauck 10:33, 18 September 2011 (UTC)

I for one have no problems with such NONFICTION additions. And I am supportive of the "become multi-language" idea - we seem to have been a bit arrogant in claiming to be "THE Internet Speculative Fiction Database" when we were clearly so English-Language biased. It's going to take more than my bumbling incremental improvements to get us there, but I hope we will. BLongley 23:59, 18 September 2011 (UTC)

The Rules of Acquisition (see Policy) currently say "In - Works about speculative fiction", but do not clarify whether it's just written SF or SF in other media as well. As a practical matter, it's hard to tell whether a book about SF films/TV spends a significant amount of time discussing written SF. At one point, I used Amazon's Look-inside to spot checked a few randomly selected books about SF films and in most cases there was some cross-pollination, e.g. a book about Star Wars would have a section on the pulp origins of space opera and an article about _Blade Runner_ discussed Dick's novel. We may want to move this discussion to the Rules page to see if we want to update the Policy. Ahasuerus 21:20, 17 September 2011 (UTC)

I think we've allowed too many in already to stop the flow, even if we wanted to - e.g. this came in today and I had no problem with accepting it. A lot were probably added due to being reviewed in Genre Magazines or Fanzines. I think the line now is "Reviews of Films" - OUT (or only included as Essays) but "Books about SF Films" - IN. BLongley 22:50, 17 September 2011 (UTC)

Aside: isn't it about time we had some more German-speaking Moderators? We only have four European Moderators, and I only grudgingly admit to being technically in Europe - I've never set foot on the main continent. I'm by no means multi-lingual like the others, and frankly some Editors would be more useful as Moderators in European Languages than I could ever be. Of course, this means that I'm not that qualified to tell which Editors are over-due for promotion either - I suspect Stonecreek and Dirk P Broer are getting there but often I can't tell if they're doing things right due to my Englishness. :-/ BLongley 23:10, 17 September 2011 (UTC)

I agree Stonecreek is getting close. I've had no problems with his submissions lately (he even corrects me now and then), and he communicates well. I think Dirk must first learn to ask questions and read documentation before submitting questionable edits. He does do a lot of valuable work on authors though, and will certainly qualify someday. I personally think Chavey should be the first candidate for moderatorship at the moment. --Willem H. 09:57, 18 September 2011 (UTC)

Except that I think Stonecreek has gotten there (here?), I agree with Willem's analysis. --MartyD 12:37, 18 September 2011 (UTC)

Well, as all of the Editors discussed seem to be too modest to put themselves forward for Moderator (maybe they know the new headaches such will cause?) perhaps someone will at least ask them if they want the job. I have had Darrah Chavey down as a future mod for some time, he certainly passes on all the good intentions and communication requirements. It's been a while since we had a new Mod, and those new Mods may not have even known how they got there, so I'll mention Moderator_Qualifications again. BLongley 23:50, 18 September 2011 (UTC)

Well, I finally sat down and read through the Moderator Qualifications and the Moderator Helpscreen pages, and I am willing to stand for nomination as moderator. I am a bit concerned that I still make too many mistakes, but I think I've gotten to the point where I'm correcting more existing mistakes in the data than I am introducing new mistakes :-) I am hoping, though, that you will allow me a couple of requests:

1. Occasionally, I think I would rather not moderate one of my own submissions, so that at least one other moderator looks at it to verify that they agree with me;
2. The moderator help screen convinced me that I should start approving submissions somewhat slowly, doing a couple of types of submissions first before I try to learn the details of other types of research/verification - so I assume it's ok if I begin by cherry-picking the submissions I look at;
3. I never want to look at another Daisy Meadows submission. I think 557 submissions on one series should be quite enough for a lifetime. (And I suspect the moderators who had to approve those will agree.)

Chavey 02:41, 23 September 2011 (UTC)

Thanks for accepting. The Moderator Qualifications say "A moderator has no obligations to do any minimum tasks", so it's entirely up to you what you want to do (there is some mental pressure of course). I can imagine you've seen enough fairy submissions by now, but occasionally Fixer finds a new bunch of vampire shaggers, also lots of fun. To be serious, moderators are allowed to ask questions, and if you want someone else to look at some of your submissions, just leave them in the queue and ask on the moderator noticeboard. I'll start the nomination process. --Willem H. 12:49, 23 September 2011 (UTC)

I think you're doing FAR more corrections than you are introducing errors - I wouldn't support the nomination otherwise. As to your other points - yes, you can still ask for a double-check from another Moderator on any edits you're unsure of: yes, "cherry-picking" is fine, that reduces the workload on the rest of us. We can't promise to protect you from "Daisy Meadows" submissions unless Ahasuerus stops Fixer from submitting them - I don't think any human actually submits such. BLongley 02:25, 24 September 2011 (UTC)

Shall we make it a double nomination then? I would be willing to start things up, I've had some pleasant experiences with Darrah Chavey, and seen a lot of Christian (Stonecreek)'s submissions. --Willem H. 12:36, 20 September 2011 (UTC)

No opposition I suppose. I'll ask them both. --Willem H. 19:31, 21 September 2011 (UTC)

Foreign language variants

Are we starting to accept submissions making records of translations into foreign languages as variants of the original English title's record? The previous policy was to merge them with the English title. Is implementation of foreign language support far enough along to change the policy? Mhhutchins 18:26, 22 September 2011 (UTC)

The policy hasn't been updated, and there's a lot of language support improvements still to go - I think Ahasuerus is making sure we continue to get more Fixer submissions first. :-/ I tend to allow (but not encourage) such for now just to see how it turns out - it's probably not harmful for obscure stuff, but I can appreciate that we still need to do a LOT of display work before we can recommend, for instance, that we rework all Asimov translations. We need user language preferences to be applied before we can really open it up. BLongley 18:43, 22 September 2011 (UTC)

My idea was to make 'Sam, of the Pluterdag' into a sort of example. I certainly do not want to have to do all editing for the umpteen translations of a given Asimov title. Paul van Herck has not written much, and is reasonably good translated. --Dirk P Broer 23:12, 22 September 2011 (UTC)

It now looks like a very good example, thanks! I don't think we can recommend such for all titles yet - maybe we can start with non-English titles translated into other languages first, then English titles translated into other languages? Sorry to sound so English-centric, but that's the way the ISFDB started and how the Wiki discussions still work. I can foresee the day when I can leave this moderating job and leave it up to the newer multi-lingual staff. Or the coding job - I may be competent in about 20 programming languages over my working career, but this Python and MySQL stuff has never helped pay my bills, just improved my self-esteem when I occasionally do something that's appreciated. BLongley 02:47, 24 September 2011 (UTC)

Fixer update - September 2011

OK, I believe I finally figured out how to keep Fixer going after October. It won't be very pretty, but it should work for our purposes.

In the process of sorting out various related issues I learned many new things, e.g.:

• Powell's has an experimental API
• Barnes and Noble has a new layout which is leveraged by Caliber plugins,
• how the Fantastic Fiction database engine works
• how to get iMacros to do what you want it to do and so forth.

There are numerous implications for what we do here, e.g. it turns out that many Nook books sold by Barnes and Noble have ISBNs and would be fairly easy to create automated submissions for, but first I need to catch up with the current backlog of software changes. Ahasuerus 01:01, 24 September 2011 (UTC)

Great news! :-) If Fixer can still submit stuff (or Bibliovore can start) then we'll have some examples for the next two new Mods to try out. (Yes, I'm assuming, but each candidate seems to be doing well so far.) And if you can then catch up on the current backlog of software changes I can give you some more to look at! ;-) BLongley 02:01, 24 September 2011 (UTC)

The new algorithm has been activated. One difference that you will notice is that the ISBN identification process has changed: when Fixer finds a new ISBN, he now also looks for related ISBNs, so you may see older books pop up in the queue. In most cases it's harmless, but Amazon may have updated prices for certain older ISBNs, so please be on the lookout for that. Other than that everything should be familiar. If you run into any bugs, please report them here. Ahasuerus 06:50, 26 September 2011 (UTC)

To be honest I didn't notice much of a difference apart from there being more than usual missing publishers. ("Paw Prints" titles again.) BLongley 18:49, 26 September 2011 (UTC)

That's exactly the idea! :-) The goal was to keep Fixer's submissions going even after October 26 when Amazon expects to put the new API rules in place. We won't know for sure whether Fixer's new logic will work flawlessly until October 18 when they temporarily activate the new API for 4 hours, but it looks promising. Ahasuerus 18:56, 28 September 2011 (UTC)

I'm not sure how many Mods would miss Fixer if he retired, BLongley 02:18, 29 September 2011 (UTC)

Fixer-generated submissions merely reflect the current state of the field. If anything, their average quality is higher than what's out there because Fixer automatically suspends most vanity publishers. Granted, we still have to deal with more paranormal romance and YA "extruded product" than most of us would prefer, but what can we do if that's what sells? Ahasuerus 03:15, 29 September 2011 (UTC)

but if he can keep going then it's good practice for our latest Mod recruits. Can you arrange a special run for all the German translations of Daisy Meadows books to give them some practice? ;-) BLongley 02:18, 29 September 2011 (UTC)

At the moment Fixer auto-suspends non-English books (or at least tries to), but that's because all of Fixer's sources are US- or UK-based [1] and the quality of their foreign language records is, more often than not, dismal. I could add the French National Library as a source, but that would require even more man-hours which we don't have at the moment.

[1] I am also working on adding Australia, but it's not ready yet. There is just too much stuff going on, from answering e-mails about ISFDB to operational issues and development/testing. I am getting increasingly burned out, but I don't see any alternatives. Ahasuerus 03:15, 29 September 2011 (UTC)

Rithebard's edits

I am working on the software today and just happened to notice that the last round of User:Rithebard's uploads looks peculiar. Is anyone familiar with her edit history? Ahasuerus 20:16, 25 September 2011 (UTC)

Same thing here, I've been close to delete the three "images" as I don't undestand their purpose (add to the fact that all this looks like self-publicity) but I refrained from it. Hauck 20:21, 25 September 2011 (UTC)

I've deleted the advertising banner and left a message on the editor's talk page. The other two images remain until I can determine if he (or she) wanted to replace the Amazon image that is currently linked to the pub record. Mhhutchins 05:34, 26 September 2011 (UTC)

New Worlds 4

Does anybody have any idea how the cover artist got credited in this pub? I only verified it, but didn't do the initial edit. I can't find any credit for him in the book, and it seems like the only art by him. Perhaps there is a reliable source? Stonecreek 15:05, 27 September 2011 (UTC) Oh, I see now that it was also verified by Unapersson. I'll try and get a response from him. Stonecreek 15:11, 27 September 2011 (UTC)

Source is probably the 1995 British Science Fiction Association Awards. The style is certainly like him (see here). There are other sites that credit Szumowski like SFcovers.net and Librarything. I think the credit should stay, with notes about the (secondary) source of course. --Willem H. 15:34, 27 September 2011 (UTC)

And someone should link the Award to the title. (We're well behind on Artwork awards). BLongley 16:01, 27 September 2011 (UTC)

I had no idea how to do this, but tried anyway. Please check the result. --Willem H. 16:40, 27 September 2011 (UTC)

Yes, that's right. I'm not sure how many awards still need fixing, it might be worth me developing a quicker method than add/delete. BLongley 19:30, 27 September 2011 (UTC)

I first tried to link the title record to the award on the "editaward" screen. That's more intuitive I think, but it didn't work (too dangerous?). Then I tried it the other way around ("Add Award Information to This Title"). That went better, but I then had to remove the original award. I'll try some more one of these days. --Willem H. 19:54, 27 September 2011 (UTC)

I'm thinking more along the lines of "Link Review to Title" and creating a "Link Award to Title" equivalent. But there is a severe backlog on development at the moment, including the area of awards, so we might well have worked-around most of the problems before such an improvement goes live. There will still be issues of awards to Publishers and suchlike to deal with, so I'm not yet running out of ideas for improvements yet! BLongley 00:22, 28 September 2011 (UTC)

Moderating Fixer submissions

When Fixer finds "To be published" titles, they sometimes have incomplete data, e.g. Amazon does not yet have all the details about the book. Do we try to keep track of such books to revisit them after publication, when Amazon has more data? Or is that missing data handled in some other way? Chavey 11:42, 29 September 2011 (UTC)

If it's too vague to approve, or you can't tell if it's Spec-Fic, add it to User:Fixer#Deferred_Submissions, otherwise let it in. We can cope with, for instance, a missing tp/pb binding for a while. I think Dissembler used to go back and revisit such but Fixer doesn't. I tend to have a look at Forthcoming Books every so often and check ones about to be published, but it's not a regular thing. BLongley 13:50, 29 September 2011 (UTC)

Thanks. The one I was looking at couldn't distinguish between pb/tp, and didn't have the "Look Inside" yet to look for a cover author. So I'll follow your advice. Chavey 20:07, 29 September 2011 (UTC)

It's also a question of getting used to Amazon's quirks. For example, if both Amazon records (US and UK) for a commercially produced book lack cover scans, it suggests that the book has been canceled or postponed. However, if the publisher is a small press, it's possible that it failed to send the scan to Amazon. And if the book is marked as "library binding", then chances are that Amazon will never get a scan. Of course, if a scan exists but shows a different title, then more research is needed.

Also keep in mind that in the past Fixer re-checked various months on a semi-continuous basis. I typically asked Fixer to download all genre books about 6-8 weeks before the months starts, at which point we got a bunch of books by major publishers and some small press records. Subsequent runs for the same month were less likely to find new books from major publishers, but more likely to find books from small presses and other oddballs. This is about to change since the Amazon API will change in late October, so we will have a more even mix of major and minor publishers. Nonetheless, we are still likely to see clusters of books from certain publishers appearing all at the same time. Ahasuerus 23:30, 29 September 2011 (UTC)

Red Exclamations, Moderator Help Screen

When I became a moderator, I suddenly began seeing red ! exclamation marks in various places that I didn't understand. They were not mentioned in the Moderator Help Screen, but with some research I discovered a 2009 posting that helped explain them. As with that posting, I suggest adding a note about this to the Moderator Help Screen (which I've been working through in modest-sized chunks :-). More specifically, I have suggested a paragraph to add to the Moderators and the ISFDB Wiki section. I hope it's accurate. The last sentence may be unnecessary. Chavey 20:22, 29 September 2011 (UTC)

I have no problems with this addition. And thanks for working through the help - it's not immutable, and probably needs updating due to all sorts of software improvements and policy changes since they were created. Carry on! BLongley 20:29, 30 September 2011 (UTC)

Noting Sources made easy Attempt 3 part 3

Following on from ISFDB:Community_Portal#Noting_Sources_made_easy_-_Attempt_3 - here's the first stab at what the Moderator will see.

Note the additional notes that were automatically added due to selecting the right checkboxes, and the two new messages indicating the primary sourcing that will lead to verification. (Those should probably be mutually exclusive.) I'm already thinking that we could format the messages better (maybe rolling them into the Notes to Moderator in the same way I rolled the Primary and OCLC sources into normal notes). And maybe we should show when they've checked "Other website (please state which in Notes)", "Later printing or edition (please state which in Notes)" or "Other source (please state which in Notes)" just in case they didn't state it in notes. BLongley 17:41, 30 September 2011 (UTC)

Those are my initial thoughts, I'd like to hear any others from people that will have to deal with this. BLongley 17:41, 30 September 2011 (UTC)

I like where you are going. In reference to previous comments... I would like to see this implemented for all edit type (which address a publication)... and only the 'auto-verification' held back when the publication isn't 'new'. And yes... I would like to see when they've claimed a website or other source but haven't provided it. (I think that will help us be consistent in guiding newer editors. Kevin 17:48, 30 September 2011 (UTC)

Ok, I'll dummy up a demo of them claiming sources they haven't noted. BLongley 18:27, 30 September 2011 (UTC)

And yes, this can be rolled out to all types of publication edit, I just want to get one right first. There's normally 4 steps to any particular edit: the data entry screen (which I've shown you), the submission part (which I have shown you but don't expect anyone to really be interested in), the moderator screen above, and the approval part (which I will show you when I get to it, but again I don't really expect many people to have views on the display, only on what it actually does.) If I can get all 4 stages right for Add Pub, then doing the same for Clone Pub and New Pub should be trivial: and even Edit Pub shouldn't be too bad, although we will have to decide on how to treat already Primary verified ones. BLongley 18:27, 30 September 2011 (UTC)

After that, we can turn to Sourcing for other types of edit: but I'll start those discussions separately. BLongley 18:27, 30 September 2011 (UTC)

(Unindent) Second attempt, slightly less shouty: BLongley 22:10, 1 October 2011 (UTC)

Another example:

which will lead to (the bit most of you don't read)

And finally

I think I've cracked most of the coding problems, unless people want to enforce some checking earlier in the process rather than asking a Mod to query such situations as both Primary and Transient being checked. As usual - comments welcome. I'm sure this will not meet with 100% approval, nothing ever does, but if it goes over 50% I'll see about adding it to Ahasuerus' ever-increasing workload. BLongley 00:00, 2 October 2011 (UTC)

I like what you've done so far, and that you've taken into account how the software will display and function throughout the various steps from submission to acceptance. I'm assuming this is only for "Enter New [pub]" and "Add Publication to This Title" functions, and that the display for the "Edit This Pub" function will be different when it comes to noting sources (taking into account that those pubs may already have been verified). Mhhutchins 18:07, 2 October 2011 (UTC)

I started with "Add Publication to This Title" for this demonstration, and have tested it with "Clone This Pub" too. If people like this it's easy to roll out those changes without affecting anything else. I can submit such changes for "Add New [pub]" as well, but there's an outstanding change there (to allow users to choose a default language that will be used when creating new publications) - so I'd like to know whether this can go on top of those changes or whether it should go in before. Only Ahasuerus can tell me that. BLongley 18:41, 2 October 2011 (UTC)

I think "Edit This Pub" will indeed function slightly differently, except maybe for the case of unverified pubs with no existing notes. I'd be reluctant to allow "lazy sourcing" for anything already verified without a separate discussion. Maybe we can implement this like Moderator Notes - try a few sample submission types first and roll it out to more if people find it useful? BLongley 18:41, 2 October 2011 (UTC)

I also see that it will still be up to the moderator to make sure that when certain boxes are checked that the editor has completed the process, i.e. recording the OCLC number, the website, and other printings in the note field. I'm willing to take on this duty and hope that other moderators feel the same. Without this check by the moderators I can see the whole process falling apart. Mhhutchins 18:07, 2 October 2011 (UTC)

Well, we can introduce checks - e.g. demanding an OCLC number if you claim you're using OCLC, and make the Primary and Transient checkboxes mutually exclusive (even I don't choose both deliberately, even when I have spare copies that I'm trying to get rid of). As always, the balance to find is between ease of entry and accuracy, and the division of work between editor and moderator and the software - I'm sure none of us want us to move to a purely Fixer-populated database, but neither do we want serious bibliographers to give up on us totally. I hope that moderators will use the new information wisely and not blindly. BLongley 19:59, 2 October 2011 (UTC)

I agree with most of the above, but I'd like to review it again later because I am badly sleep-deprived at the moment. As far as validation goes, we'll probably want to do it via pop-ups. Have you had a chance to review the JavaScript-based checks that I added to Edit Title last year, Bill? They are hopefully fairly straightforward and may serve as a template for similar checks in Add/Edit Pub. If not, I can code them myself. Ahasuerus 03:01, 6 October 2011 (UTC)

I'll have a look. I used to understand some Javascript, but got a bit fed up having to test it on multiple browsers. BLongley 18:40, 6 October 2011 (UTC)

It's not too bad if you stick to the basics, typically just JavaScript 1.2. Ahasuerus 02:15, 15 October 2011 (UTC)

Fixer follies - October 2011

Fixer is almost halfway through his first run using a new and improved (or at least "made necessary by Amazon's changes") algorithm. It's a mixtures of late 2011 editions and some older books that got picked up because of the new logic that Fixer uses to find ISBNs. Basically, when a book is reprinted (either on paper or as an ebook), Fixer will simultaneously submit other missing ISBNs for the same book. It's somewhat inconsistent because of the way Amazon works, but at least it doesn't seem to be breaking anything.

Unfortunately, the new process is significantly more labor intensive on my side, so I am still tweaking and automating it, but we are getting there.

P.S. Fixer is still not ready to submit Kindle books without ISBNs, but they are all captured and warehoused in one of Fixer's internal queues. Ahasuerus 00:08, 4 October 2011 (UTC)

We are 80%+ of the way there. Ahasuerus 22:42, 8 October 2011 (UTC)

US submitted, UK in progress. Not a whole lot of UK books in this batch since it was a cleanup pass and there are fewer small presses and such in the UK. Ahasuerus 18:29, 9 October 2011 (UTC)

UK submitted. Ahasuerus 23:02, 9 October 2011 (UTC)

Fantastic Fiction cover images

We have several cover images from Fantastic Fiction, which are normally cited with the note "Cover art supplied by Fantastic Fiction". They have recently changed how they store their images though, moving them from "www.fantasticfiction.co.uk/images" to "img1.fantasticfiction.co.uk/images/". This doesn't actually break our links, because they're using a redirect on the older links to take them to the new location. However, editors adding such pictures in the future will link directly to the "img1..." picture and put that in. Due to the details of our coding, any such images will not include the citation crediting the source. (See Riddle of Stars for an example.) It would be a good thing if someone could update the code to correctly respond to that change. Chavey 04:01, 5 October 2011 (UTC)

Good thing done. See FR 3419096. BLongley 18:13, 5 October 2011 (UTC)

Thank you. I should have thought of submitting it as a FR, but I appreciate your jumping on it as is. Chavey 13:53, 6 October 2011 (UTC)

No problem - it was a trivial change. The administration work was longer than the code changes! Now we just need Ahasuerus to catch up on sleep. BLongley 18:36, 6 October 2011 (UTC)

The deed is done. Ahasuerus 03:40, 9 October 2011 (UTC)

"Winter Warriors", by David Gemmell

Could someone check whether I've handled the Winter Warriors Preview Chapter correctly? It had previously been listed with the title Winter Warriors. However, this particular publication was published for Novacon 26 as a 17 page preview chapter for the 300+ page novel released a year later. So I unmerged it, stuck it in a chapbook, and added it to The Drenai Saga title series. Was this the correct resolution? Chavey 23:27, 10 October 2011 (UTC)

You'll have to add a chapterbook title record to the pub record. (Look at the pub record and you'll see that there is no link for the "title reference" (read: title record). Choose "Edit This Pub, then under contents choose "Add Title" and in the new fields that open add the book's title, change type to CHAPTERBOOK, add the author, then submit and accept. Mhhutchins 00:46, 11 October 2011 (UTC)

Thanks! I had made several of those Chapterbook errors in filling out the Novacon publication series, and that helped me fix them all. Chavey 02:54, 11 October 2011 (UTC)

Great. Chapterbooks are tricky and it takes awhile to get the hang of them. A quick look at the pub record will usually tell you if it's OK. There must be two elements: a title reference (which links to the CHAPTERBOOK title record) and a content record (usually a SHORTFICTION record). If one or both is not present, it's easy to add the missing element. Mhhutchins 03:09, 11 October 2011 (UTC)

Fixer holds

I've placed holds on 5 books from Fixer in the queue. I find this category of books offensive. These come from a company that prints off Project Gutenberg books (and Wikipedia articles), throws cheap covers on them, and sells them as something new. These are nothing more than a print-out from Gutenberg. Do we have to accept them? (Yes, they are spec fic.) Chavey 05:03, 11 October 2011 (UTC)

Even if you're right in finding these ways offensive or unethical, I think that we can't (and mustn't) participate in any form of censorship (on the lighter side, I'm sure that we already have some not-very-legit publications). The best you can do is not to enter or validate them. Hauck 05:33, 11 October 2011 (UTC)

We do list unauthorized editions like the Ace version of The Lord of the Rings, the Lancer version of Elric of Melniboné (aka The Dreaming City), pirated Asian reprints (which I can't find at the moment), sequels-by-other-hands apparently done without the original author's permission, etc.

That said, Fixer typically auto-suspends all books by this particular publisher, General Books LLC, simply because they are lower priority and our resources are limited. These five snuck in since I am desperately trying to adapt Fixer to work with Amazon after the looming API change 2 weeks from now. Which involves leveraging other resources and doing other headache-inducing things. Ahasuerus 05:42, 11 October 2011 (UTC)

I'm afraid we do have to accept them - you don't personally have to though. I don't think we've ever fallen into "censorship through inactivity" but we've come close at times - I don't think any of the moderators are actually interested in very juvenile SF for instance. BLongley 17:30, 11 October 2011 (UTC)

I accepted the submissions, changing them to chapterbooks, adding a shortfiction content record, and then merging them with the pre-existing title records. No judgement based on the quality of the work or the ethics of the publisher crossed my mind. Mhhutchins 18:24, 11 October 2011 (UTC)

SF Birthday Calendar

We have a current submission to add a date of birth for Sharman DiVono. I don't feel comfortable putting a hold on it, because it seems like we're still debating issues of requesting/demanding documentation for such stuff. The only online documentation I can find to support that submission is an online SF Birthday Calendar. Is anyone familiar enough with that site to know whether we should accept it? If so, it looks like there might be a wealth of data there. If not, then I would be more hesitant about this submission. Chavey 14:26, 11 October 2011 (UTC)

I see that the source is documented now. I can't say how reliable the source is - Steven H Silver is quite well known, but we do disagree on some dates - e.g. we're one day apart on Thomas Burnett Swann. BLongley 17:20, 11 October 2011 (UTC)

We also disagree about the use of periods after initialized names. :) Mhhutchins 18:26, 11 October 2011 (UTC)

To be fair, he claims "H" is not an initial, it is his full middle name. But we do seem to have put him off for good. :-( BLongley 18:46, 11 October 2011 (UTC)

Fixer's new finds

Fixer is slowly adjusting to the new algorithm and appears to be ready for the Amazon changes that will take effect in 12 days. We'll know for sure on the 18th when Amazon briefly turns the new API on to let us test.

One of the side effects of the new algorithm is that Fixer is finding ISBNs that were not readily available in the past. If you check the queue, you will find a number of items published by "ProQuest, UMI Dissertation Publishing", which publishes exactly what its title suggests. The only vaguely applicable comment in the Rules of Acquisition says "Debatable - Academia-produced magazines. Can we realistically compete with, say, the SFRD?"

My thinking is that these are incarnated as bound books and, as long as they are about SF, they should be included. If there are doubts about it, we may want to open a conversation on the Rules page. Ahasuerus 06:21, 14 October 2011 (UTC)

People seem to disagree: ISFDB:Community_Portal#Dissertations. BLongley 19:24, 14 October 2011 (UTC)

Based on the Community Portal discussion, we have green light to reject these submissions. Fixer has been updated to automatically suspend this publisher. Ahasuerus 02:12, 15 October 2011 (UTC)

Moonlight Tuber

We currently include issue #1 of this zine, and have submissions to include the next two issues. Moonlight Tuber is a downloadable .pdf publication. Our Rules of Acquisition say that items are "In" if they are:

In - downloadable e-zines (periodicals in electronic format) which have been assigned an industry standard registration identifier (ISSN or ISBN) or have a history of containing reliably stable contents.

This e-zine does not have an ISSN or ISBN, and doesn't really have a "history", since the first issue came out in May, 2010. So it seems that this should be "Out". Is that correct? Chavey 01:44, 16 October 2011 (UTC)

It's downloadable (PDF), not just a website, and it has been going for over a year. My inclusionist tendencies make me think it is "In" enough.... --MartyD 11:18, 16 October 2011 (UTC)

Yeah, I'd let it in. (I think I might have let the first in anyway). Desmond Warzel may have started with a bit of self-promotion but he's a reasonably good editor now. BLongley 23:10, 16 October 2011 (UTC)

Armchair Fiction Double Novel

Just a heads up that many "Armchair Fiction Double Novels" reprint novellas even though they are advertised as "complete novels" on the cover. You can't blame the publisher for trying to preserve the look and feel of Ace Doubles, but technically these books are collections rather than omnibuses. Ahasuerus 06:27, 18 October 2011 (UTC)

More often Anthologies, as there are multiple authors. I haven't messed with the lengths but this looks especially suspicious. BLongley 17:24, 18 October 2011 (UTC)

Right, anthologies. Sorry about that, it was 2 in the morning... Ahasuerus 20:37, 18 October 2011 (UTC)

Understandable. It's 2:34 in the morning here, and I'm torn between doing the latest "Masters of Science Fiction" entries before I go to bed, or revisiting them in the morning and seeing whether they've been completed, or coding something to make such easier. BLongley 01:37, 19 October 2011 (UTC)

Also, in case you haven't see it yet, there is a publisher-provided list of all AF doubles. Ahasuerus 06:27, 18 October 2011 (UTC)

It seems we have a big gap in their series. Does Fixer have the rest or should they be manually added? BLongley 17:24, 18 October 2011 (UTC)

Fixer had 21 "Armchair fiction" ISBNs in his internal queue ready to go, but then I ran an explicit search for "Armchair fiction" and found an additional 18 books at Amazon. All have been submitted now, but it's disconcerting that so many were not found by the "discovery" program that was run in September. I suspect that Armchair Fiction only recently uploaded their catalog to Amazon, but I don't know that for sure.

In other news, today was a big day for Fixer. Amazon turned the new algorithm on for 3 hours and Fixer was able to run successfully under the new rules. So the good news is that we will be able to continue getting forthcoming books from our lovable digital friend (at least until Amazon messes with things again.) The bad news is that the new restrictions are fairly onerous and make the whole process more time consuming, but there is not much we can do about it. Ahasuerus 20:37, 18 October 2011 (UTC)

Congratulations on Fixer's success! It's too bad that it will be more time consuming, but maybe that will improve with experience. But, in my limited experience, Fixer is crucial to staying on top of new works. (And, often, discovering older works missed the first time around.) Chavey 00:44, 19 October 2011 (UTC)

Thanks, you made Fixer's day! :-)

Fixer is a hard working robot with many thousands of records in his internal queues. They are ready to be submitted, but the main limiting factor is how quickly these submissions can be approved. Experience shows that having 200 Fixer submissions in the main ISFDB queue can have a discouraging effect on moderators, so I try to keep Fixer runs more manageable and wait for old Fixer submissions to be processed before I send more stuff over. This in turn forces Fixer to prioritize ISBNs -- audio CDs, children's chapterbooks, known self-published stuff, etc get auto-suspended -- which can be a tricky and error-prone process. It's a balancing act, but we are getting there. Ahasuerus 04:37, 19 October 2011 (UTC)

I just accepted Fixer's submissions of two Armchair Fiction Double Novels, using the standard format of entering double novels: Title A / Title B. I noticed that Fixer has given them as Title A & Title B (undoubtedly based on Amazon's listing) and corrected them. But earlier ones were accepted in this format. Looking at the Amazon "Look-Inside" I see that each work has their own title page (though not bound tête-bêche like Ace Doubles), so the slash format should be used. Right? Mhhutchins 00:59, 19 October 2011 (UTC)

I don't recall us ever having discussed multiple title-pages before, but "space slash space" is OK by me. (Not spacing seems to make us less visible to search engines.) BLongley 01:19, 19 October 2011 (UTC)

It was discussed several years back when editors were trying to standardize the entry for Ace Doubles. I suppose the reason for going with the slash and not the ampersand is that there are two title pages. The use of the ampersand might imply a single title page. I'll correct the pub records and their title records when I get a chance. Mhhutchins 04:28, 19 October 2011 (UTC)

Declaration of Conflict of Interest

As requested, I've submitted changes that allow a "Smashwords" link in "Other Sites". I created and used my own affiliate ID while testing it, so if implemented as it stands I could actually make a little money from it if people use us to buy books from there. (I doubt we'll even get to the $10 minimum to justify a payment for several years, given our limited support for ebooks at the moment, but it's a possibility.) I'm declaring it now as we could just as easily use a central "isfdb" affiliate ID (if someone is willing to manage it) and use it to defray the hosting costs, or I could direct any payments from Smashwords on my ID to someone with a Paypal Account. (A US resident would be the best choice as for a foreigner like me they'd take 30% off in taxes unless I jump through several paperwork hoops.) Comments please? BLongley 17:03, 27 October 2011 (UTC)

Also, I realised while doing this that our new ability to put links in title records (and the existing ability to put such links in any sort of notes) means that people could add their own affiliate IDs to any link they provide. Should we allow, ban, encourage or discourage this? Allowing it might encourage editors but might annoy other people doing the work for free. Banning such means we need to teach Moderators how to spot such. Again, comments please? BLongley 17:03, 27 October 2011 (UTC)

I am afraid that's above my pay grade. Let me see if I can drag Al into this discussion... Ahasuerus 21:14, 27 October 2011 (UTC)

"Pay grade"? You seem to be the only editor with a negative one (paying for the hosting as you do, rather than "just" donating many hours of labour like the rest of us). I've no intention of profiting from this, hence my declaration (and I hope I don't regret it if ISFDB becomes as popular as Arthur C. Clarke's communication satellite idea). But I'm going to refrain from adding any Smashwords titles until there's a bit of consensus about this. It does look as though we could easily add affiliate IDs to some other Amazon sites to compensate Canadian, German, and French editors (but as I don't speak some of those languages I can't be sure if there is an affiliates program for those Amazon sites). It looks as though Al might profit from Amazon US and Amazon UK still (and I for one don't begrudge him those pennies) but it really is worth a discussion. "We do it for nothing and can't be bought" might be the preferred option, in which case we should shut off Al's income stream. "Let the developer have a little money for enabling the feature" is another. "Let the editor choose to get a little kickback" is another but I think that's hardest to police. But I'm willing to fix my "title links" improvement to ban such if that's the consensus. BLongley 00:15, 28 October 2011 (UTC)

Just FYI, I've been aware of that possible abuse by an editor and have kept my (intermittent) eyes open for obvious associate type links. I don't recall seeing any coming in, nor seeing any sitting already in the database. Kevin 23:56, 27 October 2011 (UTC)

As for the associate dollars accruing to you, I'm perfectly fine with you accumulating it until such time that it reaches some reasonable amount (~$25 or the equal to 1 hardcover book is a reasonable (and appropriate) standard). As for importing the dollars, I have a paypal account that is registered as a business in the U.S. and could receive dollars to skirt the VAT incoming tax on the funds, and I can cut a check to Al or Ahasuerus when (if ever) you need to render unto Caesar. Kevin 23:56, 27 October 2011 (UTC)

It's not VAT incoming Tax, it's outgoing US taxes. I don't expect there to be enough in any reasonable timescale to cover the cost of getting the US paperwork, so redirecting to the people paying for the site seems wiser. I'm not a fan of Paypal (used them, got screwed-over by them) but if anyone is going to be compensated then that's probably simpler than waiting till we get a $75 paper cheque sent to me and have me posting it back to someone that can use it. BLongley 01:45, 28 October 2011 (UTC)

As for the topic of Ban, Encourage, Discourage, I think the answer is Only in disclaimed instances like this, only by moderators, and only when there is commitment to return the dollars to the mother ship or other group expense. This is Al's decision in the end, but these are my 6 cents or so. Kevin 23:56, 27 October 2011 (UTC)

On the same topic (Should we allow, ban, encourage or discourage this?), the only thing that I can say is that I'm against it personnally (meaning that I don't want to be "rewarded" or encouraged) but think that the one(s) paying for the ISFDB should at least see their expenses covered (if possible). Hauck 05:30, 28 October 2011 (UTC)

We should prohibit any editor from adding a link to any website from the ISFDB for which they will receive personal compensation. The person/corporation/entity/whatever who is paying for the hosting of the ISFDB (I believe it's Ahasuerus) should be reimbursed for any out-of-pocket expenses if we ever generate any income from linking. I wouldn't object to having Amazon affiliate links from the front page forthcoming pubs, but wouldn't want them on every pub record. With a 4% kickback from Amazon to the ISFDB coffers, I'd probably be personally responsible for a least a couple of months worth of ISP hosting each year! Mhhutchins 17:04, 28 October 2011 (UTC)

I think we DO have them on every ISBNed pub record, for the links to Amazon USA and UK at least. (We certainly include an "isfdb" reference on those links.) If the consensus is to put all such income towards hosting costs that's fine with me, and we should look into taking any money offered from the other sites we link to. I have no idea what this site costs or how much is being accrued (if anything - maybe those Amazon links are just to justify our being able to use Amazon Images rather than for revenue?) but there are times when I wish our service was a bit more robust. And if it turns out there's a lot of money to be had (Smashwords default is 11%!) then we can fund an ISFDB editor meetup or something. BLongley 19:18, 28 October 2011 (UTC)

I think the most important thing is that we know where the money is coming from, where it's going, and who's in charge of such. I think there was some talk of making ISFDB an official non-profit organisation at some point, for legal purposes (less likely that individual editors will get sued, an entity that can own the software copyrights, etc) and there might be a good reason for us to do that so that we all see where the money comes from and where it goes. But that's a lot of work and I have no idea if we make enough money to even set it up. BLongley 19:18, 28 October 2011 (UTC)

Unfortunately, only Al can provide the answers and I haven't heard back from him yet. Keep in mind that he owns the "isfdb.org" domain and pays for it. He is also the official customer as far as the hosting company is concerned, although I send him $30/month to offset the cost.

Re: the "non-profit organization" angle, one of the original reasons to form one (or perhaps piggy back on NESFA or some other pre-existing non-profit) -- as discussed ca. 2005-2006 -- was to ensure continuity of the ISFDB software/database in case Al became unavailable. However, now that the software and the database are freely available for anyone to download, it's less of a concern. Well, except for the Wiki and the submission/verification history, which are not a part of the publicly available backups.Ahasuerus 04:14, 1 November 2011 (UTC)

Well, we depend on YOU too! We could lose up to a week's worth of updates if you stopped posting weekly backups, and possibly several weeks if we didn't know why you stopped. And I for one would hate to attempt to rebuild the site from the backups - I may be one of the few people that has a local ISFDB running, but I haven't tackled the Wiki-Integration aspects of it. I do worry about these single points of failure. BLongley 12:55, 1 November 2011 (UTC)

But back to the subject at hand. Different sites have different terms and conditions for their "affiliate" programs. I know that Amazon keeps tweaking theirs and at least some affiliates find that they have to spent non-trivial resources to keep up with the changes. I don't know if we are in compliance with the current version of the terms and what might happen if they decided that we were not -- would they tell us to disable links to their images? They haven't bothered us so far, probably because we are small fry (and good publicity?), and I don't know enough about their business practices to tell if we are in compliance, "out of compliance but deemed harmless" or something else. Which is yet another reason why I am hesitant to get us further involved in the world of "affiliate programs" without Al, who knows much more about the business side of things. I just sit in the basement and play with the software in my plentiful spare time :-) Ahasuerus 04:14, 1 November 2011 (UTC)

With Amazon, I suspect we're non-compliant but useful. :-/ It might be difficult to justify more Amazon affiliations these days if we wanted to add such for Canada, France and Germany, but we now have moderators fluent in those languages if we want to pursue such. It seems to me that the consensus so far is that nobody should profit from ISFDB work, but that nobody should be out of pocket either - so taking $30/month from affiliates for you is fine, plus whatever Al pays for the domain registration. And if it turns out we accidentally make more than that, boost our service levels with the extra money, or put it towards the moderator's secret decoder-rings fund. BLongley 12:55, 1 November 2011 (UTC)

What he said. Mhhutchins 19:46, 1 November 2011 (UTC)

2011 World Fantasy Awards Announced

Would anyone else care to add them? I think Darrah and I have had enough experience with our current tools, I thought I'd point these for someone else to try. Feel free to point out how bad our user interface currently is - there are some changes in the pipeline, but more feedback is always welcome. BLongley 01:22, 31 October 2011 (UTC)

A good idea; it would be nice to get some additional feedback from others, although you might wish to stick to entering the books and stories -- I think we're still working out how to handle things like Angélica's lifetime award (Well deserved!). Personally, right now I'm sitting back and enjoying the results. Two of my friends won, and the friend I introduced as a GoH at WisCon 2010 won for best novel! I'm sending her chocolates tomorrow. Chavey 01:44, 31 October 2011 (UTC)

What are the odds that all five fiction awards would be won by female authors, as well as the single artist award? Mhhutchins 04:23, 31 October 2011 (UTC)

It depends on a great many factors, some of them probably too obscure to spot. There are many strange coincidences in SF -- like the Sheffield/Clarke space elevator incident or the simultaneous appearance of a bunch of Mars novels a few years back -- which we can only blame on Richard Shaver... Ahasuerus 06:11, 31 October 2011 (UTC)

Fixer October 2011 - January 2012

Prior to the Amazon change on October 26, Fixer grabbed everything that he could for September 2011 - January 2012. I am now in the middle of processing his catch and submitting it to ISFDB. There are under 200 likely US-originated submissions left ("likely" because I don't know how many will be pre-rejected or auto-suspended due to comics, calendars, non-genre, vanity, etc exclusion rules) and about 400-500 UK-originated ones. The number of UK submissions will be higher than normal because I haven't submitted the UK sub-queue in a while due to other things intervening.

One downside to grabbing so much "stuff" before the Amazon change was that we ended up with a number of January 2012 pubs with no cover scan. Sorry about that, I will try to move what's left of the 2012 ISBNs to a different queue and submit them in mid-late November when more scans should be available.

On the plus side, the Fixer reorganization necessitated by the Amazon change is mostly over. As an added bonus, I have made Fixer's internal structures more flexible and soon he should be able to create submissions using other sources. It's something that I should have done a long time ago and the Amazon change was the last drop in the bucket since I had to rewrite big chunks of Fixer's code anyway. It's quite a big of code, believe it or not -- 2MB, i.e. as much text as you will find in two randomly selected SF novels :-) Ahasuerus 03:06, 1 November 2011 (UTC)

All US ISBNs have been submitted. Ahasuerus 17:08, 4 November 2011 (UTC)

UK has been submitted as well.

This was the last batch done using the data that Fixer got from Amazon before the Big Change on 2011-10-26. The new data capture mechanism is in place and Fixer is slowly sucking in new ISBNs as they appear, but there is no hurry since we are pretty much caught up through January. I will try a few experimental submissions over the next few days, but mostly I will be concentrating on restarting development now that the development server has been rebuilt. Unfortunately, one of my fingers is experiencing problems at the moment (bad digit! no biscuit!), so typing is a bit problematic right now. Ahasuerus 06:18, 12 November 2011 (UTC)

Spam evasion maneuvers

Due to the recent increase in spam attacks on User pages, I have made further tweaks to our Wiki settings: new users are no longer allowed to edit User pages. Since it usually takes a while for our editors to add something to their User pages, it shouldn't inconvenience them too much. Hopefully... Ahasuerus 19:53, 2 November 2011 (UTC)

Comments on a Fixer submission

One of the books in Fixer's queue is "A Tale of Two Tuscan Cats". A description of the story is at the author's book page. I can't decide whether it's "just" anthropomorphic animals (hence out), or whether it's a spec fic story with better-than-average animals (as in Le Guin's "Catwings"), and hence in. So I'll leave that link for someone with more experience. But a more general question: I assume there's no easy way to attach a "note" to such a submission, while leaving it for someone else to decide, other than posting the note here. Or am I missing something? Chavey 21:59, 12 November 2011 (UTC)

"Moderator notes on submissions" might be a nice feature, I don't think we currently have anything better than what you just did. Although recently I also wished we could look up "Other Sites" links before approval rather than after. BLongley 22:45, 12 November 2011 (UTC)

That's a good point and should be easy to implement as long as the submission contains a valid ISBN. Ahasuerus 07:29, 13 November 2011 (UTC)

List of "Recent Integrations"

The list of "Recent Integrations" available to moderators should be much faster now. The list of "Recent Rejects" is still quite slow because we are using an old version of the database and it would require more tweaking of the software to speed it up. Ahasuerus 07:28, 13 November 2011 (UTC)

Fixer - first pass using the new algorithm

Fixer is in the middle of submitting additional 2011 and January 2012 ISBNs, all of them identified using the new algorithm. Please note that in a number of cases the categories and subjects assigned by Amazon may be insufficient to determine whether a book is SF. Sometimes they are just plain wrong -- Fixer has just submitted a fantasy romance which was labeled "western". The current round of submissions will have an unusually high number of these cases because most of the properly categorized ISBNs were captured using the old algorithm.

In a way, Fixer's forced transition to the new algorithm was a good thing because he can now find more ISBNs, but the logic is still a little raw, so please exercise caution. Ahasuerus 06:14, 17 November 2011 (UTC)

P.S. Fixer is also submitting certain older books which were identified using the new algorithm. Fixer uses a very different way of finding "suspected SF" now, so some patterns are likely to change.

One thing that will remain the same is the ordering of submissions, which is done by ISBN with ISBNs ending in "X" processed at the end of each run. ISBN ranges roughly correspond to publishers, so submissions are likely to be grouped thematically: one bunch may be vampire shaggers and the next one may be Golden Age reprints. Ahasuerus 06:38, 17 November 2011 (UTC)

We are about 50% through the first batch of submissions generated exclusively by the new algorithm. So far so good, although some of the older books (mostly from 1970-1989) have so little information associated with the captured ISBNs that I sometimes hesitate to submit them. Some may have been canceled and others may be duplicates of previously entered pubs that used a catalog ID instead of the ISBN. Still, I figure that it's better to have the ISBNs reviewed through the standard submission process in case they are valid. Ahasuerus 04:58, 19 November 2011 (UTC)

US completed. Ahasuerus 04:29, 21 November 2011 (UTC)

UK finished as well. Ahasuerus 19:20, 22 November 2011 (UTC)

Vanity publishers

As most of you know, we let vanity published books in as a matter of policy, but they are not a priority since other books come first. To facilitate this process, Fixer maintains a list of "known vanity publishers" whose books get automatically suspended. ("Suspended" means that Fixer keeps the ISBN's data in one of his internal queues in case we ever have the time to process it.)

Unfortunately, it can be hard to tell whether a publisher is a 100% vanity operation -- e.g. some start out as vanity shops and then become more or less legit -- and I don't have the time to research every publisher that Fixer finds. However, if a moderator notices that a particular Fixer-submitted publisher keeps popping up and that it's clearly a vanity outfit, could you please post its name here so that I could add it to the auto-suspend list? TIA! Ahasuerus 02:52, 22 November 2011 (UTC)

Fixer officially migrated

Well, it took longer than expected, but Fixer has been officially migrated. <wipes the sweat off his brow>

To test the new way of doing things, I submitted a bunch of books by Dial. Since Dial is a juvenile/YA imprint, about 50% of the identified ISBNs were in the 10-40 page range and were suspended before they could get to the queue. The rest appear to be legitimate SF, although a few are borderline, e.g. Nancy Werlin's Killer's Cousin is, to quote the note that I added a minute ago, "a YA novel with elements of psychological horror, but, according to reviews, it does have a bona fide ghost." As we all know too well, determining how much SF content certain books contain can be tricky.

Also, please keep in mind that Amazon's data for older books can be very unreliable, so you may want to check other sources when moderating submissions for pre-2007 ISBNs. Ahasuerus 05:31, 25 November 2011 (UTC)

I have submitted two dozen Gollancz ISBNs from 1990-2011, but please be extra careful as many of them appear to be 8888-00-00. Ahasuerus 18:08, 25 November 2011 (UTC)

I think I allowed one to go through as is. All the others were 8888ed. Mhhutchins 22:05, 25 November 2011 (UTC)

Another problem: this one has the image for the new "SF Gateway" ebook editions. Mhhutchins 22:08, 25 November 2011 (UTC)

Eek, that is even worse than I expected! I suppose we have pretty good coverage for Gollancz, so anything that we don't have on file is likely to be vaporware of some sort.

In the meantime I have submitted the rest of Fixer's Gollancz catch -- about 50-60% appear to be vaporware, but the rest look promising. Ahasuerus 22:24, 25 November 2011 (UTC)

What's left seems pretty dismal too. Were these done from an ISBN "0575" prefix? As "Authors: uncredited" and "Unknown Publisher" seemed to be more common than usual. I don't know how many I missed out on, I was called away on an urgent mission to drink some cider in desperate need of consumption. BLongley 02:32, 26 November 2011 (UTC)

Now that the new algorithm is live, there are over 110,000 ISBNs in Fixer's "Amazon queue". Yesterday night I ran a query sorting them by publisher and then picked a manageable one, Dial, as a test case. The test was successful, so I then picked a beefier publisher, Gollancz, with 670ish ISBN. I thought that it should keep moderators busy for a while while I was working on Python, but, unfortunately, it turned out that almost all of these ISBNs were not SF -- and what *was* SF was, well, underwhelming, as you saw. I cleaned up the mess by reviewing and manually rejecting 650 ISBNs, but now I wonder how badly the rest of the queue is contaminated with non-SF ISBNs.

Oh well, it's not like these "historical" ISBNs are particularly urgent. The main goal of the Fixer rewrite was to preserve our ability to stay current by capturing new ISBNs and that part appears to be working fine. I'll submit something else tonight (right after I finish upgrading server hardware) and then go back to Python. Ahasuerus 03:39, 26 November 2011 (UTC)

I think most moderators can stay busy with or without Fixer stuff to look at. And if not, I'm sure I can provide a few more projects! I also have ideas about using the Worldcat "getEditions" method to allow people to work on one title and fill in the gaps more easily, but that should probably wait till more language support is enabled. BLongley 05:45, 26 November 2011 (UTC)

When I last checked, the free version of WorldCat only allowed 1,000 look-ups per day, but that may be sufficient for our purposes. Also, Fixer already keeps track of this information by means of a nightly process and can be told to create submissions for the captured ISBNs. Ahasuerus 07:30, 26 November 2011 (UTC)

Using Worldcat or just Amazon? I'm thinking of extending ISFDB:Community_Portal/Archive/Archive12#Bookmarklet:_Other_ISBNs so that people can quickly find editions we don't have, and can consider entering them. (I definitely bail out when it's a non-Western alphabet.) However, the Worldcat data is sometimes too complete at present, and the Amazon data is too incomplete. BLongley 00:28, 27 November 2011 (UTC)

In the meantime, Fixer has created submissions for a bunch of ISBNs from July 2007. They include some classic reprints, some juvenile novels, an urban fantasy, some erotica and everything in between -- about par for the course. Now I can go get some sleep and attack Python tomorrow... Ahasuerus 07:30, 26 November 2011 (UTC)

[unindent] I've noticed that most of the latest submissions for older releases give the ISBN-10. The Amazon "Look-Inside" feature shows they have the ISBN-13, even as far back as the 2007 releases. So I've been correcting the ISBN after acceptance. Any way Fixer can give the correct ISBN? Mhhutchins 18:31, 26 November 2011 (UTC)

Checking the code, I see that Fixer uses ISBN-10s for anything published prior to 2008-01-01 and ISBN-13s for 2008+ books. However, according to ISBN.org, the conversion from ISBN-10s to ISBN-13 occurred on January 1, 2007. I seem to recall that many places didn't convert until mid-2007, but since the official changeover date was 2007-01-01, I will change Fixer's logic to use ISBN-13 for all 2007+ books. Thanks for catching it! Ahasuerus 23:12, 26 November 2011 (UTC)

Most British publishers converted even earlier than 2007, and I've even seen some 2008 American editions with ISBN-10s. But I use the 2007 date when I'm looking at publisher's annual lists to see which pub records need to be converted. I've changed hundreds of records that Dissembler submitted with the wrong ISBN. If I don't have access to the Amazon "Look-Inside" feature (which lately seems to be going the way of the dodo because they're substituting the ebook versions for the print versions), I use the OCLC records, which like us, give both ISBNs. In their case, they always list the stated ISBN first in their records. Is there a way to generate a list of pub records post 2007-01-01 that have ISBN-10s, or even better, a way to do a universal change of their records to give the ISBN-13? Mhhutchins 23:43, 26 November 2011 (UTC)

Yes to the list: yes to the universal fix if Ahasuerus will allow such. I'm a bit wary of updating verified 2007+ pubs if the US was slow to catch on to ISBN-13s, and it's fairly unimportant while we are still converting each way. But it could be important when we start getting 979 prefixes. (I don't think I've seen one yet.) BLongley 00:15, 27 November 2011 (UTC)

And while we're at it, it's probably time to improve the ISBN search - the Advanced one seems to only do exact search, and the simple one gives up after 100 results. BLongley 00:18, 27 November 2011 (UTC)

Checking my notes, I see that at least some US publishers started using ISBN-13s in July-August 2008, so we probably don't want to change verified pubs. Ahasuerus 01:57, 27 November 2011 (UTC)

Except when the verifier didn't update the ISBN because they didn't notice it was wrong. I've gone back over records I verified and discovered that the original record was wrong and I failed to see the mistake when I updated and verified it. And believe me, from looking at thousands of "Look-Insides", I've corrected more than a few verified pubs that gave the ISBN-10 when the ISBN-13 is stated in the pub. (Many editors update the Fixer and Dissembler records from their hard copies and more than a few forget to check the ISBN.) I'm not pushing for a universal change, but a list will at least give us an idea about how big a problem this is. Mhhutchins 02:13, 27 November 2011 (UTC)

From the latest backup I've loaded, it seems there's almost 8,000 2007 pubs with 10-digit numbers in that field. I really don't think it's worth the effort to convert them - users can't spot which is recorded anyway due to the automatic conversion for display purposes, and if such does become important then we need to cope with publications that list both. I'd concentrate on making sure that new publications get an ISBN-13, as this will get important when a 979 ISBN-13 is confusable with a 978 ISBN-13. BLongley 03:37, 27 November 2011 (UTC)

Sorry, I should have said 2007 or later. Things like this are probably wrong. There's only 1300 or so from 2009 or later so it's possible to provide a project list if people are sure about the date cut-offs. But as I said, I don't think this is a priority. BLongley 04:33, 27 November 2011 (UTC)

June 2007

June 2007 submitted (highlights only.) Ahasuerus 03:11, 27 November 2011 (UTC)

Advice needed

Am I doing and saying the right thing here, or should I go ahead and accept these submissions without question? This is becoming frustrating for both sides. My only other move is to leave them in the queue and let another moderator who has less concerns handle them. We can't afford to lose an editor, so maybe a word of encouragement from those who are working on the problem can help. I'm not in any situation to make promises about when these changes may see the light of day. Mhhutchins 17:15, 27 November 2011 (UTC)

I know we disagree on this topic, but my advice is to accept the submissions as they are. For me, at this time, the multi-language aspect of the ISFB is perfectly serviceable (bar the displaying of similarly titled works) and a presentation like this one is clear enough for any potential ISFDB user, for whom a bibliographical leaning shoud be expected (all this is IMHO of course). The vting of non-english titles is for me a non-issue as it seems to be the correct way to do things. If there's extra work to do it's only for works entered in the "old" way (e.g. non vted). There are probably lots of display issues to discuss and feud about but the fact that every title has a language and that there is logically one canonical title whose language is the one of the author (or the one of its first publication ?) and that "other-tongued" titles are variants seems to me perfectly sound. Hauck 18:29, 27 November 2011 (UTC)

I would accept them as they are. I would still strongly discourage Reworking pubs entered the old way for now as that will become much easier. And great care is needed to set the right language when entering new non-English publications, or all content titles will be set to English by default - and there's no quick way to change all contents at once. But if we start capturing non-English titles correctly at entry there won't be any extra rework needed, it'll just lengthen biblio displays until users can effectively choose their desired languages. BLongley 19:03, 27 November 2011 (UTC)

This is strange. Didn't we just last week decide to discourage varianting translated titles? (See this discussion). No wonder the editor is so frustrated, when we can't made a decision and stick to it for one week. I'm beginning to feel his pain. Mhhutchins 20:28, 27 November 2011 (UTC)

We agree on "unmerging of non-English translated titles" - at the moment, unmerging, re-merging, changing title language, re-varianting is at least a three step process, four or five at times. That I would still discourage as a waste of effort, but if a moderator wants to try it, or mentor a particular editor, I'm OK with it. I don't think we're ready to open the floodgates, but if the Euro Mods are happy with such I'd let them get on with it - just don't let editors expect as much support as we currently have for English titles. We do need more proficient moderators - I'm working on things I don't fully understand, but realise that we're moving toward a true Internet resource rather than just an English-Speaker's one. BLongley 00:35, 28 November 2011 (UTC)

The current ISFDB standard definition of "variant" is the publication of a work that changes a) title or b) author or c) both, but is essentially the same work. I believe that a translated work is not a variant; it's a new work which is related to the original work. That "relationship" should not be handled the same way that we currently handle variants. There should be software changes that can properly display this relationship. This new function can also be used to handle other relationships between titles based on text: adaptations, expansions, excerpts, revisions, rewrites, and yes, translations. We shouldn't force "variant" (as we currently use it) to perform all of these functions. Having no knowledge in software design, I don't know if such changes can be easily made, but it should be our goal. And for years I thought it was. If others feel that the current software handles translated titles well, then what actual changes are we pushing for in the software design? Mhhutchins 18:59, 27 November 2011 (UTC)

As far as I'm concerned, display issues should be put in the forefront as the pages are starting to be cluttered and that some choices (e. g. the fact that the texts in series are on top of the pages, the peculiar way to count pages or the whole CHAPTERBOOK business) are not really to my taste. But every editor has his/her own agenda (in terms of entering strategy) or his/her bibliographical preferences and the result of all this centrifugal choices is by definition a compromise. I also think that it's important not to discourage potential contributors and I appreciate your way of handling things in Rudam's case. Hauck 19:12, 27 November 2011 (UTC)

In my opinion, the new standard definition of "variant" should be the publication of a work that changes a) title, b) author or c) language (or any combination of these), and treating translations as variants is the right way. I agree that non-moderators should not yet rework translations entered in the old way, but I see no harm in accepting new translations. New editors should be warned about the extra edits needed though, especially on collections and anthologies. See this as an example of how things look when (some of) the vt's are forgotten. The software changes I want most are the translator field, setting language preference, and display issues like with this pub (all titles are vt's, but none are shown). --Willem H. 20:09, 27 November 2011 (UTC)

Language preferences are with Ahasuerus for testing: if those pass muster, then the display changes to use them can be worked on. I just don't want to get too far ahead if I need to rework the current proposals. I've also had a look at translator support and think I can add the capturing of that data fairly easily, with limitations (e.g. only adding one translator to a title at a time). How we display such is still open for debate. The display issues with same name and title in different language shouldn't be difficult to fix, but once "Working language" and "User languages" are in then a LOT of displays need adjustment. BLongley 20:29, 27 November 2011 (UTC)

The design is currently very simple: a title is either a variant (and points to a parent title) or isn't (and has a title parent of 0). How we use it is purely down to convention - as you say, a title change or author change justifies a variant. The proposed changes extend that to "a different language" is worth a variant as well. (It's still debatable about whether a different translator will be worth a variant.) The recent changes to display title languages mean (I hope) that the existing varianting process will suffice for now. We do need to make entering non-English titles easier, and conversion of titles entered the old way easier, and apply language preferences to the displays. I have ideas on these lines, but they are not going to cover adaptations, expansions, excerpts, revisions or rewrites. Those relationships require a LOT more work and come with a high risk of messing up the existing relationships - and I suspect would be far less frequently used than languages. We have editors wanting to work on French, Dutch and German titles and I think it's safe to let them do so manually - yes, it clutters the display a bit but that's what we're working on next. I am certainly not prepared to unleash the weapons of mass translation till those changes are ready! BLongley 19:47, 27 November 2011 (UTC)

My purpose in starting this discussion was to keep non-moderating editors from becoming frustrated at the decision we arrived at last week to delay the varianting of translated titles. Now I'm seeing no one really meant that. The editors are seeing this being done by moderators and are wondering why there's a double standard. (See here.) I give up. Mhhutchins 20:34, 27 November 2011 (UTC)

No worries - its just a miscommunication. We can cope with a bit of improved language support at present, but it needs a lot more. We're working on that. We are heading into areas where some Moderators are not capable of moderating some languages - I know I'm not going to be one. So long as the Mods that can cope with the editors that try, we'll be OK, I think. BLongley 00:49, 28 November 2011 (UTC)

(unindent) Sorry about all the trouble, folks, it's my fault because I haven't tested/implemented the changes that are needed to support multiple languages yet :-( Ahasuerus 22:15, 27 November 2011 (UTC)

No apology necessary - we're all volunteers here and shouldn't expect anyone else to take charge of things. The fact that we have moved on since Al left us up to our own devices is a good sign though. (Although it would be nice if the non-coders at least translated our help into French, German and Dutch for a start...) BLongley 01:20, 28 November 2011 (UTC)

I am done with Fixer for now (aside from handling new ISBNs that pop up every month), but it will still take some time before we have full fledged language support.

How about the following as a band aid solution: change the Summary Bibliography display logic so that VTs are only displayed if their language code is:

1. English
2. The same as the language of the canonical title, e.g. both are French
3. Not defined (for old title records that have not had a language code assigned yet)?

Note that this change will not affect the Title page, which will display all VTs and their languages. It's not perfect, but it can be implemented quickly, within a day or so, and give us enough breathing space to do the rest. Would that work or am I missing additional permutations? Ahasuerus 22:15, 27 November 2011 (UTC)

I'd say not to bother with this as the cluttering effect is not so strong that it requires such measures. For me the vt thing is one of the main forces of the ISFDB and the presence of vts on the SB page could be an clear signal to an non-english-speaking casual user that his/her language is also treated here. A "global" (e. g. a ticking box with "show all", "show author language" or "show only texts in XXX") solution can be found later. Hauck 04:39, 28 November 2011 (UTC)

OK, I can do that too. It won't be as elegant as the ultimate solution, e.g. it won't suppress collections and omnibuses that have no analog in the original language, but it will get us 85% of the way there and allow entering non-English titles as VTs. I have the code mostly working on the development server, but it needs a bit more polishing. With luck, I should be able to install it tomorrow night. Ahasuerus 06:46, 29 November 2011 (UTC)

Fine, that'd be my preferable way of doing it, too. Stonecreek 15:38, 29 November 2011 (UTC)

After installing the patch, I picked Wilson Tucker's The Long Loud Silence as a test case. It has a good mix of pubs: 7 English language reprints, three Italian ones (using two different titles between them) and one appearance in German, Spanish and Swedish each. Does the Summary page look OK to everybody using different choices under Language Preferences? Ahasuerus 07:18, 30 November 2011 (UTC)

Looks good. I tried it with all languages (shows all vt's) and with Italian and German unchecked (suppresses the right vt's). --Willem H. 07:30, 30 November 2011 (UTC)

Summary page looks good, although I'd like the language to be displayed. Same with Chronological display. Alphabetical display doesn't show any variants but I think that's been a minor problem for ages. And we still need to fix problems like "Solaris (1961) [also as by Stanislaw Lem ]" when the title and author are unchanged although the language is. But overall a good step forward. BLongley 17:58, 30 November 2011 (UTC)

Except that a translated publication is not a "Variant Title" as it is now displayed. "Variant Title" implies that only the title is changed and the works are essentially the same. A translated publication is a related work, but is essentially not the same. I'm puzzled that this bandwagon has been jumped on without any real discussion. Guess I missed the comments period for any input into how the software change would handle translated publications. I'd always assumed it would not use the same function that we currently use for varianting titles. Mhhutchins 21:32, 30 November 2011 (UTC)

Agreed: we already bend the "title has changed" to allow for "author has changed" even when it's just a slight change in author credit. We're bending it a little more here. We might need to bend it a bit more ("translator" has changed) before we move on to a more final solution. (I say more final as I'm sure there will be changes long after I'm gone.) BLongley 02:29, 1 December 2011 (UTC)

The approach that we are currently implementing was first outlined over 2 years ago when we started the first discussion of what we called ISFDB:Proposed_Design_Changes#Enhanced_support_for_other_languages and its finer points have been debated ever since. The discussion took place on many Wiki pages over the course of many months, so perhaps it has become so fragmented that it's hard to keep track of unless you all following all (or most of the) threads? Ahasuerus 23:35, 30 November 2011 (UTC)

Because of the technical aspect of those discussions I never followed them...and looking at those who were involved in those discussions, they seem for the most part to be between you and GaborLajos. Mhhutchins 00:14, 1 December 2011 (UTC)

The history of this change is quite long and complicated. First, I discussed using VTs for translations with Al in mid-2009 to make sure that we wouldn't be taking the software in a direction that he would later have to undo. Then I posted it on the Community Portal when we were discussing other proposed design changes and it got rather involved (I suspect this may have been the original failure point since we were talking about a dozen changes at the same time) and the whole discussion was moved to the Design page. GaborLajos had some ideas about adding translator support, but that was later, when the notion that we would be using VTs for these purposes was universally accepted. Or rather I should say that I *thought* that the notion was universally understood and agreed upon... Ahasuerus 07:59, 1 December 2011 (UTC)

How was I going to become involved in discussions to which I'm technically unable to understand? Admittedly, that's ultimately my failure. Mhhutchins 00:14, 1 December 2011 (UTC)

Point taken. We should do better at separating the technical issues from the design ones. That's why I started posting screen-shots of proposed changes - something that makes sense to me may not make sense to others, but any feedback is useful. I freely admit I don't fully understand what GaborLajos was on about technically - and he seems to be inactive now. I think the move towards getting something about improved languages in was inspired by Ernesto Vegetti, sadly no longer with us. I'm very much a cautious developer - I've f**ked up before, thankfully not too many times on ISFDB coding - and I have improvements waiting until prior ones go in. ("Default language" for editors and "Author working language" for "authors" before we unleash "Unmerge Foreign Titles"). BLongley 02:19, 1 December 2011 (UTC)

If I knew that I wouldn't have been raising a ruckus about all of the varianting of translated works that have been done in the last year. Why did we keep all of the translated titles under the original title record if we knew they'd all have to be varianted eventually, using the same function that's been at our disposal for years? We could have been doing this all along. Mhhutchins 21:32, 30 November 2011 (UTC)

Sorry about the confusion, sometimes Wikis with their topic/page fragmentation are not the clearest medium :-( Ahasuerus 23:35, 30 November 2011 (UTC)

That stills begs an answer to my original question: if it were planned all along that the title records of translated pubs were going to be handled the same exact way as we handle variant titles (actual variant titles), then why haven't we been varianting? It would have save hundreds, if not thousands, of submissions that entered translated titles under the original ones, and saved all the effort it's going to take to unmerge these pubs and create variants. And I can't believe it's simply a matter of display. If that were true, we could have lived with having them all displayed until the software changes were implemented. Maybe it's just a bad assumption on my part, but I thought the delay in implementing these changes was because there was going to be a fundamentally new way of entering translated pubs, and I was willing to wait until the programmers had the time to work on it. Treating translated pubs as variant titles seems to me, in the long run, to be a bad design choice. I'm sorry I didn't know sooner and lost any opportunity I may have had to offer that opinion. Mhhutchins 00:14, 1 December 2011 (UTC)

Sorry Michael. You certainly haven't lost the opportunity to offer an opinion - for instance if you feel strongly that a translation is not a variant, we can incorporate that in future changes and display the relationship as "translation of" or whatever people prefer. Those are big changes though - really, REALLY big changes - and this seemed to be a comparatively quick win I guess. (It's not one I designed and/or coded but it does seem a good way forward to me. My changes are still to come.) BLongley 02:00, 1 December 2011 (UTC)

Advice on Fixer submisisons with sparse data wanted

What is the current practice on handling Fixer submissions for forthcoming books with sparse data? Specifically I was looking at the submission for "Petrify (Wired Up)" by Beth Chambers (submission #1732942). The Amazon page this links to doesn't contain a synopsis or any info from which one can conclude if this is spec fic or not. By searching on the author's name, i found another Amazon page which shows a cover image and gives a synopsis which makes it at least probable that this is horror (it leaves it a little unclear if the horror will turn out to have a non-supernatural explanation) and gives a page count (72pp). But it also gives a different ISBN than the Fixer submission does. Is this enough evidence to approve the submission? Or should it be rejected? Or did we ever implement "Defer until published"? -DES ^Talk 03:57, 4 December 2011 (UTC)

You can add it to the Deferred Submissions list if you feel that it may be an eligible pub, but don't think there's currently enough evidence to let it in. You will have to then reject the submission. Fixer will not submit the pub again, so you will have to check back after the scheduled publication date to see if additional data has become available. Looking at the list now, I see some moderators are just adding items to the list without returning to confirm or delete the items listed. Mhhutchins 04:18, 4 December 2011 (UTC)

Thank you. Looking at User Talk:Fixer#List of ISBNs from rejected Fixer submissions It seems that the plan at one point was that this list would be fed back through Fixer every 6 months or so -- I guess that didn't happen? -DES ^Talk

I don't believe it did. Or if it was implemented, I missed any announcement about it. It's my understanding that Fixer will not resubmit any rejected submission. That may have changed as well. Even though I try to look at every change in the Wiki on a daily basis, I may have missed it. Mhhutchins 04:30, 4 December 2011 (UTC)

Fixer resubmissions are currently available, but they have to be run "on demand", i.e. I need to initiate the process on the server where Fixer resides. I haven't done run yet, but I can do it for older ISBNs once the current batch has been submitted (612 in the queue right now, but based on prior experience it will result in fewer than 400 submission.) Ahasuerus 05:01, 4 December 2011 (UTC)

This is useful, but would be much better if it were easy to use (e.g. pushing a "Defer" button on the moderator screen). As such, I have submitted a feature request:

"Instead of adding to the "Deferred Submissions" by hand, it would be great if on the moderator screen we had a button "Defer Submission" which put that submission onto a queue to wait until more information was available. Ideally, it would put it in a queue much like the "Submissions" queue, so that returning to that queue occasionally to look at deferred submissions would be easy to do."

To make comments on that F.R., you could go to this link. Chavey

The way things are at the moment, a submission is either "New", in which case it appears in the "New Submissions" queue, or "Integrated", in which case it appears on the "Recent Integrations" page, or "Rejected", in which case it appears on the "Recent Rejects" page. I guess we could add another status, "Deferred" and create a "Deferred Submissions" page. Hm, it may be doable, let me think about it... Ahasuerus 04:58, 4 December 2011 (UTC)

Difficulty in editing ISFDB wiki

A new contributor has some difficulties with logging into our wiki. See here for his description of the problem. Is there some help we could offer? Stonecreek 10:11, 4 December 2011 (UTC)

He may misunderstand logging into the ISFDB vs. logging into the Wiki. Other ideas:

• Knowledge of login state in the Wiki is handled with a persistent cookie that eventually expires (requiring another login). If something on his system is deleting the cookie, or if his system time is far into the future, the cookie may be going away on him.

• There may be some caching behavior in Safari (or in something between his Safari and the ISFDB site, if he were going through a proxy) that could be keeping the old, you're-not-logged-in version of the page and delivering that instead of the fresh, you-are-logged-in version. This case is highly unlikely unless he's operating out of a corporate/institutional environment.

Maybe Ahasuerus can see the logs and find some useful information there. --MartyD 13:11, 4 December 2011 (UTC)

Another possibility is that he has his browser set to refuse cookies. Ideally, the login process should check to see whether it can set a cookie and then read it. Then if the user has cookies turned off (or their clock set badly), the login process can bring up a warning message describing what's happening. But I don't know if our system does that. Chavey 14:43, 4 December 2011 (UTC)

Logging in would never work without cookies, so I don't think that's it in this case. The symptoms seem more like very rapid session expiration. Hard to tell. --MartyD 21:18, 4 December 2011 (UTC)

The user responded and seemed to think he might have logged into the ISFDB but not into the wiki. That confuses many people, I think, since the username and PW are the same, but the login cookie is not. -DES ^Talk 00:02, 5 December 2011 (UTC)

But it used to be so. Since a few weeks I have to log into the wiki when I want to add biographic remarks. That was not the case a month or so ago. --Dirk P Broer 22:07, 5 December 2011 (UTC)

It certainly was the case that separate logins with separate cookies were needed a year ago -- I will be very surprised if this was not the case a month or two ago. -DES ^Talk 22:30, 5 December 2011 (UTC)

An added nuisance about the logging into the wiki is that when you want to make a biographic or bibliographic comment for an author, the header is removed once you login. My work around is to copy the header first before I log in and then upon entering paste it into the record. This behaviour (the dropping of the header upon logging in) may account for the loss of them. --Dirk P Broer 22:01, 6 January 2012 (UTC)

Very interesting. That may explain why EVERY new person who uses the db link to create a bio or author page removes the header. Or I thought they were removing them. You may have noticed that I strengthened the message even further today, but it's possible the new editor isn't even seeing the message. Programmers to the rescue! Mhhutchins 22:10, 6 January 2012 (UTC)

Hm, that's peculiar. I will experiment with it to see if I can recreate the problem, but my knowledge of the Wiki software is decidely lacking, so I am not sure I will be able to fix it.

In the medium-to-long run, I think we want to move the Author and Bio pages to the database proper. That way they will be attached directly to the Author record and there will be no orphans, mismatches, etc. It's not that hard to do, just two new pages with a single "Note" type field each and an "Edit" link. Of course, existing Wiki pages will be migrated automatically. Ahasuerus 06:03, 7 January 2012 (UTC)

Should be fairly trivial, apart from the next-to-impossible bits. (Migrating existing pages and keeping all the links on them working...) BLongley 12:02, 7 January 2012 (UTC)

Kathleen Bryson in the moderator queue

I'm leaving a submission request in the moderator queue that I don't feel qualified to decide. Dirk suggests replacing the legal name that is currently "Kathleen Kiirik Bryson" with "Kathleen Bryson". The author's Wikipedia page says "Kiirik is an Inupiaq name which was given to her shortly after birth in Wainwright, Alaska." But I have no idea if that means it's more like a "nickname" or part of her legal name. (I'm pretty sure she could sign contracts under either name.) She uses "Kathleen Bryson" as her acting name and "Kathleen Kiirik Bryson" as her writing name, as reinforced by the additional web page on the author that Dirk has also added with his submission. I leave this question to others. Chavey 16:22, 6 December 2011 (UTC)

Never mind :-), it's already been approved. Chavey 16:24, 6 December 2011 (UTC)

Sorry. I didn't see your notice here until after I'd approved it. (Placing a submission on hold will keep it from being moderated, even if temporarily.) I based my decision on the Wikipedia page, as it seems that the name she writes under is not her birth name. But who's to say if she's legally changed her name to include it? I personally think the legal name field has a very limited purpose (in cases where a pseudonym has become the canonical name) and otherwise of very little importance, especially in the case of someone with one short story credit. Mhhutchins 16:55, 6 December 2011 (UTC)

It can be useful for distinguishing a pair of similarly named authors - e.g. we've just separated two Karen Williams, there were two British Colin Harveys (one died recently though), etc. And sometimes it's the only way to record real names with "funny characters" in, where using such for canonical name would make them difficult to search for. BLongley 17:07, 6 December 2011 (UTC)

Granted, but still, 99% of the time it's useless, meaning it serves no purpose, bibliographically speaking. Mhhutchins 17:52, 6 December 2011 (UTC)

I find it often helps mark which is the canonical name. And it can be info worth having, but is less important than many other fields, IMO. -DES ^Talk 19:54, 6 December 2011 (UTC)

If the 'legal name' is to reflect the last name that the author published under, then you have to consider this phrase in Wikipedia: "Bryson acted under the name Kiirik Bryson for several years, though she now uses "Kathleen Bryson"." --Dirk P Broer 01:01, 7 December 2011 (UTC)

I don't think we've really established what 'legal name' means here. I'm pretty sure it ISN'T "last name that the author published under". Although that might have suited Peter Morwood and Steven H Silver, who have both turned up here to complain about their treatment. BLongley 01:58, 10 December 2011 (UTC)

Patch r2011-66 installed

Patch r2011-66 has been installed. It added the following warnings to the moderator approval page for New Pubs:

1. Catalog IDs without a "#" sign are now highlighted.
2. Invalid ISBNs (10- or 13- digit) are also highlighted.
3. Binding codes that do not match what is currently listed in Help are highlighted -- this includes things like "pulp" vs. "Pulp".

In addition, if a New Pub submission includes an ISBN (valid or otherwise), a list of "verification sources" is displayed below the Hold/Approve/Reject section. It provides ISBN-specific links to all currently supported "Other Sources". Note that this list includes all sources, including sources that you may have chosen to ignore on the User Preference page. Hopefully, this will help process robotic submissions more efficiently. Ahasuerus 19:31, 7 December 2011 (UTC)

Yay for 1 and 2. Not so sure about 3 - previous discussions have petered out - see Rules_and_standards_discussions#Audio_Book_Bindings_.28again.29 and the previous Rules_and_standards_discussions/Archive/Archive05#Audio_Bindings - but I know I didn't like the current help for audio or ebooks. I didn't push it at the time as I never entered either, it was only Fixer or Dissembler that ever did. That's mostly still the case so I'm not going to push this time either, but I would like them standardised more and if we can reduce them to a drop-down list or two I would be very happy. Even if it's only for hc, tp and pb and "other". BLongley 23:48, 7 December 2011 (UTC)

I thought about that too, so the warning message says "Uncommon binding" rather than "Invalid binding". It was primarily added to make it easier to catch things like "hardcover" instead of "hc", but if it also highlights discrepancies between what the Help pages say and what most editors do (e.g. "pulp" vs. "Pulp"), well, that's just an added bonus. Hopefully it will make the problem more visible and more moderators will be able to form an informed opinion. Ahasuerus 00:02, 8 December 2011 (UTC)

If it restarts the discussion for bindings I'm OK, although I'm quite happy with the "hc, tp, pb" options. Maybe even put an autoconvert on some of the common errors - I can't recall the last time a new editor entered their first pubs with our official binding codes and "hardcover", "hardback" and "trade paperback" should be no-brainer conversions. Are you still averse to such scripts? By the way, what do the magazine mods think about the warning for "quarto"? BLongley 00:17, 8 December 2011 (UTC)

I pushed for that binding to be entered because so much of our data at the time was from Locus1 who uses "quarto" to indicate sheet-size magazines (in the US a standard sheet of paper is 8 1/2 x 11 inches). I didn't add it to the options because many magazine editors preferred bedsheet. In any case, "quarto" has been used and is being used and there should not be a warning for it. I don't want to change the help standards unilaterally but it has become relatively common in usage. Mhhutchins 00:35, 8 December 2011 (UTC)

Yes, rechecking the data from a more recent backup suggests "quarto" (at 2,113 entries) is now about three times more popular than "bedsheet" (721). I guess we've been using Locus a lot. "bedsheet" means nothing to me in binding or paper-size terms so I wouldn't miss it, but I have no dog in that fight. (Even after a day or two filling in missing Playboy issues - as a Nongenre magazine I didn't bother with "bindings" on the first pass.) I'm quite happy to have A4 and A5 in there, those are useful for dimensions for a lot of UK fanzines, and maybe "foolscap" for older ones - but again those aren't a priority to me. BLongley 01:23, 8 December 2011 (UTC)

I should look into "Binding" Advanced search - people can choose to look for "unk" but not null or blanks. But searching for blank/null fields is missing from all options that I can see. BLongley 01:23, 8 December 2011 (UTC)

The list of "official" binding codes implemented in this patch matches what's currently in Help. I am not crazy about that list myself, but I didn't want to deviate from what Help claimed was our current standard. If we decide to change Help, then I can easily change the software to use the new values. Ahasuerus 17:07, 8 December 2011 (UTC)

I didn't want the help changed the last time. :-( We still have far more publications using the "old" audio "bindings" than we do have using the suggestions DES made:

audio CD	841
audio (MP3)	759
audio cassette	409
audio	187
audio MP3 CD	146
audio (CD)	119
audio (CAS)	110
audio (LP)	10

As I started the last two discussions and got nowhere, I'll let someone else start off this time and reiterate that mass-update scripts are in order here. E.g. I want all "audio (CAS)" to be "audio cassette". If we can't do it in one thread, then let's fight each case on its own. It's a horrible fudge at present. BLongley 02:16, 10 December 2011 (UTC)

And for unnumbered point 4 - I think "Verification sources" is a misnomer. We could turn some of them into things a verification can be made against, but I wouldn't value anyone verifying any pub against any bookseller site. (Maybe against the definitive publisher's site, but I'm already thinking about those - if we can record the pattern for ISBN searches against a publisher on the publisher page then we can keep those out of "Other Sites" but still have a link when the publisher is entered "correctly". But that's a long way off.) BLongley 00:18, 8 December 2011 (UTC)

True, they may not be "real" verification sources, but I couldn't think of a better term to describe them in this context. Ahasuerus 17:01, 8 December 2011 (UTC)

In the meantime, it does look mildly useful - if only to have a quick link to Amazon UK for US pubs, and vice versa. And can only get more useful if Fixer starts browsing Canadian, French, German and Japanese Amazons - although I hope before we get that far that I can turn off the sites, or the languages, that are beyond me. BLongley 00:18, 8 December 2011 (UTC)

Fixer has acquired biblio data from 7 non-Amazon sources, but the submission generation process hasn't been enhanced to take advantage of it (yet). The main challenge is reconciling conflicting data elements, e.g. if Book Depository and Amazon UK claim that the price is £17.99, Amazon.com says that it is $24.95 and FishPond lists it as AU$41.99, how can Fixer tell which one is correct? Presumably Fixer will need to add all of these various permutations to Moderator Notes, but it will require a fair amount of juggling. Ahasuerus 16:56, 8 December 2011 (UTC)

I thought that was Bibliovore rather than Fixer? I think some Bibliovore submissions might make a nice change for some of us. Although I also feel that we don't have much Mod support for audio or ebooks (I'd be quite happy not to moderate those). BLongley 02:24, 10 December 2011 (UTC)

Bibliovore handles publicly available library catalogs (Open Library, TALIS, the Library of Congress, etc) while the 7 "non-Amazon sources" above were bookstores like Barnes and Noble, W. H. Smith and FishPond. They have quite a few ISBNs that Amazon doesn't have (ca. 100,000 although many are for comics and such), but I just don't have the time to do them justice at the moment :-(

We also need to capture genre ISBNs from Amazon.fr and Amazon.de, but that shouldn't be too hard since I can relatively easily (8-16 man-hours) adapt the data capture software that was developed to find genre ISBN at Amazon.com and Amazon UK after the last round of Amazon changes. And we need to do it before 2012-01-01 when Amazon may further tighten the rules at its European stores to match the recently implemented Amazon.com restrictions. Ahasuerus 04:13, 10 December 2011 (UTC)

Differentiating submissions

A couple of days ago I was about to approve a submission when I realized that the submitter had been recently moderatorized. Oops! At that point it occurred to me that given the relatively high number of active moderators lately, it may be beneficial to color-code or otherwise mark submissions created by moderators. While we are at it, we might as well add a way to distinguish submissions created by bots.

What would be the best way to do this? I am color-blind, so I can pick a color that most color-blind people can see, but it won't be 100% foolproof since there are different types of color-blindness. Any other ideas? Ahasuerus 20:40, 7 December 2011 (UTC)

Would a suffix in the submitter column (-m, -b) work as an additional clue in addition to any color coding? --Ron ~ Rtrace^Talk 21:17, 7 December 2011 (UTC)

I like the idea of color coding the submissions. How about a further step of coding differently the submissions by new editors, say less than 500 submissions? I couldn't tell you which colors to use that would help those moderators who may have one form of color-blindness or another. Just as long as they're not too bright, I have no preference in what colors you choose. Mhhutchins

We're really bad at suggesting colours, which is why internal Wiki-links are that distasteful shade of green - DES and I thought if we made them really horrible, people would be inspired to suggest something better! (Not too late to change that I believe, now DES is back.) BLongley 23:36, 7 December 2011 (UTC)

I think we'd better mix colours and text changes. An (M) for Mods or (N) for New users should be fairly clear, and there's space for that especially if you trim some of the unnecessarily wide columns like Submission, State, and Date/Time. If you're going to work on that area then I'd like Subject to be a bit wider too. But I agree we're going to step on each other's toes more and more often as we grow and it would be nice to do something to ease that. BLongley 23:36, 7 December 2011 (UTC)

You could use a font face. E.g, italics for mods, bold for new folks. Or shading of the font (a light gray). --MartyD 02:49, 8 December 2011 (UTC)

Fixer - December 2011

Sorry about the overload of less-than-exciting titles lately, but, as you know, submissions are based on ISBN ranges. I think the worst is over and we are now in the Harper territory. Ahasuerus 00:16, 8 December 2011 (UTC)

I'm becoming more and more concerned about the ebooks. We seem to be getting "Sign up to be notified when this item becomes available" too often for my liking. (I'm afraid I did convert a few to the first physical publication when it was missing, without putting the ebook back in Fixer's deferred submissions: I think they were "Vanguard" publications entered with an ebook publisher name). The "Angry Robot" titles seem to have settled down a bit, but I'm still not sure we're covered for Kindle books with ISBNs, we're not an affiliate for those. BLongley 00:31, 8 December 2011 (UTC)

Spot the problems, suggest improvements

See this example:

There's several problems that come to mind to me.

1. The existing record isn't really English.
2. The proposed variant has no language.
3. The proposed variant has only a verified pub but there's no warning of that being affected.
4. I get to see the submission even though I claim no ability in either French or German.
5. There's no way of "holding" the submission for someone else to look at, who would do far better than me.

There's some features I'd like too:

1. I'm getting to be able to guess some languages - "In the Blizzards of Time" is supposed to become a variant of "The Temponauts"? If so, I'd quite like to know the literal English translations of the foreign titles.
2. And I'm sure another feature that's desired is that these should be "Add" or "Create Translation" rather than "Make Variant".

But I'm willing to help. Obviously I can't promise anything, I'm not the final arbiter of any change. But I'll throw out these ideas:

1. Mass update scripts to be applied for when the language is clear. (Advice needed on when it's absolutely clear, though. I don't propose to change all titles starting "Die" to "German" for instance.)
2. Get agreement on "Variant Title" versus "Variant Author" or "Variant Language" (i.e. "Translation".)
3. Language choices to apply to moderator queue.
4. Searching by language. Obviously not too useful until more title languages are set, but I'm thinking ahead.
5. Our Help pages to be translated into our most-used other languages to encourage new editors with those skills. I'd guess we have French, German, and Dutch skills fairly covered now, but I'm sure those experts could do with some help. And I think we've lost our Italian and Hungarian experts for instance, and I'm not sure we ever had Spanish and Portuguese covered.
6. Default language for Editors to be implemented to save them mistakenly adding English titles that aren't actually English. (OK, I coded that, but I did so because I think it would help.) And feedback on whether such still helps on editing a publication would be good.
7. "Unmerge Foreign title" to be implemented for those that understand how to use it. (I haven't even submitted that yet, as the existing Unmerge bug isn't fixed yet, but it's not too far off if I get the go ahead.)
8. "Translator" support.

This has been a bit of a rant, I'm afraid, but hopefully it will be taken in the spirit of progress. There's a certain level of tension between Mods at the moment that I wish wasn't there - and if I can do a bit of coding, designing, testing to help I'll do that. Some of the future language features are beyond my skills, especially if the alphabet is different, and I don't think you'd want ME to do Help Page translations. I shall go sleep now and await comments. BLongley 05:33, 9 December 2011 (UTC)

For #1 on your third list, we can do mass updates of titles based on publishers rather than titles, e.g. all titles published by Heyne could be mass updated to German. (Unless they ever published anything in another language.) Of course, that can not be done for any content title records. Mhhutchins 16:23, 9 December 2011 (UTC)

I'd be interested in knowing how dangerous a script Ahasuerus will allow. We can't just mass update all titles published by Heyne unless they've never been published by anyone else, as most are still under an English parent title. The script would have to unmerge each "Heyne" publication, set the title language, and make it a variant of the original title. Possibly auto-merge some created foreign titles too. I've got "Unmerge Foreign Title" in the pipeline that will cover all but the last step, but on a pub-language-within-title by title, Moderated, basis. So somebody could, for instance, go to Earth Abides and unmerge only the publications they are sure of. That seems to have French and German titles we have Mods for, and Spanish or Portuguese and Russian and Italian ones we don't. BLongley 00:44, 10 December 2011 (UTC)

Or on publication series, that's what I've suggested. For the few thousands of French titles that I've entered, I've deliberately done this by series to allow such a mass modification. Hauck 17:19, 9 December 2011 (UTC)

"Publication series" might be a safer subset than publisher, if scripts are allowed. If not, they'll still be useful in finding them one-by-one. BLongley 01:17, 10 December 2011 (UTC)

A script that would accurately determine which pubs need to be unmerge and then do it on your behalf (especially when there are VTs involved) could be a little dangerous. I'll have to think about it... Ahasuerus 01:46, 10 December 2011 (UTC)

I'd say more than a "little" dangerous, which is why I've not tried any so far. Maybe a halfway measure where a bot (Fixer, Bibliovore or Data Thief) submits things to be double-checked by humans? In small quantities of course. I'm trying to balance danger with convenience - reworking all our existing non-English pubs is a daunting task and wasteful of effort as it stands, but BIG mass-updates are so risky I'd really like such to be double or triple tested, and we seem to be short on testers. BLongley 02:38, 10 December 2011 (UTC)

Also, as the verifier (and creator) of the french title, please note this was done before the advent of the Language field. Hauck 17:19, 9 December 2011 (UTC)

Yes, we've been very late in implementing ANY language support, and so there is much rework to be done. I'm hoping to minimise that rework. I'm not blaming anyone for entering things the old way, and thank those that did for making it possible to suggest mass scripts. There's lots of options - safest would be to post project pages and let everyone do unmerges, title language changes, and make variant in various steps. I think that would overload the Mods. Even if we allowed Mods to select the languages they are happy to moderate. Providing a tool for "unmerge to foreign language", as I've been experimenting with, would cut the number of edits by two-thirds, but still requires moderation and maybe merges of variants (which are a bit clumsy compared with the simple "Check for Duplicate Titles" functionality). Or if anybody can suggest any fool-proof rules for changing pubs by ISBN, or Publisher, or Publication Series, or Author's main language, or title-words that only appear in one language, we (which currently seems to be me and Ahasuerus) can develop mass-update-scripts tested by at least two of us before we unleash them with possibly devastating effects. Or we can create a tool that allows Mods to do a mass-update when they decide that "all Puffin books were written in English" or suchlike. And we can build in other checks to make sure we never affect a verified pub, and let the verifiers do their own rework. Lots of options really - which is rather why I'm asking for opinions before working on something that won't be desired or permitted. BLongley 01:17, 10 December 2011 (UTC)

I like the idea of a bot submitting changes of language! I think it is the only way to have no major mistakes made. As it turns out every 'secure' way of mass-changing has some inbuilt failures you didn't think of before. What about five (or up to ten) proposals per day per language (that'd be Dutch, French & German at the moment - although it should be possible to make some judgment for other languages, too: just by looking at the publication the work is in). Stonecreek 12:59, 10 December 2011 (UTC)

Let's examine the numbers. We have:

• 2429 pubs whose ISBN-10 starts with a '2' (France)
• 1517 pubs whose ISBN-10 starts with a '3' (Germany)
• 97 pubs whose ISBN-10 starts with a '4' (Japan)
• 65 pubs whose ISBN-10 starts with a '5' (Russia)
• 8 pubs whose ISBN-10 starts with a '6' (Iran, Romania, etc)
• 14 pubs whose ISBN-10 starts with a '7' (China)
• 611 pubs whose ISBN-10 starts with a '8' (India, Brazil, Slovakia, etc)
• 701 pubs whose ISBN-10 starts with a '9' (the Netherlands, Finland, Greece, etc)

The grand total is under 5,500 pubs -- with ISBNs -- that need to be reviewed. Ahasuerus 23:54, 10 December 2011 (UTC)

I think that suggests France and Germany are worth an ISBN-based script, but the Netherlands are too risky. BLongley 01:10, 11 December 2011 (UTC)

We also have a significant number of pubs without ISBNs that need to be reviewed and a few non-English pubs that were published in the US (mostly in Spanish.) They will be harder to find, but in most cases European pubs were published by "known SF publishers" like Heyne and Opta, so, as previously suggested, we can create a list of "known publishers" to help us find suspected non-English titles. There will be exceptions, e.g. this Argentinian pub, and the only way to find them will be by comparing the pub record's title with the title record's title. Unfortunately, it will generate a lot of false positives, but we can eliminate most of them programmatically. Ahasuerus 23:54, 10 December 2011 (UTC)

Publisher-based scripts are fine by me, if people can agree on which are definitely monolingual. It would be nice to set some definite English titles rather than assume they're English by default. I'm pretty sure Gollancz is English and J'ai lu is French. BLongley 01:10, 11 December 2011 (UTC)

Finally, there are a couple thousand magazine issues -- see Fantascienza for a representative sample -- which will need to be redone. They were mostly entered by the late Ernesto Vegetti, so they are primarily Italian. Ahasuerus 23:54, 10 December 2011 (UTC)

I'd love to be able to put Italian rework at the top of our list in memory of Ernesto, but practically, that's going to be hardest. I'll look into improving my potential "Unmerge Foreign Title" for contents - Ernesto left a lot of useful Notes that unfortunately I don't think we can use programamtically. BLongley 01:10, 11 December 2011 (UTC)

Well, the Italian magazines would be clearly a case of monolingual, I think. They have no ISBN, but should be quite easily identifiable by title; 'Urania' seems to be another one. Stonecreek 12:55, 11 December 2011 (UTC)

I might be able to help with the Italian rework. Ernesto left the translated titles in the notefield. I will probably be a lot of (manual) work, but if the "unmerge/edit/set language/make variant" can be done in one submission, that would make a big difference. I think the Mondadori series is the largest one.

The Dutch pubs should pose no problem. I can correct the few that were in the database before language support with ease if they can be identified (and I'll probably see most of them as I enter my collection). --Willem H. 13:57, 11 December 2011 (UTC)

Submitted Link Error

There may be a bug in the author webpage submissions. Could one of our developers take a look at this discussion and the associated edit that's on hold? Thanks. -- JLaTondre (talk) 22:04, 9 December 2011 (UTC)

Will do... Ahasuerus 22:07, 9 December 2011 (UTC)

It's not outputting "&section" correctly. I'll see if I can fix it tonight. Ahasuerus 22:10, 9 December 2011 (UTC)

"&sect" seems to be reserved in HTML, even without a trailing semi-colon. You can't use "amp;sect" to get round it as the software uses the semi-colon to separate multiple websites submitted in one step. And if you leave out the semi-colon, it looks right but doesn't include it in the click-through, which still uses "amp". But in this case, leaving out the whole "&section" part seems to take you to the same page, and I'd also question whether that site is one we'd like to link to as an "Author Web Page" anyway. If it works for the Wiki I'd just leave it on the Author Wiki page rather than spend hours figuring out the whole encoding/decoding/re-encoding/re-decoding and display shuffle. BLongley 01:32, 10 December 2011 (UTC)

We should make the server side escape ampersands in the links when it makes the page (cgi.escape() is probably good enough). It's currently just throwing the links up there as-is. --MartyD 12:09, 10 December 2011 (UTC)

On the development server I added XMLescape() to PrintMultField in library.py and it apparently did the trick. However, PrintMultField is called in many other places, so the change will affect a lot of other pages and will require a significant amount of testing. It may be better to create a Bug report on Sourceforge and fix it later, after we are caught up with the current backlog of changes. (And definitely after I am over this flu since I am leery of testing anything major until then.) Ahasuerus 20:22, 10 December 2011 (UTC)

Check it in and let the rest of us do some testing while you recover. BLongley 01:31, 11 December 2011 (UTC)

With respect to the applicability of the link, it seems to meet the "If the author is deceased and has no official or authorized sites then alternate web sites with significant bibliographical and biographical data may be entered" criteria. However, I've started a related discussion at Rules and standards discussions#Web Page Entries as the disconnect between the Help and practice has been rattling around in my brain for awhile. With respect to the "&", that's a valid URL character and is used to separate parameters in query strings. Since it's only a display issue, it can be made low priority, but it is something that should be fixed as it is likely to occur again. -- JLaTondre (talk) 15:02, 10 December 2011 (UTC)

I'm always wary of URLs with query string parameters, unless they're in a documented API. To me, they suggest they're likely to be unstable, or that we're deep-linking, possibly without permission. It would be nice if we had a bot that could check our website links, I'm sure many are dead now. I did have a stab at cleaning up Geocities and AOL Hometown links after those services closed, but it's a boring thankless task. BLongley 16:36, 10 December 2011 (UTC)

Query strings are just one way of organizing website content and have no more inherent instability or deep-linking issues than non-query URLs. -- JLaTondre (talk) 21:36, 10 December 2011 (UTC)

I know they're not inherently unstable, but we've been caught out with such in the past. Ask the other Bill, who wasted a lot of effort in adding BLIC links that only work while you're in the same session. BLongley 23:18, 10 December 2011 (UTC)

As for checking links, I already have a script that checks author Wikipedia links (from a database dump) for deleted Wikipedia pages. I can work on expanding that to check all links. -- JLaTondre (talk) 21:36, 10 December 2011 (UTC)

Please do! And a script that checks whether Wikipedia links to us might also be useful - I suspect some pages wouldn't be deleted if they had a link to us, although I may be over-estimating our importance. BLongley 23:18, 10 December 2011 (UTC)

The output of an initial version is at URL Issues. It still needs some work as I need to add checking of IMDB links, the 3xx status codes can probably be ignored if I interpret the codes correctly, some websites may be blocking unrecognized user agents (some of the 403 and 406 code responses), a bug in unescaping Wikipedia links (causing a few false missing results), and probably others. -- JLaTondre (talk) 21:42, 12 December 2011 (UTC)

Approving series-wide changes

A word of caution: Please be extra careful when approving series-wide changes. As the "Series Editor Note" says:

• Changing the Name field changes the name of the series for all books currently in this series.
• Changing the Parent field does NOT change the name of the parent series.
• If the Parent exists, changing the Parent field relinks the Named series to that parent.
• If the Parent does not exist, a new Parent series will be created, and the Named series will be linked to that parent.

Earlier today the Honor Harrington Universe had its name changed to "Star Kingdom" and became a part of a newly created (and otherwise empty) super-super-series due to an accidentally approved submission. The change has been identified and reverted, but it's a good example of how dangerous series-wide changes can be. Ahasuerus 01:42, 11 December 2011 (UTC)

Another gotcha with series: series names with double-quotes in get truncated on the edit page, so an edit to, say, The "Conroyverse" will change the name to just The if they're not careful. Watch out for those too. BLongley 17:22, 12 December 2011 (UTC)

Beth Bernobich's Biographic and Bibliographic comments

With this author both comment types point to the same page, editing one edits the other. --Dirk P Broer 09:52, 12 December 2011 (UTC)

Someone had put a REDIRECT on the Biblio page to point to the Bio page. I've fixed that. BLongley 15:56, 12 December 2011 (UTC)

Now both pages are blank. The bio data was incorrectly removed from this page as being a duplicate of the bio page, but it is actually the bio page. Look at the page history. The bibliographic comments page (which we call Author for some arcane reason) should be deleted because it's empty. Mhhutchins 18:36, 12 December 2011 (UTC)

All restored to the condition the pages should be in. --Dirk P Broer 08:56, 13 December 2011 (UTC)

Game Sourcebooks

I have two submissions on hold updating two records for a sourcebook for a game. Nothing wrong with the edits but I thought Games and related material strictly a no-go? Should these even be on here? --~ Bill, Bluesman 19:07, 14 December 2011 (UTC)

They shouldn't strictly, but it seems people have let in a few books with background material for a known SF Universe, e.g. search for GURPS Nonfiction titles. I do zap all RPG modules, and rulebooks where the game came first. BLongley 19:15, 14 December 2011 (UTC)

Given the comments on Amazon people use these Jane's pretty much as I would read them. That is to say: as reference to the series, and not as game backup.--Dirk P Broer 21:10, 14 December 2011 (UTC)

That's pretty much the borderline for me. If I'd read it for background to a fiction series it's based on, I'd leave it: if it's the background for an RPG I'd zap it. I think previous discussions have suggested a "Which came first?" rule - the game or the books? So all the "Dungeons and Dragons" and "Warhammer" sourcebooks are out, while the fiction based on such is in. (I won't start on "Choose Your Own Adventure" books, a related but just as controversial area.) BLongley 01:50, 15 December 2011 (UTC)

Fixer - end of year cleanup

Fixer has another 50ish US ISBNs in its internal queue and then about 170 (mostly mainstream and therefore more interesting) UK ISBNs. At that point we should be pretty much caught up with 2011. Ahasuerus 06:34, 16 December 2011 (UTC)

UK submitted. Ahasuerus 02:37, 19 December 2011 (UTC)

Can you have another look at the new "Verification sources" section on the approval screen please? I spotted a few UK-sourced entries that should have been from the US site, but using the Amazon US link didn't find anything pre-approval - it does after approval though. I think you need to link by ISBN-10 rather than ISBN-13 for most sites still. BLongley 17:32, 19 December 2011 (UTC)

Thanks, I'll look into it. Ahasuerus 03:37, 21 December 2011 (UTC)

Various transliterated spellings of non-Anglo authors' names

Is there a procedure for selecting a 'canonical' name from all the variant spellings of a non-Anglo author's name? I'm currently looking at three variations of the same author on ISFDB: Rokeya Sakhawat Hossain, Rokheya Shekhawat Hossein and Roquia Sakhawat Hossein, and then Wikipedia transliterates her name differently again as Roquia Sakhawat Hussain. Given that all transliterated variations of such an authors' name are equally correct, how does one go about selecting one as canonical and the rest as variant spellings? The Wiki article mentions she Anglicised her given name as 'Rokeya' but gives no indication as to the remainder of her name.

In this instance, Hossain/Hossein/Hussain has only one short fiction entry but it's an important one in the history of speculative fiction: 'Sultana's Dream' from 1905, and which currently appears to have three unconnected entries which ought to be consolidated. Thanks for any pointers. PeteYoung 21:57, 19 December 2011 (UTC)

I have used WorldCat to decide transliterations before, but that doesn't help us here since, not having written a book, she does not appear in there. The French Wikipedia uses the Hussain transliteration of her last name, but uses "Rokeya" instead of the English Wikipedia's "Roquia". After looking around a bit, it seems to me that a valid authoritative source for this is the Greater Rangpur Association of Britain, which transliterates her name as "Roquia Sakhawat Hussain", i.e. the same as the English Wikipedia. (And then, of course, the others would be variants of that.) Chavey 23:08, 19 December 2011 (UTC)

Variants of variants

Will the moderators who are making variants of translated works please go to the Data Cleansing Scripts and run the script "Find Variant Titles of Variant Titles" (third from the bottom). I will clean them up (and have been), but I think we should all be certain that when we're making variants that we're linking to the parent title and not to a variant title. Thanks. Mhhutchins 23:02, 22 December 2011 (UTC)

Ouch! I hadn't realised it was an increasing problem. It should be possible to improve "Make Variant" to make sure this doesn't happen, but I understand that some people might want to distinguish translations of "Harry Potter and the Philosopher's Stone" from translations of "Harry Potter and the Sorcerer's Stone" for instance. But for now, variants of variants are an error (hence the Cleansing script wot i rote.) BLongley 06:41, 23 December 2011 (UTC)

There's only about four or five each time I run the script, but the majority have been translated titles. Just thought I'd send out a heads up. Mhhutchins 15:41, 23 December 2011 (UTC)

I would be very surprised if you ever found one of my variants there, and I would like to know about that. --Willem H. 09:36, 23 December 2011 (UTC)

In most cases, I could only speculate about who may have approved the varianting, so I'm not going to be pointing fingers. (And I'm not going to go through the integrations list looking the person who did it.) Other than a Bill Longley test record, there are currently only two items on the list, both German publications. One of them is verified by a non-moderating editor. Mhhutchins 15:41, 23 December 2011 (UTC)

Two of them were mine. Hauck 18:36, 23 December 2011 (UTC)

Publication recs with no title recs

The non-genre Asimov collection The Key Word and Other Mysteries contains two publication records. When I follow the links to those two pubs, they don't have the normal links back to the title record. Trying the tactic used for chapterbooks to add the title record as a content fails to create the title rec, and then removing that content item gives me the warning "Can't find the title rec for this book", i.e. the system really doesn't know where the title rec is. How can I fix this problem? Chavey 06:22, 26 December 2011 (UTC)

Can't be fixed, it's inherently part of the system until we come up with a separate way of designating NONGENRE COLLECTIONS, NONGENRE NOVELS, NONGENRE ANTHOLOGIES, and NONGENRE SHORTFICTION. Mhhutchins 06:28, 26 December 2011 (UTC)

It's been on "the list of things to do" for some time, but it will be a time consuming task, so it's not at the top of the list. Ahasuerus 00:20, 28 December 2011 (UTC)

Because everything falls under the NONGENRE type, you'll find the title record becomes one of the content records for NONGENRE collections and anthologies. Look again at the pub record and you 'll see the title record listed under the content records. The only way around this is to change the pub record's type to match the title record's type. In this case they don't match. That's why the title reference link is missing. Mhhutchins 06:28, 26 December 2011 (UTC)

I wondered if it was something special about NONGENRE. Thanks for explaining it. Chavey 06:38, 26 December 2011 (UTC)

Thanks for the quick explanation, Michael! Are you sure you don't understand the programming here? As that was a very succinct summation of the problem. I think it's somewhere on the "To-Do" list, but our "Who-To-Do-It" list is far shorter. :-( BLongley 07:00, 26 December 2011 (UTC)

As a user of the program, I know its results, i.e. how it works. I have no idea how to make it work. I could describe Michelangelo's David down to his fingernails, but that doesn't mean I would ever know how it was created. Mhhutchins 16:30, 26 December 2011 (UTC)

I've never really understood why some people can program and some can't. To me, most procedural computer languages are just a dialect of English. I freely admit that Prolog and APL were a bit too baffling to even start on, but Python wasn't that bad. I've never even read a manual or tutorial on it as lots of it was as clear as day anyway, to me. Mind you, I also have an opinion on Michelangelo's David - you just take a big block of marble and chip away anything that doesn't look like David. ;-) BLongley 04:32, 28 December 2011 (UTC)

I never had the desire to learn how to program, though with enough economical incentive I don't think it would be beyond my ability to learn. Now sculpture, that's another matter! Mhhutchins 05:13, 28 December 2011 (UTC)

Apparent Galician novel

That was interesting! I just approved my first Galician language book (submitted by Fixer). The publisher is located within Galicia, where the two official languages are Spanish and Galician. So the fact that translate.Google auto-detected the language as Galician isn't too surprising. Of course I doubt we have any Galician-speaking editors to verify this, so I'm going on the authority of Google's software, and added a note to that effect. (Amazon gives it no language, and WorldCat doesn't have it yet.) Their translation of the back cover is certainly consistent with the cover of the book, and justifies their claim for the language. (Translating as Portuguese, a related language, gave reasonable but incomplete results.) However, using some English-language chauvinism, I included an English translation (cleaned up) for the synopsis, instead of the publisher's Galician synopsis. Question: Is this the right way to handle such a book? Chavey 04:09, 28 December 2011 (UTC)

Congratulations! I'd never even heard of Galician before. I'd say you did right - if we let one in then we might encourage more Galician editors. Whether we have mods to cope with such is another matter, but we're beginning to move that way at last - when I started here a Dutch, German, French or Finnish title would probably baffle the few mods we had. (They still baffle me, but I'm beginning to learn a few new words.) The language support software changes we introduced this year seem to have been embraced by several editors, even when the programmers and testers have no skills in some of those areas. I'm not sure how many languages we have expertise in now - not as many as the software supports, and I think the lack of "Klingon" is still a software problem - but I think we'll end 2011 with a much better ISFDB, language-wise. BLongley 04:55, 28 December 2011 (UTC)

Thanks! I can read a little bit of Galician based on basic Romance-language skills, but mostly I relied on Google to do the translation. I'm not sure we necessarily need a "Klingon" language option, but I do think we need an "Other language" option. For example, there are editions of "The Arabian Nights" in 21 languages that we do not include in our language list -- none of which I've entered yet, because I had no way to enter them correctly. (Aside: My students speak 4 of those "obscure" languages. But even I was surprised when, at our monthly ballroom dance, I was playing a song in a 5th one of those languages, and the student I was dancing with began to sing along to it, having learned the song in temple as a child!) Of course we can't possibly support "all" languages, but I think some catch-all category for the occasional books in these less common languages would be useful. Chavey 05:13, 28 December 2011 (UTC)

It so happens that I have been working on Amazon.es lately and I see that its SF section includes 18 Galician, 17 Catalan and 15 Basque titles. (There are no immediate plans to have Fixer submit Amazon.es's data, I just want to capture the ISBNs while we can.) This particular ISBN comes from Amazon UK and ordinarily I would have suspended it prior to submission since Amazon.com/UK's data for non-English titles tends to be spotty, but it must have slipped under my radar. The record looks pretty good, though, so no harm done, but we probably want to change the price to the one given by Amazon.es.

Just for the heck of it, I added 59 Galician spec fic books. (Most of them are translations of English titles, and I only added original Galician publications if I could find a description that made it clear that they were eligible.) I'm guessing I got some of those 18 ISBN's, but not all of them. After doing this, I realized how hard it would be for someone to find those 59 books. So I'll be submitted a feature request that we add "language" as one of the options in the Advanced Searches. Chavey 02:48, 2 January 2012 (UTC)

As far as the 21 missing languages go, they are on my list of things to add this week. There may not be much genre SF published in Amharic and Minangkabau, but non-genre SF (Dickens, Nabokov, etc) may be a different story. Ahasuerus 18:16, 28 December 2011 (UTC)

I changed the price to that on amazon.es. That site also verified the language and gave me the page count, which amazon.uk didn't have. UK says it's tp; ES says it's hc, so I changed it to hc. The two sites also disagreed on the author's name, and neither of them had quite the name which is clearly printed on the cover -- so I went with the name on the cover. And if you want me to moderate some of those other Galician titles, let me know ; at least there we don't have to worry about duplicating what someone else has in another database :-) Chavey 19:25, 28 December 2011 (UTC)

I didn't know Amazon had a Spanish site, it's not one we link to. (Yet. Another thing for the to-do list.) No doubt we'll have to let Fixer loose on more sites, but I think we can wait till we have more volunteers for moderating specific languages. And probably till we can choose to suppress which languages we each moderate. BLongley 20:12, 29 December 2011 (UTC)

Links to Other Sites fixed

The links that were recently added to the moderator approval screen for New Pub submissions have been fixed. They originally used ISBN-13s in their URLs, but most sites still require ISBN-10s. All links should work now (fingers crossed.) Ahasuerus 06:37, 29 December 2011 (UTC)

Fixer: January-February done (first pass)

Fixer has just finished submitting forthcoming books for January-February using the latest and greatest algorithm. A much higher percentage of self-published books was auto-suspended and the submission queue was generally much cleaner, although the final dozen currently sitting in the queue are iffier because they are based on a manual cleanup of the "dregs".

Of course, Amazon adds many (and I mean *many*) books either immediately before they are published or after the fact, so we will be revisiting these months, but it's progress nonetheless. For example, Fixer knows of 1,213 outstanding ISBNs for December 2010, but only 253 of them are eligible for submission based on the new algorithm. Ahasuerus 05:00, 30 December 2011 (UTC)

Second pass done. Ahasuerus 06:25, 6 January 2012 (UTC)

Third pass done. It included a few classic reprints, but many items were pretty close to the bottom of the barrel. Thankfully, there weren't that many of them, so I let them through this time around.

Overall, it's been a useful exercise demonstrating how we can stagger and prioritize submissions. I have made some adjustments to Fixer to reflect what we have learned. Ahasuerus 05:56, 7 January 2012 (UTC)