User talk:Ahasuerus/Archive/2020

< User talk:Ahasuerus‎ | Archive
Revision as of 12:52, 7 March 2021 by Ahasuerus (talk | contribs) (Page creation)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Web of Angels

Cover artist of this is Terrance Lindall, see here or this book (page 1) about the artist. Horzel 16:12, 4 January 2020 (EST)

Thanks for the heads-up! Ahasuerus 16:56, 4 January 2020 (EST)

Series name

Hi Ahasuerus, your diplomatic skills are needed here, I'm afraid.--Dirk P Broer 08:07, 5 January 2020 (EST)

See also this. ···日本穣 · 投稿 · Talk to Nihonjoe 12:40, 5 January 2020 (EST)
Thanks for pinging me! I have posted a response. Ahasuerus 14:32, 5 January 2020 (EST)
Thanks!--Dirk P Broer 16:17, 5 January 2020 (EST)

The Fourth "R"

The initials RLC on this are those of Richard Courtney, see for example Count Brass. Horzel 14:59, 31 January 2020 (EST)

Thanks for the heads-up! Ahasuerus 23:56, 31 January 2020 (EST)

The Holiness of Azédarac in Smith's Lost Worlds, Volume 1

We have the Clark Ashton Smith story "The Holiness of Azédarac" listed in your verified printing Lost Worlds: Volume 1 without the accent over the e ([1]). I have verified that the first printing of the collection contains the accent and have corrected that publication. I would expect that your copy does as well and if so, the titles should be swapped. I'll leave this message on the other active permanent verifier's page as well. Thanks. --Ron ~ RtraceTalk 12:27, 1 February 2020 (EST)

I have checked my verified copy. You are right, both versions of the title -- in the table of contents and on the title page -- have the accent. If the other verifiers agree, we can swap the titles. Ahasuerus 08:43, 2 February 2020 (EST)

Off Center

Cover artist of this is Chris Achilleos, see Farewell, Earth's Bliss with the same cover art. Horzel 10:40, 10 February 2020 (EST)

Updated, thanks! Ahasuerus 11:58, 10 February 2020 (EST)

Locke & Key

should Small World and Dog Days be added as short stories in this series? Susan O'Fearna 17:32, 17 February 2020 (EST)

I am not really familiar with Joe Hills' work, but I think I recall working on his bibliography at one point. If memory serves, the reason that I kept the graphic novels -- even though we rarely include them -- was to help distinguish between them and the non-graphic stories/scripts that also belong to the "Locke & Key" series. I guess there would be no harm in adding another couple of graphic novels to the series. Ahasuerus 19:10, 17 February 2020 (EST)
I'll try to get them in then... Susan O'Fearna 11:56, 19 February 2020 (EST)

Fitzroy editions

Further to this conversation: Okay, I'm mostly caught up on the Verne translations and finally got around to this. I have submitted a change to the Publisher Series note. I've included the full list as we don't have most the the editions (about 25%). I know most of the rest and could enter them, BUT I don't know if the extent of the edits and abridgements and can't say if they warrant a separate title or to be folded up with the translation it's based on. So that part of the project is on hold. But once the update is approved, you can remove the I. O. Evans and Jules Verne pages. As it is a publisher series, I made the Ace paperbacks a separate (sub)series ../Doug H 15:53, 18 February 2020 (EST)

Done. Thanks for working on Verne's bibliography! Ahasuerus 11:20, 19 February 2020 (EST)
Publisher's series don't provide a hierarchy (yet) so they're just separate series for now. And still a long way to go. ../Doug H 21:23, 19 February 2020 (EST)

Publication Edit History

Found an oddity for the publication edit history. For this pub, the displayed edit history is for an unrealted new award. I can understand that this pub is before the edit history started being available, but wouldn't have expected unrelated history in that case. Thanks. -- JLaTondre (talk) 11:59, 23 February 2020 (EST)

Thanks, I'll take a look... Ahasuerus 16:22, 23 February 2020 (EST)
Fixed. Thanks for identifying the pesky arthropod! :) Ahasuerus 19:33, 23 February 2020 (EST)

Derek Godat

Hi, I have two submissions on hold -and already rejected a third from Derek Godat, whom I remember from a previous and similar request. Unfortunately I can't find the previous request, nor the answers that have been given to him, as he uses a new userid now. The old id seems to have vanished. Could you please look into it?--Dirk P Broer 16:43, 23 February 2020 (EST)

Ah, yes, I remember the last request. His last Talk page with my response is User talk:Derekgodat. Ahasuerus 18:27, 23 February 2020 (EST)
Thanks, I'll copy it in my reply to him!.--Dirk P Broer 21:05, 23 February 2020 (EST)

Image Linking Permission

Please see this conversation where Jennifer R. Povey would like to grant permission to link to images on her site. I've gone ahead and accepted the link assuming it gets worked out. I'll monitor and convert it over to Amazon if needed. -- JLaTondre (talk) 11:15, 7 March 2020 (EST)

Sure, I'll add it later today. Thanks for the heads-up! Ahasuerus 11:50, 7 March 2020 (EST)
Done! We may also want to migrate her Wiki Bio page to database. Ahasuerus 18:13, 7 March 2020 (EST)
Done. Thanks. -- JLaTondre (talk) 16:01, 8 March 2020 (EDT)

Mistyped NONFICTION Cleanup Report?

It might be worth having a clean-up report that looks for NONFICTION records that should actually be ESSAYS. As of this weekend's database dump, the following is a count of publications, by publication type, that contained a NONFICTION title record:

PubType Count
NOTE 1: The counts were created with a join so if multiple publications contained the same title record, it was counted for each publication containing it.
NOTE 2: I've corrected some of these (quite a few of the ANTHOLOGY ones) already so current counts will be slightly lower.

While OMNIBUS can be presumed to be correct (there are probably a few errors in there, but not worth having to go through all), the others are much more problematic. Example: Uncollected Prose and Poetry is a COLLECTION with 8 records that should be essays instead of nonfiction. I was thinking a report that finds NONFICTION title records that are in a publication type other than NONFICTION or OMNIBUS might be in order. -- JLaTondre (talk) 15:58, 8 March 2020 (EDT)

That's a good point. Let me see what we currently have. Ahasuerus 19:29, 8 March 2020 (EDT)
Done -- see the Community Portal announcement. Ahasuerus 20:32, 10 March 2020 (EDT)

Fitzroy editions of Jules Verne

You were part of a conversation many years ago about the Fitzroy editions. I've been mucking about with it and wondered if you thought it was now in a state that could be left or if further splitting was warranted? Further to the conversation - the Brits reprinted 6 in Panther paperbacks and 1 in Consul, but as far as I can tell, didn't refer to them as Fitzroy editions like the Ace ones did. The two series are The Fitzroy Edition of Jules Verne and The Fitzroy Edition of Jules Verne (Ace). ../Doug H 15:46, 14 March 2020 (EDT)

Unfortunately, I have been sick the last few days, so I am not in a very good position to comment. That said, the current separation between the two publication series seems reasonable. Perhaps we could update the Note fields of the Panther and Consul publication records to indicate that they were reprints of the Fitzroy editions but didn't use the same designation? That way we would be covered on all sides. Ahasuerus 19:33, 14 March 2020 (EDT)
Panthers do so state (a pre-existing one simply states Arco first published it). Consul I'm less sure about something - a site claims they did, but only refers to one book. Checking our records there are two others. I think I'll move on. ../Doug H 19:55, 14 March 2020 (EDT)
Sounds good! Ahasuerus 21:15, 14 March 2020 (EDT)
A recent discussion on varianting translations and abridgements has me wondering about something - I. O. Evans translated some of the titles, but for others he edited an existing translation, usually in the form of an abridgement. The 'general' rule for abridgements is that editors get credit (if specified) and the result is not varianted to the original. If we took this approach, a number of the Fitzroy editions would not appear under their original titles, but others would. Do we 'fix' these titles, is this an exception to the rule, or does the (unwritten?) rule need tweaking? Any suggestions on whether, where and how to raise this? ../Doug H 09:19, 7 April 2020 (EDT)
I guess I would raise the issue on the Community Portal. Ahasuerus 15:41, 8 April 2020 (EDT)

A Googol search for something lead me to this at which seems to be mirroring our site in some way. A search on the site name brought up 4,080 hits, more of our pages. ../Doug H 23:18, 15 March 2020 (EDT)

A number of sites have tried to mirror the ISFDB database over the years. Some are apparently research projects, the rest vary (monetization?). They don't last very long since they get out of sync with our database after the initial synchronization. Nothing to worry about :) Ahasuerus 09:24, 16 March 2020 (EDT)

Offsite blog

The link to this is on the Main Wiki page, but doesn't appear to have been updated since 2016. If we've done without using it that long, can we drop it? ../Doug H 11:43, 23 March 2020 (EDT)

Well, the original idea was that we would use a third party blog as a way to notify ISFDB users if and when the main server became suddenly unavailable. We haven't had many extended emergency downtimes in the last 3-4 years, so the blog has just been sitting there, unused. It may still come in handy at some point.
Having said that, we may also want to consider using Twitter. It works well for simple announcements. Ahasuerus 14:43, 23 March 2020 (EDT)
'Many' isn't 'any' and it wasn't used for those few. I suspect that most people don't have a record of the site to check. Aside from Twitter, isn't there a Facebook site? It's visible even to those without an account. ../Doug H 08:36, 24 March 2020 (EDT)
My primary exposure to what is called "social media" these days has been Usenet and Reddit, so my knowledge of Twitter and Facebook is superficial at best. A quick peek reveals that is/was the Twitter presence of the apparently defunct "Indian Short Film Database". I am certainly open to suggestions, but it may be best to raise these issues on the Community Portal. Ahasuerus 18:29, 24 March 2020 (EDT)
Our Facebook presence is here, but I'm unsure which one of us is managing it! The last entry was 6 February. PeteYoung 09:24, 25 March 2020 (EDT)
OK, I have compiled a list and posted it on the Community Portal (plus some questions.) Ahasuerus 13:23, 25 March 2020 (EDT)


Hi. Hope you're managing to avoid COVID-19.... I thought you might be interested: For work, I just upgraded our MediaWiki. The starting point was a bit better than ISFDB's is, but not much:

  • MediaWiki 1.16
  • PHP 5.6.x
  • MySQL 5.7.x

I had previously upgraded MySQL for some other reason. Anyway, it went unbelievably smoothly. I just laid down a new mediawiki, copied over images, and copied and updated LocalSettings.php (and got some new versions of extensions) -- some syntax changes around skins and extensions. I did have to upgrade PHP, and I went to 7.4.4, tweaking the new php.ini to reflect my previous settings. Then I ran maintenance/update.php, which ground on for a bit, and I was all set.

It does force passwords to be 10+ chars (don't know if this can be changed -- I did not look). There's also a handy WikiEditor extension that provides a friendlier wikitext editing interface.

I tried it all out first in a separate environment on a copy of my live installation (I duplicated the original environment, then worked through the changes needed to upgrade that) to work out the kinks. It was surprisingly easy.

If you ever are inclined to try upgrading -- in your copious free time, no doubt -- let me know. I'd be happy to help. --MartyD 08:58, 24 March 2020 (EDT)

Thanks for sharing! I'd very much like to avoid close encounters with COVID-19 since my health has been less than stellar the last few years, which increases the risk of a bad outcome. Time will tell how successful I will be.
Unfortunately, I know almost nothing about PHP. At one point Al said that he was going to work on upgrading PHP and the MediaWiki software, but it's been a couple of years and he has been effectively unavailable. I use Cygwin on the development server -- in part because Fixer needs Windows -- and I don't have either PHP or MediaWiki installed on it. I will probably have to install Linux, either as a dual boot setup or on a separate PC, before I can start looking into it. And, I am sad to say, it gets increasingly hard to learn new things as one gets older/sicker. At some point I'll probably need to bite the bullet and give other technically savvy editors like you and Annie access to the main server and take a step back. Assuming you have the time and the inclination, of course.
For now, the big challenge is Fixer. As I wrote on his Talk page a few days ago:
  • Fixer is having a bit of an existential crisis at the moment. He had been using version 4 of the Amazon API for many years, but that version is no longer available. The only notification that I received was on 2020-03-18 and it read "As of March 9, 2020, we started the shutdown of Amazon’s Product Advertising API version 4.0 (PA API 4)." Apparently, Amazon announced their plans in late 2019 and then extended the deadline by a couple of months, but they never notified me. ... the complexities of migrating from version 4 to version 5 [include] JSON vs. XML, a new authentication method, HTTPS/POST vs. HTTP/GET, a different set of supported request types and returned "groups" of data elements, etc
Everything else will have to wait until I can sort it out, but I will keep your experience in mind. Thanks again! Ahasuerus 18:46, 24 March 2020 (EDT)
Yes, please do stay as healthy as possible! If you think I can help with anything around the Amazon API changes, let me know. Making code crawl sites and interact with web APIs is my professional life.... FWIW, we run MediaWiki and PHP on Windows. I suspect you could work out the kinks without having to go to Linux. The primary issue I ran into is that newer PHP was required for newer MediaWiki, but newer PHP was not entirely backward compatible, so the old MediaWiki version would not run on the new PHP (package and method changes, etc.) -- I was forced to update both at the same time. Anyway, if experience or a helping hand could be of use, you know where to find me. :-) --MartyD 09:01, 25 March 2020 (EDT)
Thanks again! I have updated SR 51 ("MediaWiki upgrade") with your comments to make sure that they don't get lost.
Re: Fixer, there are two different (although related) issues here. The first one is the short term task of keeping Fixer alive for the foreseeable future. Fixer's current code has a great deal of functionality that I would be loath to lose until we have something better.
The second issue is Fixer's long term future. The current code base is one kludge sitting of top of another. There are various unwise software dependencies, ancient (and in some cases no longer maintained) technologies, poor design decisions, multiple databases, etc. In addition, Fixer concentrates on finding new ISBNs/ASINs by existing authors and existing publishers, but he doesn't do a good job of finding bestselling books by new authors and new publishers. It's a growing concern in a world increasingly shaped by indie authors/publishers.
I started working on a new, Python/MySQL-based, design in late 2019. I ran into various issues and had to adjust the design a number of times, but I was making sporadic progress until Amazon dropped its latest bombshell on me.
Immediate Amazon-necessitated changes aside, I think it would make sense to discuss Fixer's long-term design changes via email, perhaps using a mailing list of some sort. I already have Annie's and ErsatzCulture's buy-in. Would it be OK to add your email to that list? If so, which one? Ahasuerus 13:43, 25 March 2020 (EDT)
Of course! Happy to help. Use my yahoo one, which you should already have. If not, let me know. The sourceforge one is ok, but I don't send mail from that, so if we have a list for purposes of a discussion, using that isn't buying anything. --MartyD 10:10, 26 March 2020 (EDT)
Great, thanks! I'll see if I can start something later today. Ahasuerus 11:57, 26 March 2020 (EDT)

New cleanup report?

We already have a 'Variant Title Type Mismatches' report. Is it possible to add a 'Variant Title Length Mismatches' report? I found a number of these lately, short story, novelette and novella as variant of each other, and I can imagine these as variants of short fiction v.v. Thanks, --Willem 17:02, 26 March 2020 (EDT)

There will be thousands of those... if not tens of thousands. Not that we should not clean them up but... it may take awhile. I had always wondered why this is not tied like the series - so variants cannot have their own values :) Annie 17:13, 26 March 2020 (EDT)
A few things:
  • There are 1,765 mismatches at this time
  • It would be easy to create a new cleanup report
  • As I recall, at one point there were disagreements re: the best way to handle translations which contain fewer or more words that the original version. For example, if a German title contains 18,000 words while its French translation is only 17,000 words, do we use "novella" in both cases? Did we ever reach consensus?
Ahasuerus 17:52, 26 March 2020 (EDT)
Ha, I expected more considering how often I see the yellow warnings on those... For #3 - when it is novel/novella or novel/collection and so on, we go with the type of the original... no reason why this would be different here, is there? Annie 17:57, 26 March 2020 (EDT)
That's fine by me, although we may want to codify it in Help before we create a new cleanup report. Ahasuerus 18:08, 26 March 2020 (EDT)
Absolutely. Someone should post over in R&S :) Annie 18:24, 26 March 2020 (EDT)
I've been away from the site for a few days (low energy, recovering from a mild flu. I don't think it was corona, but since I've not been tested, I'm not sure). I'll ask the community on R&S. --Willem 16:08, 29 March 2020 (EDT)
Glad you are doing better! Ahasuerus 11:16, 30 March 2020 (EDT)
Oh, man! Get better and stay safe, please, Willem! Annie 15:22, 30 March 2020 (EDT)

(unindent)It's been a week now, and I'm finally feeling somewhat better. Energy still low though. We had the a disappointing three votes here, and it looks like there's a majority for having the length of the parent title leading, with a note when the variant should have been a different length based on the word count. This means we won't need an ignore function. Can someone adjust the help text? I'm not sure I can do much the coming days. --Willem 14:33, 6 April 2020 (EDT)

I am sorry, I was going to comment on the R&S page but then I got distracted. I have now added my opinion (also in favor of keeping length values in sync) to the discussion.
Hope the low energy situation will resolve itself soon -- I know how bad it can get! Ahasuerus 22:40, 6 April 2020 (EDT)

See this pub. The Amazon cover for this one is from a host which the software doesn't recognize as Amazon. It links to the host (vice which will give a permissions error if accessed. -- JLaTondre (talk) 08:58, 11 April 2020 (EDT)

This is how Amazon had been linking the ebooks and audio books covers for the last few months (at least) so we have hundreds of them. We should probably just say Amazon on that screen. The code are the same so we can also enforce the old URL but that will mean an additional step for any editor (unless we can do something on the software side). Annie 12:01, 11 April 2020 (EDT)
Thanks, I'll take a look. Ahasuerus 14:40, 11 April 2020 (EDT)
OK, has been added to the list of recognized domains. No more yellow warnings! should be credited whenever a image is displayed. Ahasuerus 16:06, 11 April 2020 (EDT)
There were no yellow warnings anyway - I think it was recognized as a sub domain of It was just the link at the bottom of the pub pages. That begs the question what other patterns we may have with a similar display problem. :) Annie 16:59, 11 April 2020 (EDT)
I pulled all the Amazon domain names from yesterday's database dump and looked at a sample for each. They all display a link to Amazon. -- JLaTondre (talk) 08:52, 12 April 2020 (EDT)

Nitchevo Factory, Vonnie Winslow Crist, Jennifer R. Povey, & Urania Mania are all cover sources that are linked to by the software. However, I'm not seeing them on Template:Image Host Sites. Should they be added? -- JLaTondre (talk) 10:01, 12 April 2020 (EDT)

I believe so. Good catch! Ahasuerus 10:55, 12 April 2020 (EDT)

Words of wisdom

Please take a look at this discussion. You are good at defusing situations. Maybe there are words of wisdom you can add that will help us from loosing an editor that has 1299 verifications across 4 years and who has been doing good work (example). It may be too late, but it is still worth the attempt in my opinion. Thanks. -- JLaTondre (talk) 18:04, 13 April 2020 (EDT)

That was a very well written response. Thank you! At the moment, I can't really think of anything else that I could add. If I think of something tomorrow, I'll throw in my 2 cents. Ahasuerus 22:50, 13 April 2020 (EDT)

Page Counts & Numbers

The publication page count and the publication contents page number fields are becoming a mess.

For page counts:

  • [2] page counts having a () or {} instead of a [] (there are some that might be considered valid, but most are typos)
  • [3] page counts having a - instead of a + (one of them looks like date was copied to page count by mistake; two are intentional, but non-standard)
  • [4] publisher in wrong field
  • [5] field contains "pages" (including an ebook that shouldn't have a page count)
  • [6] unpaginated (and variants) instead of []
  • [7] publication with invalid roman numeral (looks like an inadvertent character at the start)

For page numbers:

  • 37 cases of "del%" which probably indicate users meant for the title record to be removed
  • 162 cases where "&#448;" was used instead of "|". The software doesn't handle the HTML representation (see [8]).
  • [9] & [10] are cases where start & stop ranges where provided for content (maybe a standards discussion for that?)
  • [11] is an oddball case where the ToC uses "Rm" but the actual page numbers do not (per Amazon Look Inside)

There are plenty of non-standard cases also.

The page counts would be a pretty easy clean-up report. It could look for entries that contain values other than numbers, roman numerals, [], + and space. The following REGEX query does that:

select pub_id, pub_pages from pubs where pub_pages REGEXP '[^\]\[0-9ivxlcdm+ ]'; (could leave out "d" & "m" since page counts probably wouldn't get that large)

There would probably need to be an ignore to deal with a few of the non-standard cases.

The page numbers are more complex. I tried a similar query (adding in the additional characters that can be in that field) along with skipping bp, bep, etc. but there is such (valid) variety in publications that there are still a lot of false positives. There may be a way to tighten that up more. Or perhaps only query for the known problems like "del" and "&#448;"? -- JLaTondre (talk) 10:01, 14 April 2020 (EDT)

I see what you mean. I agree that it would be useful to have nightly cleanup reports for invalid page counts and Contents-level page numbers. I agree that the second report could be limited to "del%", "remove" (2 matches) and "&#448;" for now.
I am less sure about the first report. Perhaps we could use the REGEX expression above and see where it takes us without the "ignore" functionality. Once we whittle it down, we could then either further tweak the REGEX or add "ignore" links. Ahasuerus 13:52, 14 April 2020 (EDT)
Done -- see this Community Portal announcement. Ahasuerus 12:47, 17 April 2020 (EDT)


I saw your comment on Biomassbob's page [12].

I was recently looking at some of my non-fiction titles - I read a lot of physics books - and found, to my surprise, that some of them are on the db. Specifically these:

996656 997893 103270 1809399 103249

and other titles by authors who have an Author Record but not the following titles:

Albert Einstein; The World As I See It 99141
Brian Greene; The Elegant Universe; The Fabric of the Cosmos; The Hidden Reality 140278
Paul Halpern; The Great Beyond 21249
Stephen Hawking; A Briefer History of Time 37485
Banesh Hoffman; Einstein 104505
Lawrence Krauss; A Universe From Nothing: Why There Is Something Rather Than Nothing 21148
F. David Peat; Superstrings : and the search for the theory of everything 8478
Simon Singh; Fermat's Last Theorem 37402
Leonard Susskind; The Black Hole War 37512

Whether or not these are eligible for inclusion is definitely a trivial matter for my part, I have no opinion one way or another whether any of them should or should not be in the db, I'm only offering the above to maybe illustrate the issues "new editors" might face - perhaps taking the view that precedence grants justification?

The decision line has a definite quantum uncertainty feel to it; what to think when encountering titles like this?

As physics is so woven into the genre, does that make inclusion / exclusion harder to decide?

The help pages aren't. Maybe addressing the issues there properly first would help in restricting new incoming?

Anyway, I hope this is helpful. Thanks, Kev. BanjoKev 17:27, 16 April 2020 (EDT)

Re: "physics is so woven into the genre", it is indeed the main cause of the current situation. Many SF magazines -- going back to the middle of the 20th century -- have a habit of reviewing popular science books. Our original policy was to enter reviews of non-fiction books, non-genre fiction, and other ineligible titles as REVIEWs. We then created title records for reviewed books. The result was the addition of hundreds of books that wouldn't have been eligible otherwise. When we changed the policy to enter reviews of otherwise ineligible works as ESSAYs instead of REVIEWs, we should have gone back and converted pre-existing REVIEWs to ESSAYs and deleted the ineligible title records.
We already have a cleanup report that looks for "reviews of CHAPBOOK, COVERART, INTERIORART, INTERVIEW and REVIEW records". I think we need another one for "reviewed non-fiction titles". I'll post on the Community Portal later today. Thanks for the reminder! Ahasuerus 16:21, 17 April 2020 (EDT)

Moderator trouble again

Sorry I have to disturb you, but unfortunately Stonecreek is, again, trying to chase an editor away from the database. Perhaps it’s a good idea to have your comments here before things get out of control. For now I will try not to interfere, but my fingers are itching. I'm truly sorry I ever nominated him for moderator. Willem 14:32, 20 April 2020 (EDT)

Will do! Ahasuerus 14:45, 20 April 2020 (EDT)
Done. Ahasuerus 15:59, 20 April 2020 (EDT)
Thanks! --Willem 16:08, 20 April 2020 (EDT)

Chapbooks with multiple fiction

Please see this conversation for another possible clean-up report. Thanks. -- JLaTondre (talk) 13:40, 24 April 2020 (EDT)

Thanks, I'll take a look. Ahasuerus 15:59, 24 April 2020 (EDT)


When editing the title record for an interview, if you click on the help sign next to Interviewee, it takes you to Template:TitleFields:Author and not Template:TitleFields:Interviewee (like it does when you do the same from the publication editor). Since the rules for interviewee are different (record canonical & not as displayed in pub), it would be best of the title one linked to the latter also. Thanks. -- JLaTondre (talk) 18:02, 26 April 2020 (EDT)

Bug 752 has been created. Thanks! Ahasuerus 18:25, 26 April 2020 (EDT)
Fixed! Ahasuerus 09:52, 5 May 2020 (EDT)

Question regarding series details that conflict with author web site

Hello :) I've been on hiatus for a long time, but recently got back to this site now that I've moved and my books are in the same room as my computer. I don't know if you're the correct person to bring this issue to, but I couldn't find the discussion page for mhutchins, the guy I remember helping me from many years ago. If this needs to be posted elsewhere, just let me know.

Anyway, I was adding in some Christopher Stasheff books to my database, and pulled up Christopher Stasheff's ISFDB page for "The Warlock In Spite of Himself" series, and I also pulled up the author's website list of books, <>. There are conflicts between the two, such as series name Rod Gallowglass (ISFDB) vs The Warlock (author website), "Escape Velocity" being included in the main Warlock series (ISFDB) instead of the DDT*Verse series (author's website), and "A Wizard in Abstentia" being #3 in the Rogue Wizard series on ISFDB vs. #1 on the author's website. What is done in this situation? Should changes be made for the books on ISFDB, or just leave it alone?

Thanks for any help you can give. --AndonSage 21:24, 27 April 2020 (EDT)

Welcome back! :-) The best place to ask questions about ISFDB data entry rules would be ISFDB:Help desk. Let me copy-and-paste your question to that page real quick. Ahasuerus 22:16, 27 April 2020 (EDT)

publisher without publications

HI. I am surprised to find publisher Tauchnitz Collection of British and American Authors, now without recorded publications, because I have understood that Publisher entities and their data are lost when they have no publications.

Certainly I contributed the Note. Certainly not the URL, whose target is online images of the May 1905 Complete Catalogue as "British and American". --Pwendt|talk 14:19, 29 April 2020 (EDT)

IT was an HTML problem (someone added /ul at the end. Once cleaned up, everything showed up again, Annie 14:21, 29 April 2020 (EDT)
Thanks for fixing it! Ahasuerus 15:52, 29 April 2020 (EDT)

Hawaii pre-1900

Can you change the check for valid birthplace to allow Hawaii as a country pre-1959 (as Hawaii Territory from 1900 to 1959 and as just Hawaii or Hawai'i before that)? And then add it to the help page :)

I can make this work if I add ",USA" at the end but this is not how treat territories (see Puerto Rico) and sooner and later we will get someone from before 1900. Thanks! Annie 19:02, 8 May 2020 (EDT)

Actually, the cleanup report already has an exception for "Kingdom of Hawaii" :)
The "territory" issue is more complex. Puerto Rico is a Commonwealth and its relationship with the US is that of an "unincorporated territory". Hawaii, on the other hand, was an "incorporated territory", just like Indiana, Arizona, Florida, Oklahoma, and many others places which eventually became states -- see this Wikipedia list. We always add ", USA" when we enter incorporated territories. We are less consistent in our use of the word "Territory", e.g. see this list of authors born in Oklahoma in the 1900s. (And then there is Alaska, which was a "Department", a "District" and a "Territory" before it became a state :-) Ahasuerus 19:27, 8 May 2020 (EDT)
Then can we document these in the Help page? Not everyone knows the US history intimately (and even if they do, figuring out the types of territories can give you a headache). :) Annie 19:50, 8 May 2020 (EDT)
PS: I just added the Russian ones to the help page - I found them years ago in a Talk post and had them since on my page - can you verify that they are still the correct checks in the code? Annie 19:53, 8 May 2020 (EDT)
Will do. I should probably post the current rules on the Rules and Standard page and then document them in Help. Ahasuerus 22:08, 8 May 2020 (EDT)
I have confirmed that the cleanup report logic matches what you have added to Template:AuthorFields:BirthPlace. I have also updated Rules and standards changelog with links to the 2016 discussion and the Help template. I plan to post re: US territories shortly. Ahasuerus 11:20, 9 May 2020 (EDT)
Done. Ahasuerus 12:32, 9 May 2020 (EDT)

Title Dates after Publication - check days

I believe that this still compares only years and months but not days inside of the same month. If so, can it be changed to also check the dates (still ignoring 00 everywhere)? Thanks! Annie 01:26, 10 May 2020 (EDT)

I think it should be doable. I'll take a closer look tomorrow. Ahasuerus 20:38, 10 May 2020 (EDT)

Suggestion: reduce time window of forthcoming books on homepage

Hi, can I point you at this discussion on my talk page between me and Annie:

I know you added the caveat text to the forthcoming books, but I don't think that really explains to an unwary visitor why some of the books showing on the homepage are there. A quick look at an old copy of biblio/ shows that there's a test of maxmonths > 11 when pulling forthcoming books, maybe that could be temporarily reduced to 1 or 2, to avoid showing the spotty medium to far future releases?

No worries if everyone else thinks this is a terrible idea :-) ErsatzCulture 16:22, 11 May 2020 (EDT)

I wonder if it may be a good idea to change it to "Forthcoming and Recent books" for awhile - grab what is in the future this month but if you are under 20, grab the books from earlier this month (and last month if this month does not have enough). That way we will have 20 recent and coming books there. When we start loading Forthcoming again, it will fix itself as it will always have 20 coming ones but hat way we have a backup if something weird happens (like now)... Annie 18:18, 12 May 2020 (EDT)
I agree that the way the front page looks at the moment is odd. On the other hand, I am not sure changing the behavior of the front page -- adding recent books -- for a few months and then changing it again would be the best way to handle the situation. Something to discuss on the Community Portal, perhaps? Ahasuerus 11:27, 13 May 2020 (EDT)
You are probably right. Alternatively we can try adding a week early -- while 2 months out are too much,a week or two should be safe-ish. Let me think on it a bit more. Annie 12:43, 13 May 2020 (EDT)

Forced Reject Page

When a clone results in a forced reject (like this one), the moderator screen does not list who submitted the edit. It lists the contents & then the reject notice. It would be helpful if the submittor information was still listed. Often, I will want to leave the submitter feedback and having the link to their wiki page would be useful. Thanks. -- JLaTondre (talk) 08:59, 24 May 2020 (EDT)

Done! Ahasuerus 13:01, 24 May 2020 (EDT)
Thanks. -- JLaTondre (talk) 16:28, 29 May 2020 (EDT)

Series labels

If a series has only short stories and 1 non-fiction book, the whole series shows up under "Nonfiction Series" on the author page. That does not sound like something we want to happen. See this author and the "Dragon Masters" series. Thanks! Annie 03:27, 27 May 2020 (EDT)

Hm, an interesting scenario. There is a hierarchy of title types which controls what kind of series is created on Summary pages. At this time NONFICTION titles outrank SHORTFICTION titles in this hierarchy. Perhaps it should be the other way around. Ahasuerus 16:33, 27 May 2020 (EDT)
We will have more and more of these as we bring the juvenile "novels" to code and all kinds of companions popup. The first collection will fix them but not all of them get collected, especially these 96 pages ones for very young readers... Annie 16:45, 27 May 2020 (EDT)
While we are on the topic of reordering these, I won't be very upset if the chapbooks get kicked towards the bottom of the page... Probably should post in Community. Annie 21:50, 27 May 2020 (EDT)

The Banner Art credits page states that new art should be brought to the attention of Alvonruff. Is that still the case? ../Doug H 15:01, 27 May 2020 (EDT)

Al has been mostly unavailable, but there is no harm in sending him an e-mail. If he can't help, then I would post about it on the Community Portal. Ahasuerus 16:31, 27 May 2020 (EDT)
He's left no email. Portaling we go. ../Doug H 16:39, 27 May 2020 (EDT)
You may be able to contact him by pulling up his User page and then clicking on the E-mail this user link on the left. Ahasuerus 16:45, 27 May 2020 (EDT)
This is where you get "This user has not specified a valid e-mail address, or has chosen not to receive e-mail from other users."... Annie 16:47, 27 May 2020 (EDT)
Ah, OK. He used to have his email address posted on his User page, but it's no longer available. I'll respond on the Community Portal. Ahasuerus 16:50, 27 May 2020 (EDT)

Variant Date Display

When a variant has a different title, the publication listing will display the date of the variant and the parent (assuming they are different than the pub). However, when the title is the same, but the author credit is different, it doesn't which results in a display like this one. It would be nice to have the dates in these cases also. Perhaps like:

  • The Envoy, Her • (1951) • short fiction by H. B. Fyfe [as by Horace Brown Fyfe]

and if the variant was first published in a different year:

  • The Envoy, Her • (1951) • short fiction by H. B. Fyfe [as by Horace Brown Fyfe (2010)]

-- JLaTondre (talk) 08:24, 31 May 2020 (EDT)

An interesting thought. I have been thinking of a related issue recently. Consider Encrucijadas del Espacio, a Spanish translation of one of Groff Conklin's anthologies. Here is how it displays Fyfe's "Let There Be Light":
The display order is:
  • Variant/translated title
  • Canonical author name
  • Canonical title
  • Alternate name used by the variant title
It's a rather weird order. It looks even odder when dealing with non-Latin alphabets, e.g.:
in this Russian anthology.
I wonder if an alternative, two-line, arrangement (inspired by Tercera Fundacion) may be better:
Something to discuss on the Community Portal, I guess. Ahasuerus 10:06, 31 May 2020 (EDT)
It looks even weirder if there is a series in a different alphabet from the title in that line... I like the 2 lines idea. It also will give us space to add the series in the "story" language (if/when we get around to supporting that)... Annie 03:55, 2 June 2020 (EDT)
At first thought, two lines isn't my favorite idea as publication listings already can be pretty long. However, I know width vs length is dependent on ones preferred screen size. Best thing would be to mock up some ideas. I'll put together a table that has all the different possibilities (title, author, dates - plus translator & series as should be looking forward) with examples of what they look like today and post that on the community portal. We could then see what suggestions there might be. That will probably take me a day or two. -- JLaTondre (talk) 07:41, 2 June 2020 (EDT)

Audio Book Format Cleanup Report

This report shows pubs that contain a narrator template in the pub notes, but are not audio formats. There are a couple of valid cases (the other and the hc are both ones with a both a book and an audio version -- not sure if there is a standard on how that should be handled), but most are not (one looks like translator should have been used instead, the others look like audio books). Perhaps a clean-up report that looks for publications with either a narrator template or an Audible-ASIN and do not have an audio format is in order? I've already pinged the verifier of the two primary verified versions. I'll clean-up the others once you take a look. -- JLaTondre (talk) 20:25, 9 June 2020 (EDT)

I think it's a good idea. FR 1357 has been created. Thanks! Ahasuerus 13:00, 10 June 2020 (EDT)
If we are going to create reports for audiobooks, can we add one more: "audio format without "Narrator" template"? While I can fish most of them out via search, I cannot find the ones with no pub notes at all for example. :)Annie 13:54, 10 June 2020 (EDT)
Let's see... 8302 audio books without a narrator. Ouch! Let me create an FR... Ahasuerus 14:29, 10 June 2020 (EDT)
FR 1358 has been created. Ahasuerus 15:34, 10 June 2020 (EDT)
Done and done! Ahasuerus 12:33, 11 June 2020 (EDT)

Jo L. Walton has granted ISFDB permission to link to cover image on (see here). Thanks. -- JLaTondre (talk) 08:27, 21 June 2020 (EDT)

Done. Ahasuerus 15:00, 21 June 2020 (EDT)

Is it possible for you to move a PV record to another publication?

Hello Ahasuerus! Markwood wants to delete a duplicate pub with this submission - I happen to agree, but unfortunately, there's one PV by ScottLatham. So I was wondering if you could move that PV record to the other, UK edition publication record, so as to preserve this user's PV (even though he's no longer active) once the duplicate gets deleted. Thanks! MagicUnk 15:56, 21 June 2020 (EDT)

It's possible, but I try not to make manual database changes because they can be dangerous. I am also not sure that the primary verifier would be happy to see his verification linked to a pub which may or may not be identical to the one that he originally verified.
We may want to have a Community Portal discussion of different ways to address this issue. Some options that come to mind:
  • Create a bureaucrat-only option which would let bureaucrats move a primary verification to a different publication record as requested above
  • Make a policy decision re: this type of scenarios without changing the software. For example:
    • Add a line to the Note field indicating that the record is a likely a duplicate of another record and link to the latter
    • Delete the likely duplicate record and leave a note on the primary verifier(s)' Talk page, including a copy of what was deleted
Ahasuerus 20:37, 21 June 2020 (EDT)

Help Link Updates

At the top of the publication editor screen, it has:

Help on editing publication records
List of supported templates and HTML tags in notes

At the top the new publication screen, it has:

Help on adding new publication records: Help:Screen:NewPub

The second draws the eye more than the first, but the templates & tags is also applicable to new publications. How about combining the two with the following for the Publication Editor:

Help on editing new publication records:
  • Help:Screen:EditPub
  • List of supported templates and HTML tags in notes

and the following for new pubs:

Help on adding new publication records:
  • Help:Screen:NewPub
  • List of supported templates and HTML tags in notes

Thanks. -- JLaTondre (talk) 09:15, 28 June 2020 (EDT)

Good idea! FR 1361 has been created. Ahasuerus 14:30, 28 June 2020 (EDT)
Done! Sorry it took so long -- I have been mostly working on Fixer the last few months. Ahasuerus 17:08, 12 August 2020 (EDT)

Double Verification

We have a publication with a double verification (see this discussion). Not a big deal, but as the software theoretically is designed so you cannot do that, I thought I'd point it out in case there is a bug you want to look at (maybe caused by a double click on the verify button). -- JLaTondre (talk) 17:37, 11 July 2020 (EDT)

It looks like the problem has been corrected on the main server, but I can see it on the development server. The two verifications happened during the exact same second, so I suspect that it was a timing issue. Luckily, there is only one other "double verified" pub in the database -- The Time Dweller. Ahasuerus 18:43, 11 July 2020 (EDT)

Server config email

I tried to email this to you, but it got blocked due to policy content restrictions....

Got it.  So the code doesn't ever use more than one query parameter (and then uses space-encoded-to-+ to
separate multiple arguments).

I think you could do this:

<IfModule rewrite_module>
    RewriteEngine On
    RewriteCond %{QUERY_STRING} ^(.*)(?:^|&)fbclid=(?:[^&]*)((?:&|$).*)$ [NC]
    RewriteRule ^(.*) $1?%1%2 [PT,NE,L]

PT will run it through the processor again without going back to the client via a redirect, so it will be
invisible to the link-clicker.  I would also use NE to avoid potential double-encoding.

If you wanted to do something more strict, such as insisting that a query string is all-numeric and either
tossing anything that's not or truncating at the first non-numeric (or issuing an error), you can do that
sort of thing, too.  If you know, for example, the ISFDB code will never have multiple query parameters,
you could simplify the above to strip any additional arguments:

    RewriteCond %{QUERY_STRING} ^([^&]*)&.*$
    RewriteRule ^(.*) $1?%1 [PT,NE,L]


p.s. I don't know if you have any log space considerations, but mod_rewrite can be awfully chatty.  I
like to add:

    LogLevel rewrite_module:crit

to get just the critical errors.

--MartyD 16:32, 14 July 2020 (EDT)

Thanks for looking into this! One thing that I should have mentioned is that the Advanced Search logic is different. It extracts the search parameters from cgi.FieldStorage , which uses a different format. It should be valid to include "?fbclid" in a search string when searching for copy-pasted Facebook links. For example, this Advanced Author Search discovers that our Carol Emshwiller record has a link with an embedded "?fbclid" value.
Re: "insisting that a query string is all-numeric", the vast majority of legal values are numeric, but there are exceptions. For example, Summary and other author pages accept both author record IDs and author names as arguments (although non-Latin-1 names don't work correctly.) Publication pages accept both publication record IDs and non-numeric "publication tags" (not to be confused with title tags.) They also accept an optional second parameter, which is non-numeric.
We may be able to create additional Apache rules to accommodate these exceptions, but I suspect that the amount of work, especially the amount of testing that would be required, wouldn't be that much lower than what would be needed to implement the same functionality in Python, at which point the advantages of doing it in a Web server-independent, consistent and maintainable way would outweigh the advantages of a quick Apache-specific fix. Ahasuerus 11:25, 15 July 2020 (EDT)
Advanced Search should be encoding any equals sign, so I'm pretty sure it would be safe enough and would leave any attempt to search for "fbclid=" alone. But if you're worried, you could limit the effect to just Facebook's parameter added to the end of normal links. For example:
     RewriteCond %{QUERY_STRING} ^([0-9]+)&fbclid=^&+$ [NC]
     RewriteRule ^(.*) $1?%1 [PT,NE,L]
That would require the parameter string start with a number, then have "&" and the Facebook click ID thing, and then nothing else.
Anyway, yes, it would ultimately be better if the code did its own input validation. One other approach you could consider that is more work up front but that would be more flexible down the line would be to have an "external links" form that does not use query parameters and does not expose the underlying implementation technology. The web server could convert them and for now ignore any provided parameters. E.g., instead of a link could use Not suggesting you go there, just food for thought. --MartyD 07:46, 16 July 2020 (EDT)
An interesting point, thanks!
Unfortunately, a blanket "parameter string start with a number" requirement wouldn't work for author and publication pages since they accept non-numeric values as well as numeric IDs.
The more I think about doing it at the Web server level, the more concerned I become about possible obscure permutations, especially considering the fact that I know very little about Apache. One of the worst things about getting older is how much longer it takes to learn new things :-( Ahasuerus 15:27, 16 July 2020 (EDT)

Missing cover art for the Tor Double #0 "The Other Sky/ The House in November"

The dos book here has only one of the covers shown. I have a copy, and I could scan it, but I don't know how to get it into an image of both covers, as I see in other dos pubs. Jack Sjmathis 16:39, 15 July 2020 (EDT)

A publication record can have only one image directly associated with it. That image can be either a single cover scan or a composite image, e.g. a collage of two different scans. There are different tools that let you combine multiple images into one. If you use Windows, Photoshop is the gold standard, but it costs money. I believe Microsoft Paint, a free program that comes with Windows, can also do it -- here are the instructions -- but I haven't touched it in many years. Ahasuerus 18:04, 15 July 2020 (EDT)

Also, the cover shown looks like it's in better shape than mine, and I don't want to lose it. Can you tell me how to do this? Thanks! Jack Sjmathis 16:39, 15 July 2020 (EDT)

There is no way to have a publication record display two different images directly. The closest approximation that I can think of would be creating a composite image as per the discussion above, uploading it to the ISFDB Wiki (see the Special:Upload page), then adding a link to the newly uploaded image to the publication's Note field. Ahasuerus 18:04, 15 July 2020 (EDT)
Since the book already has an ISFDB image, it would be better to upload the composite image as a replacement for that image (using the "Upload a new version of this file" at Image:BKTG08857.jpg). The template would need to be tweaked for the second artist, but manual upload would require doing that also. -- JLaTondre (talk) 18:37, 15 July 2020 (EDT)

Non-Database Web Pages

When I was looking for something, I used a Google search and stumbled across this page. There seem to be a number of these (see this query). As these are static, out of date pages, should they be removed? Or the robots.txt setup to tell search engines to ignore them? -- JLaTondre (talk) 18:40, 15 July 2020 (EDT)

They are indeed static pages created during the 1990s and 2000s. Most of them are obsolete, but a few include records which the main ISFDB database doesn't have. For example, this Wandering Jew bibliography includes a number of obscure SF stories and poems. I am thinking that we should archive the ones that are clearly out of date (and add them to the SVN repository?) and ask for volunteers to incorporate the rest into the database. Ahasuerus 08:35, 16 July 2020 (EDT)

Variant Title Length Mismatches

The cleanup report is finished. Took a bit longer than expected, but I have a fiery excuse. Is there merit in extending the report to short fiction with no set length? (parent title has a set length, variant does not, or the other way around). I corrected a few thousand of these while working on the report, most were translations. --Willem 14:44, 27 July 2020 (EDT)

Excellent, thanks a lot! Hope things have settled down after the fire!
Re: extending the report to SHORTFICTION variant/parent pairs where one title has a "length" value and the other one doesn't, well, let's think about it. Based on the decision made a few months ago, we don't allow length mismatches, right? The only reason why we needed a cleanup report for the existing mismatches was that there was no easy way of telling which value was correct programmatically, hence the manual review. The fact that you were able to clean everything up without running into an "edge case" with a valid mismatch suggests that our decision was correct.
Now, for variant/title pairs with one title in possession of a length value and the other one length-less, I think we have two separate scenarios:
  • The parent title has a length value, but one or more of its variants doesn't (10,100 matching parents in the database)
  • One or more of the variant titles have a length value, but the parent doesn't (4,000 matching parents in the database)
My thinking is that the first scenario can be handled programmatically: simply take the length value of the parent title and assign it to any variants which do not have one. The second scenario may need a manual review since translated variants may have been assigned a length value based on the word count of the translation.
What do you think? Ahasuerus 15:36, 27 July 2020 (EDT)
Sounds good, especially if it saves 10.000 edits. I can do the 4000 from scenario 2. By the way, I did run into a few titles that needed separation (excerpts varianted to the complete story, different versions of stories varianted to each other etc.) Not many though, and most were easy to repair. --Willem 16:22, 27 July 2020 (EDT)
Done! The output is currently limited to the first 1000 matches. Ahasuerus 18:07, 27 July 2020 (EDT)
Considering that we can never have the values in a child different from the parent per policy, it may make sense to change the software and have this field behave like the series field - once those 4,000 ones are sorted out of course (so we do not propagate errors in them). Annie 17:39, 27 July 2020 (EDT)
I have been thinking along the same lines for the last few hours. The proposed behavior would be similar to what we currently do for series names and series numbers: only parent titles can have them, so the variant creation process moves the series information to the parent record.
It's certainly doable, but we may want to discuss certain issues on the Community Portal first. For example, removing "length" values from VTs would affect Advanced Search. Is that the expected behavior? Ahasuerus 18:13, 27 July 2020 (EDT)

Author image URLs supported

Apparently, Amazon is starting to use another URL structure that we don't support. See this submission. Do we want to add support for it? ···日本穣 · 投稿 · Talk to Nihonjoe 15:57, 29 July 2020 (EDT)

Let me take a closer look... Ahasuerus 16:09, 29 July 2020 (EDT)
Spot-checking various Amazon pages suggests that this is a new URL format supported by Amazon. FR 1365 has been created. Thanks for identifying the issue! Ahasuerus 18:17, 30 July 2020 (EDT)
Done ! Ahasuerus 15:37, 31 July 2020 (EDT)
Awesome! I've approved the submission. ···日本穣 · 投稿 · Talk to Nihonjoe 12:40, 3 August 2020 (EDT)

Unexpected Unmerge Result

I had an unexpected unmerge result. For this pub, I unmerged the novel record from this title record in order to move it over to this title record (have been sorting out publication vs. database errors). When I unmerged it, it converted it from being only by H. Rider Haggard to being a novel by H. Rider Haggard and Baron Friedrich de La Motte Fouqué. Given that the container title is an omnibus, I wouldn't think the novel should "inherit" the ombibus credits for an unmerge. Thanks. -- JLaTondre (talk) 16:21, 29 July 2020 (EDT)

It had always worked that way - it picks up a title and the authors from the omnibus... When it is an omnibus, removeTitle works better. Very annoying indeed. :) Annie 16:44, 29 July 2020 (EDT)
As I recall, the current functionality was something that was requested ca. 2010. The way it worked in 2006-2010 was causing other issues, but I am afraid I no longer remember what they were :-( Ahasuerus 18:09, 30 July 2020 (EDT)
If we cannot/would not fix it, can we remove the omnibuses from the list of things you can unmerge from when you are unmerging anything BUT the omnibus record itself? That will force people to use the removeTitle plus edit/import instead. With a note on the unmerged screen that this is how you do it i we suppress the showing of omnibuses. Annie 18:15, 30 July 2020 (EDT)
I'd have to look into it. There may be other dependencies and permutations which I can no longer recall. Back in 2010, I had to code something like 3 or 4 patches to account for different scenarios. Ahasuerus 20:47, 30 July 2020 (EDT)

2020-08-08 backup

Is today's 5.5-compatible the correct file? The file name is backup-MySQL-55-2020-07-25. Thanks. -- JLaTondre (talk) 18:42, 8 August 2020 (EDT)

Oops! Sorry, it was a copy-and-paste error, now corrected :-) Ahasuerus 19:07, 8 August 2020 (EDT)

Primary Verification Warning on Title Removals

It would be good to add a primary verification warning on title removal submissions. Similar as shown on publication edit submissions. Currently, when moderating a title removal, there is nothing to indicate that the pub is verified which makes it easy to miss edits that should have been run by the verifier. Thanks. -- JLaTondre (talk) 19:36, 17 August 2020 (EDT)

Good point. FR 1367 has been created. Ahasuerus 10:15, 18 August 2020 (EDT)

A downloading data enquiry

I've had a private Messenger enquiry from a new editor Ybahjatt who wishes to use the ISFDB to download information. Specifically his enquiry is as follows: "Do you know how I can download from ISFDB? I am trying to download all data related to the Hugos as I am trying to do some inclusivity study on the Hugos. Specifically this is the data I am trying to get: Year, Award, Name of Nominee, Number of nominations, Number of Votes, Gender of Nominee, Year of Birth, Age of Nominee when nominated, Ethnicity of Nominee, Religion of Nominee, Country of Nominee, Original language of the Nominated work, Nominated language of the Nominated work, profession of Nominee."

I've already advised him that some of those categories are off-limits to the ISFDB or we simply do not have the degree of information he wants, but that others are searchable. For those that we have, I think his enquiry is asking is there a way he can download information quicker than doing individual searches? Thanks. PeteYoung 17:31, 29 August 2020 (EDT)

Apologies for intruding, but I have a slight awareness of this. I was DMed by someone I follow on Twitter, who in turn had been contacted by Yassar Bahjatt (or someone on behalf of him, I don't know the details). I mentioned that all the data analysis stuff I'd done was from public sources (mainly ISFDB, but also dragging in stuff from Wikipedia, Twitter etc), and that I have various scripts/reports that are publicly accessible via a GitHub repo, and that whilst I wasn't especially interested in getting involved in whatever he/they are doing, I'd help anyone with issues related to my scripts. I didn't hear any more, but this blog post a week or so later seems to cover much the same territory as you mention.
I don't know any more than the above, but I thought it worth mentioning, if only to avoid the possibility of people duplicating effort... ErsatzCulture 17:49, 29 August 2020 (EDT)
He can grab the full DB backup and extract from there directly. Annie 17:54, 29 August 2020 (EDT)
That's right -- our backups are available on the ISFDB Downloads page. Author- and award-specific tables should contain some of the requested data elements. The following data elements may be unavailable, imprecise or ambiguous:
  • Number of nominations -- not captured although sometimes mentioned in the Award Note field. The Hugo site may have more information.
  • Number of Votes -- not captured although sometimes mentioned in the Award Note field. The Hugo site may have more information.
  • Gender of Nominee -- not captured, but the Note field may contain clues.
  • Age of Nominee when nominated -- since nomination dates are generally not captured, calculations based on the author's date of birth and on the year of the award may be off by a year, possibly even more.
  • Ethnicity of Nominee -- not captured
  • Religion of Nominee -- not captured
  • Country of Nominee -- only the country of birth is captured in the structured way. The Author Note field may contain clues like "UK-based author".
  • Nominated language of the Nominated work -- can be derived from the language of the translated title that the award record is linked to. Some older award records may be still linked to the original title, though.
  • Profession of Nominee -- not captured in a structured way, but author profession(s) are occasionally mentioned in the Note field. Sometimes the data can be ambiguous, e.g. "photographer" may refer to a professional photographer or an amateur photographer with a couple of occasional sales.
Ahasuerus 18:29, 29 August 2020 (EDT)

Re. quoting synopses, how about enabling cite attribute support on <blockquote> and/or having a template for them?

Hi, just reading the synopsis discussion on the Rules & Standards page, and a somewhat related thought occurred, but as it has technical implementation issues, I thought it best to ask you here first - could the system do either or both of:

  • Support the cite attribute of the <blockquote> element, so that an editor could do <blockquote cite="">Synopsis text goes here</a> to indicate the source in a semantic way? I just tested it on a local instance, and passing the cite attribute throws up a slightly misleading "Unrecognized HTML tag(s) present" on the post-submit page, I dunno if the cite attribute is removed altogether? This would probably need some way for a moderator to see the URL being cited, and (probably) tweaks to the CSS rules for that attribute to not indent the way it does by default.
  • Have a template for quoted synopses, that would render to a <blockquote> similar to the above. I've not looked at the code, but this might be non-trivial, as the current set of templates (as far as I can see) all only take a single parameter, whereas you'd want two here, one for the cited source, and one for the synopsis text). One possible useful thing about having a template is that you could in theory allow cited sources that aren't webpages, for these you could maybe render differently i.e. render the source as some text, rather than a cite attribute on the <blockquote> element.

NB: I'm not advocating that a quoted synopsis absolutely must be submitted in either of the above ways, just that it'd be a nice ability to have. I'm happy to look at what might be involved in implementing either of the above, if needed. ErsatzCulture 18:26, 5 September 2020 (EDT)

A few things come to mind:
  • Adding support for the cite attribute when using the "blockquote" element. Currently only "a", "table" and "td" can have attributes, but it would be trivial to add "blockquote" to the white list.
  • We use CSP (Content Security Policy) which blocks many types of things. We'd need to see if "cite" would cause any issues.
  • Entering the HTML code above into a random Synopsis field verbatim generates an HTML validation error. We'd have to figure out what's going on there.
  • At one point I looked into adding support for multiple template parameters. It wasn't a straightforward proposition.
Ahasuerus 15:36, 6 September 2020 (EDT)

Fixer and Spectrum Literary Agency / Bujold novellas

Hi, not sure if this might be better directed at Annie in the first instance, but I'm sure she'll see this regardless ;-)

Anyway, I just noticed that this Bujold novella published in May isn't in the database. There was a similar issue with one of her novellas from last year which first got added (via Fixer) when it came out in physical a year after the initial ebook pub, and I manually submitted the original pub when I saw that it was missing.

It looks like this 2020 novella is also a digital-first title (ebook and Audible), but given Bujold's rep/history, I'm wondering if there's anything that could/should be done to make them more likely to get processed by Fixer? (I've not checked, but the browser extension I wrote to highlight stuff that isn't known to ISFDB and/or Fixer implies that the ebook isn't in Fixer's database, albeit I haven't yet installed the latest data you shared a week or so ago.)

A very quick glance at some years' output of Spectrum Literary Agency shows they put out pubs by pretty well known authors, so unless the stuff that gets into ISFDB is atypical of their wider output, it would naively seem to be to be a safe bet to add them to a list of priority publishers? (Assuming my understanding of how Fixer operates is correct.) ErsatzCulture 10:43, 25 September 2020 (EDT)

As of April, we lost the eISBNs when the amazon API changed. Which means that ebooks are added only when we know about them and specifically pull them one by one from the dB or I ask for specific authors and so on. Sorting for ASINs is being worked on by Ahasuerus but it takes time to get Fixer fixed - we had never been setup for it. So a bit more patience will be required here. :) We know we are missing the ebooks and books with ISBNs starting with 979 - they will get added as soon as we are able to sort them. Which should be soon I am told. In the meantime as of a few days ago we can at least pull ASINs by publisher name so let me finish my current batches and will see if I can get this specific publisher sorted out. :) Annie 11:31, 25 September 2020 (EDT)
Thanks for chiming in, Annie! And yes, I am working on Fixer's ASIN-specific modules as we speak. Ahasuerus 11:55, 25 September 2020 (EDT)
Thanks both. I'd noticed that recent US titles were only getting physical pubs created and not ebooks, but as I've been almost entirely focussed on the UK side of things, I'd not given any thought or consideration as to why exactly that was the case for US stuff; I'd assumed it was maybe some prioritization thing, rather than needing manual intervention. ErsatzCulture 12:23, 25 September 2020 (EDT)
At this point we are adding ebooks in two ways: either they were in Amazon pre-April (so we have an eISBN from earlier and we can sort them based on that) or when I (or Ahasuerus) specifically go for it based on ASIN (or as part of my "known authors" project or a few other ways we pick up cleanup targets). Same for Audible books. And some got their ebooks/Audibles manually during my "check if the book came out" reviews or (manually again; Fixer-driver based on ASIN lately) as a clone from me when it was there and I was adding the paper one anyway. Between the craziness of the publishing industry in 2020 and the Amazon API change (which got us unprepared), it had been challenging. :) But at least progress is being made and we had been discovering a lot of ways to streamline the process (not to mention all the books that had been shoved into corners and were brought back to the light) so there is some silver lining in the whole thing. Annie 12:33, 25 September 2020 (EDT)

Gunner Cade

I'm intending to move Gunner Cade from the title with Both Kornbluth and Merril credits to the one with the Cyril Judd pseudonym. Per this page under the Alternate Names bullet, we should use the original name when both names are present. Please let me know if you have any concerns. Thanks. --Ron ~ RtraceTalk 06:54, 2 October 2020 (EDT)

I verified this pub back when the rules were different, so the change works for me. Ahasuerus 11:08, 2 October 2020 (EDT)

Not this August

While working on art books, encountered another publisher error. Not This August has cover art by Barclay Shaw that is credited to Tom Kidd/Gnemo -- I added credit and notes. Susan O'Fearna 17:03, 22 October 2020 (EDT)

Thanks! Ahasuerus 21:16, 22 October 2020 (EDT)

Daily Outage

Every morning, there seems to be a 5 minutes (or so) period where the ISFDB is not available. I assume there is some sort of maintenance going on. While regulars are probably used to it at this point, any non-regulars would not be. Would it be possible to display a page saying "maintenance underway, check back in 5 minutes" vs. totally not working? -- JLaTondre (talk) 09:43, 31 October 2020 (EDT)

That's right, the daily backups run daily starting at 9:30am.
There are two parts to the backup process. The first part makes the database unresponsive, so any ISFDB Web page that a user tries to access during that time will hang. During the second part a "Database connection" error is displayed.
I am going to change the error message from "Database connection error" to "The ISFDB database is currently undergoing maintenance. Please try again in a few minutes", which should help with the second part. The first part is a more difficult problem. I'll see what I can do. Ahasuerus 12:36, 31 October 2020 (EDT)
The error message has been updated. Ahasuerus 12:59, 31 October 2020 (EDT)
Thank you. -- JLaTondre (talk) 17:51, 31 October 2020 (EDT)
My pleasure! Ahasuerus 17:57, 31 October 2020 (EDT)

Variant of Variant

This variant has two variants. Looking at the edit histories, the three variant submissions were submitted close together. The logic to prevent this on edits wouldn't have caught it because 2795475 wasn't a variant when the other two submissions were made. Maybe there needs to be a check on moderation and not allow a variant to a variant to be approved? -- JLaTondre (talk) 17:22, 7 November 2020 (EST)

Hm, so basically a "hard reject" "this submission is no longer approvable" error message for submissions which would result in the creation of variants of variants? Ahasuerus 18:29, 7 November 2020 (EST)
Yup, with an explanation of why. Thanks. -- JLaTondre (talk) 08:49, 8 November 2020 (EST)
OK, FR 1377 has been created. I plan to start working on the accumulated FRs tomorrow, once the next Fixer download starts. Ahasuerus 08:45, 9 November 2020 (EST)
Done. In addition, "Add variant Title" was updated to use the same auto-reject logic. Ahasuerus 15:54, 11 November 2020 (EST)
Thanks! -- JLaTondre (talk) 19:15, 12 November 2020 (EST)

Display differences

Hello Ahasuerus, just an FYI. I noticed that the language is not displayed in square brackets for omnibuses if the title series works only are listed, as opposed to being part of the author works listed. For example The Call of the Forgotten (sub-)series here - Dutch is shown as language for the omnibus. Compare that with this series display, and you'll notice that Dutch is not shown. (I didn't check but this might happen for anthologies too). Regards, MagicUnk 11:15, 13 November 2020 (EST)

There is a subtle difference between the way the software handles languages on Summary Bibliography pages and on Series pages. On Summary Bibliography pages, languages are displayed for:
  • VTs whose language is different from the parent title's language
  • canonical titles whose language is different from the displayed author's "working language"
On Series pages, VTs are handled the same way, but canonical titles never have a language displayed. The reason is that we don't have a single "working language" because series can be -- and often are -- written by multiple authors who may have different working languages. This is why The iron fey: Deel 5-7 says "[Dutch]" when it appears on Julie Kagawa's Summary page, but has no language displayed when it appears on the Series page.
That said, one could argue that the software which handles Series pages should be upgraded to display the language name for canonical titles if the language is different from the working language of at least one of its authors. We could certainly start a discussion on the Community Portal. Ahasuerus 15:57, 13 November 2020 (EST)
Ah, I understand - oh, the subtle delights of software design :). Not sure if it's worthwhile to start a discussion right now, as defining at least one of its authors does look like it will end up in discussions about corner cases... MagicUnk 15:54, 16 November 2020 (EST)

I was looking around for ideas on Banner Art and (long story short) ran across this discussion. The image is still on the wiki, but the link hasn't been added to the publication. I don't know if you're thinking has changed in 12 years. Anyway, back to looking for suitable magazines to add to the Banner Art, suggestions welcome. ../Doug H 16:48, 21 November 2020 (EST)

No new ideas re: banner art, I am afraid. I am colorblind, so I find that I tend to pay less attention to art than most people.
Re: the 2008 discussion of NSFW images, much has changed in the last 12 years. We have a lot more European covers than we did back then and some of them can be fairly NSFW by NorAm standards. It's hard to come up with a consistent approach when both your data and your users come from multiple countries with different workplace standards. I don't think we have discussed this issue lately. Ahasuerus 17:19, 21 November 2020 (EST)
Re: banner. I've no problem with the art, my problem is identifying "seminal magazines of the genre" to "homage". I came to it in the 60's and never really looked back. Also tended to avoid magazines. ../Doug H 19:29, 21 November 2020 (EST)
Something along the lines of:
although linking to the source documents presents a bit of a problem as they haven't been entered. I guess I could point to the TITLE where they are noted. ../Doug H 15:38, 23 November 2020 (EST)
Looks pretty good! Ahasuerus 16:48, 23 November 2020 (EST)
Ah, I see. I could certainly create a list of the more important genre magazines, but "seminal" can be hard to pin down. For example, take Captain Future. The writing was formulaic and hasty since Hamilton was churning out one (short) novel per month. The art was fairly crude -- see the cover of the first issue. Etc. And yet one could argue that the magazine was quite influential. It was aimed at older children and younger teenagers and managed to draw many of them to science fiction. While "Astounding" was clearly a vastly better magazine, its adult tone and advanced science articles didn't work for children while "Captain Future" was just right for that age group. Does it make Captain Future "seminal"?
That said, I see one glaring omission on the Banner Art Credits page. We have Startling Stories, but we don't have its sister publication Thrilling Wonder Stories. The two of them gave Astounding a run for its money in the late 1940s. Ahasuerus 10:58, 22 November 2020 (EST)
I'll check out the art work and see if there's anything I can work with. In the meantime, I've got something for Magasin d'Éducation et de Récréation going. As for 'seminal' - the word was used for the original art. We might get agreement on some other term. ../Doug H 23:16, 22 November 2020 (EST)

Author Names with Double Quotes

I believe the software is supposed to convert double quotes in author names to single quotes? It doesn't appear to be doing that on new pub submissions. In this submission, you can see the story on page 13 had double quotes. It was only when I edited it to correct other things that it automatically converted them. -- JLaTondre (talk) 17:58, 25 November 2020 (EST)

I suspect that it's Bug 666, "Author names can include double quotes ", which reads:
  • Normally double quotes are automatically converted to single quotes at submission creation time. However, if a double quote is HTML-encoded, it can get past the filters and be included in the author's name.
Let me see if I'll have more luck with this bug this time around. In the past it proved to be rather slippery. Ahasuerus 18:05, 25 November 2020 (EST)
They might have been smart quotes, which are encoded differently than regular double quotes. They are the bane of any coder's existence. ···日本穣 · 投稿 · Talk to Nihonjoe 18:25, 25 November 2020 (EST)
The ISFDB software has an input filter which replaces every recognized type of "smart" quotes with regular single or double quotes (depending on the type of "smart" quote used.) We also have a separate filter for author names which replaces double quotes with single quotes. It's possible that certain types of double quotes are not caught by the second filter because of an unexpected interaction between the two filters. Ahasuerus 18:32, 25 November 2020 (EST)
It's good to know the scrubbers are active. Don't want a "Little Bobby Tables" incident. (^_-) ···日本穣 · 投稿 · Talk to Nihonjoe 18:42, 25 November 2020 (EST)
The "Little Bobby Tables" problem is a different can of worms, which we addressed some years ago :) Ahasuerus 20:41, 25 November 2020 (EST)

Mismatched template braces

This submission has "Mismatched template braces" warnings on both the title note and pub note. However, that seems to be an error in the detection logic. Both notes have {{Tr|{{A|Lore Straßl}}}} which is two sets of opening and closing. There are no other braces. -- JLaTondre (talk) 10:17, 27 November 2020 (EST)

Let me take a look...
OK, I see what's going on. Nested templates are not formally supported, but a single template nested inside of another template "sort of" works and has been used by our editors. The "Mismatched template braces" logic doesn't account for this fact. I'll see what I can do. Ahasuerus 10:53, 27 November 2020 (EST)
Fixed -- see the Community Portal announcement about the impact of the change. Ahasuerus 13:55, 27 November 2020 (EST)

Mark Brake / Dr. Mark Brake

I am trying to variant Dr. Mark Brake as an interviewee to Mark Brake, in an interview titled "Dr. Mark Brake", so that it can appear on Mark Brake's biblio page. However I'm unable to do so: it appears only the interviewer can be varianted. Summary here. Please advise! Thanks. PeteYoung 14:39, 7 December 2020 (EST)

I am afraid it's a software limitation. To quote Help:Screen:NewPub#Interviews:
  • Interviewee - The canonical name of the person being interviewed, whether or not it matches the form(s) of the name used in the interview. This is done so that the interview is displayed on the interviewee's summary bibliography page.
  • Interviewer - The name of the interviewer. This name should not be corrected if mis-spelled, or given in a variant form; it follows the same rules as the rules for author name in a content record.
It would require a lot of work to make the software support variants/pseudonyms for interviewees :-( Ahasuerus 14:50, 7 December 2020 (EST)
OK, thanks. One less thing to worry about! PeteYoung 16:29, 8 December 2020 (EST)

Sweet Reincarnation

It looks like Fixer grabbed some manga. All three volumes this series are manga adaptations of the first three novels (you can use "Look inside" to verify this). Is there a way to have fixer check the categories to see if "manga" is in the name of any of them, and then flag the entries or something? ···日本穣 · 投稿 · Talk to Nihonjoe 18:05, 11 December 2020 (EST)

I started adding the novels. There are 16 of them so far (though I've only added 5). You can see the series here. ···日本穣 · 投稿 · Talk to Nihonjoe 18:42, 11 December 2020 (EST)
Considering how often light novels and novels based on manga (and sometimes ones that have nothing to do with manga...) have "Manga something or another" as one of their "long category list" (the page shows a few; the API returns more sometimes), I suspect that filtering everything with that carries "manga" may be counterproductive. Not sure if that was the case here or not but... just an observation. Annie 19:08, 11 December 2020 (EST)
The vast majority of Fixer's recent data comes from Amazon and Amazon is pretty bad at distinguishing between manga and light novels. They have gotten better at it over the last few years, but they still make mistakes, especially when the same series has both a manga version and a light novel version. I think it's best to leave this determination to humans, at least for now. I try to double check all Fixer-identified light novels and to manually reject manga at the submission creation time, but occasionally things slip through the cracks :-( Ahasuerus
Yeah, I agree. I was thinking more of some sort of flag or something so that it's easier to spot ones that might be manga. Maybe a submission note that says "This was in a manga category" or something. And yes, AMazon is often wrong with categories. :) ···日本穣 · 投稿 · Talk to Nihonjoe 12:07, 14 December 2020 (EST)

Numeric External Identifier Ranges Cleanup Report

I'm seeing a bit of oddness with this report. The report for Reginald1 has stopped showing numbers above 14330. I know that the numbers from 14334 through 15884 are all entered and have shown up previously in this report. Perhaps I had a typo in that range that is causing the report to fail. It's not urgent. Thanks. --Ron ~ RtraceTalk 07:51, 19 December 2020 (EST)

I've got a bit more information. I've entered the remaining numbers above 14330. Now the highest number displayed is 14328. Clearly there is something going on with the final grouping. --Ron ~ RtraceTalk 09:19, 19 December 2020 (EST)
Thanks, I'll take a look. Ahasuerus 12:11, 19 December 2020 (EST)
I discovered still a bit more information. Immediately prior to observing the missing final group, Dirk informed me of a Worldcat ID that I had mistakenly marked as Reginald1. This effectively had a numerically large number in the Reginald sequence making it the final group by itself. When I corrected that record the range of 14334 through 15884 became last and thus disappeared. Today, as an experiment, I added a dummy Reginald1 ID of 99999 to another record. Doing so caused the formerly hidden final group, 14174 - 15884, to appear. The dummy number which is the new final group does not show. I'll leave the dummy number there while I'm working with Reginald1 but remove when I'm done for the night. I did one more experiment. Reginald3 showed 38070 as the highest number. I added 38131 which now does not show. However, 38127, which was previously hidden, now does. I would guess that the final grouping of each of the external IDs is probably hidden. Again, this isn't urgent on my account. Especially since I found a work-around. If that dummy Reginald1 number causes you any issues when you do investigate this. Let me know and I'll remove it, or feel free to do so yourself (I'll be using this pub to hold it). Thanks. --Ron ~ RtraceTalk 18:31, 21 December 2020 (EST)
Very useful, thanks! Please feel free to keep or remove the dummy number depending on what works best for you. Ahasuerus 21:54, 21 December 2020 (EST)
And fixed! Ahasuerus 17:56, 25 December 2020 (EST)

Project Gutenberg link

Project Gutenberg updated their website and their links have changed. Instead of, it is know Thanks. -- JLaTondre (talk) 18:48, 19 December 2020 (EST)

Thanks, I'll see what I can do. Ahasuerus 21:06, 19 December 2020 (EST)
Fixed! Ahasuerus 21:15, 19 December 2020 (EST)