User talk:Ahasuerus

See User talk:Ahasuerus/Archive for discussions prior to 2023.

PLEASE NOTE: If you're writing to inform me that you've either added a COVER IMAGE or NOTES to any of my VERIFIED PUBS, please follow THIS LINK and add it to the bottom of the list. A link to the pub record would be appreciated. Once the pub has been reviewed, I'll remove your note from the list. Thanks!

Cornish

Hi, and Happy New Year! Sorry to add to your plate, but it looks like we need Cornish for Myghal Palmer and Rebellyans. In the obituary, his brother-in-law says this was published in Cornish and that at the time (2005) Palmer was responsible for 4 of the 6 (sic) books published in revived Cornish.... --MartyD (talk) 14:02, 3 January 2023 (EST)

No worries! Cornish is an ISO 639-2-recognized language, so it doesn't cause any issues. I added it a couple of minutes ago. Happy New Year! Ahasuerus (talk) 14:23, 3 January 2023 (EST)

Such speedy service! TYVM! --MartyD (talk) 20:51, 4 January 2023 (EST)

AddPub Approval page issue

When the title of the new publication does not match the title of the record you are adding into, there is no warning anymore. See here. It cannot happen outside of an API call (as the field is greyed out) but that makes processing Fixer's AddPub submissions extremely finicky (and is potentially going to lead to more missed mismatches if anyone else uses the API to submit these and the handling moderator relies on the screens to tell them of a mismatch). Annie (talk) 14:28, 27 January 2023 (EST)

PS: I think that the problem is that you never show the Publication title now on the screen - so there is no place for the warning to be shown. Annie (talk) 14:30, 27 January 2023 (EST)

I see. Thanks for reporting the problem. Let me finish testing the patch that I put together this morning and then I'll work on fixing the bug. Ahasuerus (talk) 15:04, 27 January 2023 (EST)

OK, I think I have it fixed on the development server, but it's 6:30pm on the East Coast and I don't trust my brain to test things thoroughly at this hour. I'll do the testing tomorrow morning. Ahasuerus (talk) 18:38, 27 January 2023 (EST)

No worries - I know to watch out for it and I do not think we have anyone active now that uses the API so we should be fine for a few days. Annie (talk) 19:11, 27 January 2023 (EST)

OK, I think I fixed everything. Please let me know if anything looks wrong. Thanks. Ahasuerus (talk) 13:31, 28 January 2023 (EST)

Looks good. Will take a second to get used to it being down there but it makes more sense with how the submission page is structured anyway :) Annie (talk) 17:59, 30 January 2023 (EST)

I figured that having a complete Title record in the Title table and a complete Publication record in the Publication table made more sense. The reviewer no longer needs to guess which Title field values are re-used by the Publication table.

I also realized that, for manually created AddPubs, it's possible for the pub's title/authors to get out of sync with the associated title's title/authors. All it takes is for an Edit Title submission to be approved before the AddPub submission is processed. Ahasuerus (talk) 18:07, 30 January 2023 (EST)

Mismatched Quotes Weirdness

There is a mismatched quotes warning for the synopsis field in this submitted edit. The warning appears in the proposed value column with the warning cell completely missing. I realize that this is because the missing closing quote makes the browser think that the anchor tag is not finished. I don't know if we want to consider escaping values here, or if this is a rare enough occurrence that we shouldn't worry about a cosmetic issue. In any case, I'll leave it un-approved so you can take a look. Thanks. --Ron ~ Rtrace^Talk 11:05, 16 February 2023 (EST)

I see. I think this issue is limited to Notes and Synopsis fields because they are the only ones where we allow embedded HTML (although I see that Help:Using Templates and HTML in Note Fields doesn't mention Synopsis.) I suppose the ideal solution would be to tell the browser that it should stop parsing HTML at the end of the current table cell, but I am afraid I don't know of a way to do it. Ahasuerus (talk) 12:04, 16 February 2023 (EST)

Yellow warning request

I think we need one more yellow warning on the approval screen: when AddPub had been already submitted and someone changes the type of the original we are adding to, you don't get any indication of it. Example here. I swapped the type from novel to chapbook pre-approval so I can make it in one step so I knew to check but if the approver does not open the resulting new publication, there is nothing to tell them that the work just approved needs work (And with the type missing from the top, they cannot even spot it on the screen). Annie (talk) 11:48, 23 February 2023 (EST)

PS: I think we may end up in the same situation with a ClonePub as well - I just don't have one handy as an example. Thanks! Annie (talk) 11:56, 23 February 2023 (EST)

Excellent points. FR 1561, "Enhance AddPub/ClonePub post-submission pages". Ahasuerus (talk) 17:54, 23 February 2023 (EST)

All done. Ahasuerus (talk) 23:41, 7 March 2023 (EST)

Awesome! Can you look at this one. See the top of the screen that says The Scourge Between Stars • (2023). Historically, and everywhere on the server, the year is shown that way anywhere ONLY if the title date is different from the one of the publication you are working on - so different from 2023-04-04. So seeing a year in bracket, always sent me doing sanity check on dates to see if one may need adjusting. But they match here. Is that intentional? Annie (talk) 11:20, 9 March 2023 (EST)

It looks like it was an oversight in SVN patch 1108, which implemented FR 1227, "Display the full title line for Add/ClonePub submissions", on 2023-03-05. I am busy today, but I hope to get it fixed tomorrow. Thanks for identifying the issue! Ahasuerus (talk) 14:30, 9 March 2023 (EST)

No rush and if it ends up being too complicated, I can get used to it behaving this way going forward. It looked like a quick way to see that the ebook date is not the same as the paper we are adding it to when I do the ebooks for Forthcoming books - that made me recheck the date on the paper one as well a few times today. I can just start comparing the dates on my own instead and ignore the year up there - both dates are on the screen now after all. :) Annie (talk) 18:23, 9 March 2023 (EST)

It should be fixed now. Ahasuerus (talk) 17:03, 11 March 2023 (EST)

Author birthplace question

The editor originally entered the birthplace for M. L. Weems as 'Anne Arundel County, Province of Maryland, Thirteen Colonies, British America'; triggering an exception for noncompliance with this template. I tried substituting 'England, Kingdom of Great Britain' for 'Thirteen Colonies, British America', but the software still see it as an exception. Is 'Province of Maryland' causing the software to look for USA? What is the correct fix? I'm stumped. John Scifibones 16:55, 25 February 2023 (EST)

Checking the code, I see that it expects the names of US states to be followed by ", USA" with a few exceptions: California allows "Baja California", Hawaii allows "Kingdom of Hawaii" and "Republic of Hawaii", Montana allows "Bulgaria", etc. Some of the original 13 colonies allow ", British Empire", but Maryland is not one of them. I'll need to update the code to support ", British Empire" for Maryland and the rest of the colonies. Thanks for identifying the problem! Ahasuerus (talk) 17:12, 25 February 2023 (EST)

Bug 828 has been created. We'll probably need to update the Help template as well. Ahasuerus (talk) 17:14, 25 February 2023 (EST)

Should I go ahead and update the author in anticipation of the change? John Scifibones 17:42, 25 February 2023 (EST)

Please do. It may take me a couple of days to get to it since there is a lot going on right now. Ahasuerus (talk) 17:45, 25 February 2023 (EST)

I have installed a patch to address the issue. All 13 colonies should be handled properly now. The counts will be updated when the cleanup reports run at 1am EST. Sorry about the delay. Ahasuerus (talk) 19:40, 3 March 2023 (EST)

Self-Approver

Hello, can you please comment on https://isfdb.org/wiki/index.php/ISFDB:Community_Portal#Self-Moderation_Request? --Stoecker (talk) 06:44, 11 March 2023 (EST)

Done. Ahasuerus (talk) 10:03, 11 March 2023 (EST)

Offutt EILSB

https://www.isfdb.org/cgi-bin/view_submission.cgi?5606444; Do you think the day of publication or the ISBN should be entered? --Username (talk) 08:49, 13 March 2023 (EDT)

Help:Screen:NewPub#Date says that:

• The base date optionally may be made more precise (e.g., supplying the month or day of publication) using information from a secondary source, if that source's date is otherwise consistent with publication's stated date. The source, and which details of the date were obtained from that source, must be recorded in the publication notes.

so we could make the date more precise and state the source in Notes.

Re: ISBN, Help:Screen:NewPub#ISBN allows padding a 9-digit SBN with a zero to make it a valid 10-digit ISBN, but creating a 10-digit ISBN based on a much shorter catalog ID ("64-490") seems like a stretch. We could always raise the question on the Rules and Standards page. For now I have approved the submission. Ahasuerus (talk) 13:49, 13 March 2023 (EDT)

Problem with an approved edit

When you have a few minutes, would you mind taking a look at the thread Neal Stephenson / Snow Crash? Thanks John Scifibones 19:31, 18 March 2023 (EDT)

That's odd. I will restore the full backups (which include submission history) on the development server tomorrow morning and see what I can find. Thanks for reporting the issue. Ahasuerus (talk) 23:55, 18 March 2023 (EDT)

It turns out that this and 4 other publication records were corrupted in 2007-2008. At the time we had major issues with server stability and our software was still going through a "growing pains" phase. I am going to create a database patch and apply it to the live server in a day or two. Until then these 5 publication records' Notes fields will remain uneditable. Ahasuerus (talk) 17:06, 19 March 2023 (EDT)

Thank you for taking care of this, I know you're busy. John Scifibones 17:11, 19 March 2023 (EDT)

Thanks for investigating. I'll wait until you give the all clear and then re-submit the edit. Teallach (talk) 16:56, 20 March 2023 (EDT)

Glad to be able to help. I have fixed the data and left a note on your Talk page. Ahasuerus (talk) 16:15, 21 March 2023 (EDT)

Yes, the pub record looks fine now. Thanks for the fix. Teallach (talk) 17:36, 21 March 2023 (EDT)

Possible minor bug on title-pub date check when adding collections

Hi, I just added the UK ebook of Kelly Link's White Cat, Black Dog collection, which seems to have preceded the US pubs by ~3 weeks. I was expecting it to warn me about the pub date being before the title date, but didn't get any yellow warning, which can be seen here. It did spot that 2 of the stories which inherited the US title/pub date are after that pub date though.

My guess is this might be down to one of two things:

1. It's a collection, whereas normally when I get the warning it's on a novel - do those title types perhaps use different code? 2. The check/warning wasn't done here because I cloned the pub, rather than doing an AddPub as I usually do. Again, maybe different checks are being done?

I won't update the title date for the collection, or the 2 stories, until you've had chance to look into this - I think if I fixed things, it would cause the yellow warnings on the linked submission page to disappear? Regards ErsatzCulture (talk) 09:51, 14 April 2023 (EDT)

Thanks for reporting the problem! Let me try to replicate it on the development server and see what I find... Ahasuerus (talk) 14:10, 14 April 2023 (EDT)

The problem has been identified and corrected. ClonePub was missing some code that AddPub had. Thanks again for reporting the issue! Ahasuerus (talk) 15:07, 14 April 2023 (EDT)

External IDs and Submission Conflicts

Figured I'd mention this, in case you have some good and simple idea for it. Yesterday I processed a lingering submission that had a host of changes, one of which was adding an LCCN external ID reference. There were no external IDs at the time of submission. After that was submitted, someone with self-approval rights (happened to be a moderator) edited the same pub and added OCLC/Worldcat and Reginald-3 external ID references and accepted that submission.

No surprise, when I went to process the lingering submission, that considered the external ID situation to be add LCCN and remove both OCLC/Worldcat and Reginald-3. I know the same would have been true had there been a conflicting edit to any other field. It got me to thinking, though, that if we treated external ID removal as some sort of negative addition (a la the use of 0 for removing a variant parent ID) rather than absence of the identifier, then these two additions to the External IDs would not have conflicted.

Thanks for listening. --MartyD (talk) 13:15, 20 April 2023 (EDT)

I am afraid I am not sure what "negative addition" would mean in this case. The current logic simply deletes all old External IDs, then adds the ones that were submitted. How would the proposed logic work?

To use the scenario described above as an example, if a publication has an OCLC ID and a Reginald-3 ID and a new submission has an LCCN ID, how would the software know whether the submitting editor meant to add the LCCN to the two IDs already on file or to replace them with the LCCN ID? Ditto for "Web pages" and other multi-fields.

I suppose we could create new options to "Remove IDs/Web pages/etc", which would be similar to the "Remove Titles from This Pub" option. Once it was in place, we would change the behavior of EditPub to add new values without deleting old values. Is this in line with what you were thinking? Ahasuerus (talk) 14:22, 20 April 2023 (EDT)

Yes, I was thinking that "blanking" a value would actually submit some sort of change-this-to-a-special-value that would indicate it should be removed (a la parent ID = 0), rather than simply causing it to be absent. 0 or -1 or DEL might be a good candidate. The back-end processing would have to be more complicated, as it couldn't just delete and add. But it could, for example, delete all of the IDs in the submission (rather than all of the IDs on the pub) and then add the ones in the submission back, except for the special-value ones. You could probably structure the insert's SQL cleverly to not have to filter in the code (e.g., insert into xxx values (...) where [new id value] <> 'SPECIAL-DELETE-THIS-VALUE';). --MartyD (talk) 09:56, 21 April 2023 (EDT)

The reason we can use "0" as a special value is that our Title IDs cannot be 0. On the other hand, we have no way of telling what strange value a third party may decide to use. We wouldn't want to find ourselves in the same position as a certain government agency which used "Test" as a dummy last name (for testing purposes) until they discovered that people with that last name actually existed :-)

Luckily, we don't need to use special values. We can use a new XML element name for a list of "External ID Type-External ID value" pairs to be removed.

That said, I am not sure that it will cover all possible scenarios of External IDs getting added, removed and replaced. Let me draw a few diagrams and see what scenarios I can come up with... Ahasuerus (talk) 12:54, 21 April 2023 (EDT)

Thanks for thinking about it. It's obviously NOT hugely important. --MartyD (talk) 11:16, 25 April 2023 (EDT)

ISBN Search Case Sensitive

ISBN search is now case sensitive which is a bit of a pain.

• https://www.isfdb.org/cgi-bin/se.cgi?arg=162755646x&type=ISBN returns nothing
• https://www.isfdb.org/cgi-bin/se.cgi?arg=162755646X&type=ISBN returns the pub

Can we get it back to being case insensitive? Probably would just need to convert search term to uppercase when ISBN selected. -- JLaTondre (talk) 20:24, 24 April 2023 (EDT)

That's odd. I'll take a look. Thanks for reporting the issue. Ahasuerus (talk) 21:02, 24 April 2023 (EDT)

And fixed. Ahasuerus (talk) 17:52, 25 April 2023 (EDT)

Thank you. -- JLaTondre (talk) 18:57, 25 April 2023 (EDT)

Merge variant with parent problem

See this submission, but Keep ID and Drop ID are both present. The Drop ID is a variant of the Keep ID, but in the XML it's trying to set the merged record's Parent to Drop ID, which is rather odd. And I also think maybe the message means to say that the proposed parent does not exist? --MartyD (talk) 13:46, 6 May 2023 (EDT)

Thanks, I'll take a look. Ahasuerus (talk) 13:52, 6 May 2023 (EDT)

Earlier this morning I recreated the problem on the development server. I then created FR 1565, "Disallow Merge Title submissions which would create circular VTs". The FR reads:

Currently, the software lets you create a MergeTitle submission which would merge a variant with its parent and make the resulting title record a variant of itself. Since circular variants are not supported by the ISFDB software, these types of submissions are invalid.

When it happens, the submission review software makes the submission unapprovable and displays the following error message:

"The proposed parent title is the title record which will be kept after the merge."

We want to change the software to catch this problem at submission creation time. The error message listed above should be retained in case a title record is turned into a variant between the time the submission is created and the time it's reviewed/approved. However, the text should me made more explicit:

"This submission would result in a title that is a parent of itself, which is not allowed."

Thanks for identifying the issue! Ahasuerus (talk) 12:46, 7 May 2023 (EDT)

Thanks, and thanks for looking at it. --MartyD (talk) 16:24, 7 May 2023 (EDT)

Done -- see the Community Portal announcement. Sorry about the delay: there was another self-variant scenario which I found during testing and had to address. I also rewrote the whole script to make it more maintainable. Ahasuerus (talk) 16:40, 18 May 2023 (EDT)

Excellent, thanks. And never any need to apologize for how long it takes to get to something. --MartyD (talk) 20:54, 20 May 2023 (EDT)

Suboptimal behaviour on art award categories for "proper" titles

This isn't a bug IMHO, but it's definitely not ideal - if you're adding an award entry such as this one for the Locus Poll Art/Illustrated category, then only the author gets shown, when it's probably the case that the artist(s) is at least as responsible for the work being nominated. However, if the nominated work was never entered as a title/pub, then the artist does show up along with the author in the untitled award record (example).

(I'm guessing a similar situation would be in play for translated works, and for audiobook awards if we record any.)

I can't think of any easy solution for this - obviously the ideal would be if/whenever proper "roles" for a work are addable to the database - but I'm just mentioning it, in case there's anything you can think of as an interim workaround. Maybe something could be done with the award note field, but it'd still be a bit clunky?

Apologies in advance if this is something that is already known about and has been considered... ErsatzCulture (talk) 16:40, 13 May 2023 (EDT)

When we create a "title-based" award record, we link it to one -- and only one -- title record. As long as this limitation remains in place, we can't link an award record to multiple title records at the same time.

That being said, I am not sure we have this title entered correctly. If you look at this publication record and pull up its Amazon UK page, you will note that it is described as follows:

• From the Eisner and Bram Stoker-award winning team of Snow, Glass, Apples comes a delightfully humorous and charming new graphic novel adaptation.

This suggests that this is not just a reprint of Gaiman's 1992 story "Chivalry" illustrated by Coleen Doran, but a new (derivative) graphic novel by Gaiman and Doran. Once we create a new SHORTFICTION title record for the graphic novel and enter Gaiman and Doran as its authors, we should be able to link the award record to it. Does this make sense? Ahasuerus (talk) 20:16, 13 May 2023 (EDT)

Thanks - I've pinged the PVing editor who submitted that title & pubs. ErsatzCulture (talk) 08:16, 14 May 2023 (EDT)

As the PV and editor that created all of these pubs, I may indeed have attributed the authorship wrong. The copy I have in-hand shows the following on the title page: "Story and Words" [over] "Neil Gaiman" plus "Adaptation, Art, and Illuminated Manuscript Lettering" [over] "Colleen Doran" plus "Lettering" [over] "Todd Klein". Relooking at this, I can see why the author's should be both Gaiman and Doran. If I just change the title record, will that flow down to the pub records as well? Phil (talk) 08:39, 14 May 2023 (EDT)

I am afraid not. I think the quickest way to address this issue would be to do the following:

• Remove the 1992 SHORTFICTION title from each of the 4 affected CHAPBOOK pubs
• Change the authorship of the CHAPBOOK title record to "Neil Gaiman" and "Colleen Doran"
• Change the authorship of the 4 CHAPBOOK publications which include this title record to "Neil Gaiman" and "Colleen Doran"
• Edit the "Chivalry" INTERIORART title by Coleen Doran to be a SHORTFICTION [sic!] title by Coleen Doran and Neil Gaiman; make sure to set the "graphic format" flag

I think that should do it, but we can review the records after the changes are made. You may also want to add a link to this discussion to the Moderator Note field of your submissions. Thanks for chiming in! Ahasuerus (talk) 11:32, 14 May 2023 (EDT)

Those changes are submitted except for setting the graphic format flag which has to be done after the type is changed to SHORTFICTION. In this process, I also discovered I had consistently misspelled Colleen (including in the other titles and cover artist) so there are also submissions to correct that. Sigh. Phil (talk) 15:06, 14 May 2023 (EDT)

Everything has been approved and the "graphic format" flag has been set. I believe the only thing that we still need to decide on is whether the previously discussed award record, which is currently linked to the CHAPBOOK title, should be linked to the SHORTFICTION title instead. Ahasuerus (talk) 17:31, 14 May 2023 (EDT)

The prior award record has been deleted and a new one created against the new shortfiction record. Thanks both. ErsatzCulture (talk) 17:55, 14 May 2023 (EDT)

Excellent! :-) Ahasuerus (talk) 18:32, 14 May 2023 (EDT)

There is no need to delete an award that is pointing to the wrong record. While on the 'Award Details' screen, select 'Link Award' from the 'Editing Tools' menu and enter the correct record #. This will preserve any useful information. This one for example. I won't need to repeat the research. John Scifibones 20:12, 14 May 2023 (EDT)

Odd & now missing approval

Hi Ahasuerus. I just approved this submission[1] and got an error message, but when I checked it went through, except it's not in the recently approved or in author history. Thought I should let you know.Kraang (talk) 10:52, 15 May 2023 (EDT)

I see that the submission errored out -- see this list of Errored Out and "In Progress" Submissions, which can be accessed from the "Moderator Links" section of the navigation bar on the left. "Errored out" submissions are typically due to server problems that we have no control over. Some months the server that we rent experiences more of them than other months. For example, it happened 19 times in December 2022, but only 16 times in January-May 2023. Ahasuerus (talk) 11:31, 15 May 2023 (EDT)

Invalid Page Numbers

Please see this pub and the first page number (`42). Is non-standard punctuation something that should be added to the cleanup report? -- JLaTondre (talk) 12:32, 21 May 2023 (EDT)

It's 142, I fixed it in my (pending) edit. Whoever entered contents messed up several page numbers. --Username (talk) 12:43, 21 May 2023 (EDT)

(edit conflict) From the last database dump:

mysql> select pubc_page,pub_id from pub_content where pubc_page like "%`%";
+-----------+--------+
| pubc_page | pub_id |
+-----------+--------+
| `42       | 303295 |
| 1`        | 389098 |
| |9`       | 472957 |
| `51|28    | 249793 |
| 80`       | 919331 |
| 6`        | 927024 |
+-----------+--------+
6 rows in set (6.89 sec)

So seems a common enough occurrence of people missing the 1 key and hitting ` instead. -- JLaTondre (talk) 12:45, 21 May 2023 (EDT)

Thanks for identifying the issue. I will update the cleanup report once I finish the current task. Ahasuerus (talk) 16:42, 21 May 2023 (EDT)

I have updated the cleanup report to include values with a "backtick" (`) character. The new data will become available tomorrow morning. Unfortunately, I couldn't think of any other punctuation characters that could be included because we record page numbers as they appear in publications, including oddball values like "C-1", "C!1", etc. Ahasuerus (talk) 10:46, 23 May 2023 (EDT)

Figuring out a Russian name

This story is by a Russian from the Soviet era (1954 or before). Any chance you can figure out who it is? ···日本穣 · 投稿 · Talk to Nihonjoe 19:00, 26 May 2023 (EDT)

[2]; right-click, translate to English, if that helps. --Username (talk) 19:22, 26 May 2023 (EDT)

As far as I can tell, there's no one with the name Арефьев in the database. I couldn't find any connection other than that one page. ···日本穣 · 投稿 · Talk to Nihonjoe 19:35, 26 May 2023 (EDT)

Here it is: Тайна полигона by С. Арефьев (I went brute force based on the Japanese name -- there are only that many possible spellings in Russian of that last name. I will update the record. :) Annie (talk) 19:33, 26 May 2023 (EDT)

Cool. Thanks (to you and Username). ···日本穣 · 投稿 · Talk to Nihonjoe 19:37, 26 May 2023 (EDT)

And now we have a more-or-less complete entry for the first Japanese science fiction magazine! ···日本穣 · 投稿 · Talk to Nihonjoe 19:39, 26 May 2023 (EDT)

And I just tracked down the original Russian publication of that story as well :) Annie (talk) 19:43, 26 May 2023 (EDT)

Awesome! That's a pretty quick turnaround on translating a story and publishing the translation (end of August to December in the same year). ···日本穣 · 投稿 · Talk to Nihonjoe 19:47, 26 May 2023 (EDT)

Yep - it is possible that the translation was done based on pre-publication proofs. I am seeing a few reprints in Russian magazines and newspapers in September 1954: serialization in very early September in За тяжёлое машиностроение (based in Yekaterinburg) and a full story in Молодой сталинец (in Tbilisi) in the middle of the month so I suspect the story was making the rounds before it officially came out. I cannot find anything before that August publication (and that magazine makes sense as a first publication) but now I am curious just how many I will find (I see at least one later serialization in 1957-1958). Maybe there is an earlier one out there in the same year - Fantlab puts in in 1954 (with no details on where) so it is unlikely to find it earlier than that I'd think. :) Always fun to track down old stories :) Annie (talk) 20:17, 26 May 2023 (EDT)

(after edit conflict) Quick turnaround indeed. I am not an expert in the area of Soviet-Japanese relations, but the first thing that comes to mind is that the Soviets were trying to improve their relationship with Japan around the same time. In July their First Deputy Foreign Minster Andrei Vyshinski claimed that the USSR wanted to promote trade and cultural exchanges with Japan, ultimately leading to a "normalization" of the relationship. In September Vyacheslav Molotov, the Soviet Foreign Minister, said that the USSR "expresses its readiness to make normal its relations with Japan". It culminated in a Sino-Soviet "Joint Communique" in October in which both countries "express[ed] their readiness to take steps to normalize their relations with Japan". Perhaps the editor(s) thought that it made late 1954 a perfect time to publish a Russian story. Ahasuerus (talk) 20:26, 26 May 2023 (EDT)

Weird Duplicate Submissions

This was probably just some weird hiccup, but I thought I would mention it: I made an edit to a publication after accepting the submission. From my side, I got a single moderator approval window. But when I went into the moderation queue after approving that, I saw about 30 duplicate submissions (see Recent Rejections from 2023-05-28 08:39:42 to 2023-05-28 08:39:44). As far as I know, I didn't do anything different than normal when submitting (ex. holding down a key while submitting). I certainly didn't inadvertently pop-up 30 different tabs or anything like that. Thanks. -- JLaTondre (talk) 08:56, 28 May 2023 (EDT)

Thanks for letting me know. It is exceedingly odd. On rare occasion, something happens between Apache and MySQL that results in a duplicate submission, but I don't think I have ever seen 30 duplicates.

Checking Apache's log file, I see:

• [Sun May 28 08:39:42.498256 2023] [cgid:error] [pid 1597558:tid 1597806] (104)Connection reset by peer: AH01270: Error reading request entity data, referer: https://www.isfdb.org/cgi-bin/edit/editpub.cgi?952950

which matches the time frame and the publication ID. So something clearly happened at the Apache level, but I am not sure what it was. There are online reports of a regression bug in this area under Apache 2.4.39, which was fixed in 2.4.40, but we are running a later version. I guess we'll have to keep an eye on it. Thanks again. Ahasuerus (talk) 12:41, 28 May 2023 (EDT)

I just noticed I got 3 dupe NewAwards when I was adding this year's Clarke Finalists yesterday; nothing as excessive as described above, but mentioning in case it's of any use. 5685820 and 5685819 are the same as 5685821, 5685816 is the same as 56885817. The submit times on the first three and the second pair are the same, the edits which were accepted have approval times a couple of seconds later. I don't recall noticing anything atypical at the time. This was on latest Firefox on Linux.

No expectation on you doing any investigation on these, just providing some datapoints if you need them. One thought though: if this keeps happening and no other solution could be found, could you maybe have some sort of hidden field in the edit forms with an UUID generated by JavaScript when the user hits the submit button, which might allow dupes to be detected server side? ErsatzCulture (talk) 17:05, 8 June 2023 (EDT)

Please come share your thoughts

See here. Thanks! ···日本穣 · 投稿 · Talk to Nihonjoe 19:37, 1 June 2023 (EDT)

Done. Ahasuerus (talk) 21:54, 1 June 2023 (EDT)

Unreject link

It looks like the unreject link doesn't show up in the moderator view (see here), but does show up in this view. Can the unreject link be added to the moderator view page? ···日本穣 · 投稿 · Talk to Nihonjoe 19:42, 1 June 2023 (EDT)

Good catch. FR 1567 has been created. Ahasuerus (talk) 21:55, 1 June 2023 (EDT)

Done. Ahasuerus (talk) 16:15, 2 June 2023 (EDT)

Thank you! Now I won't be confused trying to figure out where it is. (^_^;; ···日本穣 · 投稿 · Talk to Nihonjoe 17:56, 2 June 2023 (EDT)

I live to serve! Ahasuerus (talk) 20:33, 2 June 2023 (EDT)

Submission language

Discussion moved to ISFDB:Moderator_noticeboard. Ahasuerus (talk) 07:57, 3 June 2023 (EDT)

Minor bug: editing an existing pub's date does not have the yellow same warnings re. title date as when adding a new pub

If you add a new pub that has an earlier date than the title date - which I've had a lot of on stuff like Star Trek ebooks - you get a yellow warning about being earlier than the title date, which reminds you to update the title record.

However, I've just noticed that if you edit a pub to bring the date forward, you don't get the same reminder. Last night I had to fix a couple of pubs for various things including earlier UK pub date, and it was only today when I went to add one of the corresponding ebooks that I realized I never got the title date warning when editing the existing pubs. I've just done some tests to confirm that's the case. Not a big deal in an of itself, but may be a fairly simple copypaste to add the same check when editing an existing pub?

(I linked the message(s) I've left on Annie's page, as this is a US pseudo-publisher of what are actually UK pubs, which seem to be prone to multiple incorrect fields in the Amazon data that Fixer pushes through; I dunno if there's anything that could be done to warn the approver to be extra careful with submissions for that publisher?) ErsatzCulture (talk) 16:33, 5 June 2023 (EDT)

Thanks, I'll check the date warnings.

Re: Mobius, I have tweaked Fixer to let the reviewing moderator know that "Mobius" pubs are suspected UK imports. Ahasuerus (talk) 18:30, 5 June 2023 (EDT)

I have run a few tests on the development server and here is what I see. If you change one of the title dates in the Content section, the software compares the new date with the publication date. If the former is later than the latter, a yellow warning is displayed. However, if you change the publication date value in the Metadata section, the software doesn't compare the new vale with the dates of the titles associated with the pub. FR 1569 has been created. Thanks for identifying the issue! Ahasuerus (talk) 15:18, 6 June 2023 (EDT)

Image linking permissions

This author would like to add his website to the list. Would you mind outlining the procedure here. Thanks, John Scifibones 16:29, 7 June 2023 (EDT)

Done. Thanks for the heads up! Ahasuerus (talk) 19:36, 7 June 2023 (EDT)

Software and Help updated. Ahasuerus (talk) 15:31, 9 June 2023 (EDT)

I appreciate it. John Scifibones 17:06, 9 June 2023 (EDT)

Fixer and Clay Harmon's The Flames of Mira

See User_talk:Anniemod#Alice_James_-_Grave_Danger_.2F_Clay_Harmon_-_Flames_of_Mira for a bit more background/context on this.

This is a Solaris/Rebellion title that came out in hc, ebook and audio last summer, and has a tp due at the start of July. None of those pubs came from Fixer submissions, which I thought was a bit odd, given that (IMHO) Solaris/Rebellion is a reasonably high profile publisher, albeit not Big 5.

I download the Fixer files intermittently, and I can see that the ISBNs and ASINs were known. The hc ISBN (9781786185419) was priority 1, but I submitted it first, 2-3 weeks ahead of publication. The ebook ASIN (B0B1QKFFG9) was priority 1 for several months, but I added it earlier this year, several months post publication. The tp (9781786189615) has been in 'n' state since at least February; I have scraped data for it from Waterstones from November last year, and it hasn't changed pub date in that time. Naively I'd have assumed that it would have been given priority 1 as with the hc?

Obviously I'm more than happy to submit these pubs (*), but I'm just mentioning it in case Fixer is doing something funny with them. Here are relevant bits from the Fixer files I have; the dates are split into 2 groups due to me getting a new dev box recently:

   $ grep -P "(B0B1QKFFG9|B09TQ3ZCT7|B09TQ2K1HG)" */ASIN*txt
   _fixer_20220611_/ASINs2022-06-11.txt:B09TQ3ZCT7||n
   _fixer_20220611_/ASINs2022-06-11.txt:B0B1QKFFG9||n
   _fixer_20220820_/ASINs2022-08-20.txt:B09TQ3ZCT7||n
   _fixer_20220820_/ASINs2022-08-20.txt:B0B1QKFFG9||1
   _fixer_20230204_/ASINs2023-02-04.txt:B09TQ3ZCT7||n
   _fixer_20230204_/ASINs2023-02-04.txt:B0B1QKFFG9||1
   $ grep -P "(9781786189615|9781786185419|9781786185426|9781786185402)" */*ISBN*txt
   _fixer_20220611_/ISBNs2022-06-11.txt:1786185415|9781786185419|1|
   _fixer_20220820_/ISBNs2022-08-20.txt:1786185415|9781786185419|8|
   _fixer_20230204_/ISBNs2023-02-04.txt:1786185415|9781786185419|8| 
   _fixer_20230204_/ISBNs2023-02-04.txt:1786189615|9781786189615|n|
   $ grep -P "(9781786189615|9781786185419|9781786185426|9781786185402)" *ISBN*txt
   ISBNs2023-04-29.txt:1786185415|9781786185419|8|
   ISBNs2023-04-29.txt:1786189615|9781786189615|n|
   ISBNs2023-06-03.txt:1786185415|9781786185419|8|
   ISBNs2023-06-03.txt:1786189615|9781786189615|n|
   $ grep -P "(B0B1QKFFG9|B09TQ3ZCT7|B09TQ2K1HG)" *ASIN*txt
   ASINs2023-04-29.txt:B09TQ3ZCT7||n
   ASINs2023-04-29.txt:B0B1QKFFG9||8
   ASINs2023-06-03.txt:B09TQ3ZCT7||n
   ASINs2023-06-03.txt:B0B1QKFFG9||8

(* With the proviso that the Solaris/Rebellion website is pretty horrible - at least for our purposes - so I've never bothered writing scrapers for it, and rarely look at it. As such, I'm less likely to spot anything that Fixer hasn't caught, than for pubs from the other main UK publishers) ErsatzCulture (talk) 09:00, 8 June 2023 (EDT)

Checking this ISBN in Fixer's "data stores", I see that this is a known issue with the way Fixer processes UK publications. It's caused by the fact we don't have access to the Amazon API on the UK side. When Fixer finds a new ISBN/ASIN at Amazon UK, we query the US API since it's the only option that we have. If the ISBN/ASIN is unknown to the US API as of the time when it appears on the UK side, Fixer records the date when the US API was queried. The date is later checked to prevent Fixer from querying the US API over and over again, which is good if the ISBN/ASIN is never made available on the US side, but bad if it pops up a few weeks (or months) later.

In the past I would have said that this was the best that we could do given our limited access to the Amazon API. However, there is an additional twist. A couple of years ago Fixer began capturing UK- and US-specific lists of Amazon "browse nodes" for each published (as opposed to announced) ISBN/ASIN. Importantly, this list of browse nodes is built separately and doesn't rely on the Amazon API. In the case of ISBN 9781786189615 it means that Fixer knows that it is associated with browse nodes 9803, 16190, 16205 and 10159265011 on the US side.

What this means is that I can create a script to examine Fixer's data stores and create a list of ISBNs/ASINs which the Amazon API was not aware of at some point in the past, yet they have at least one associated US browse node on file. I can then send a list of these ISBNs/ASINs to the Amazon API and get their data.

I am currently sick and can't really code, but I'll add this issue to my list of things to do. Thanks for reporting the problem! Ahasuerus (talk) 12:04, 8 June 2023 (EDT)

Thanks for responding. First of all, please don't feel any obligation to respond to this comment, or this item in general.

A question which can hopefully be easily answered with just a yes or no answer: Is the content of the XML files that are in the Fixer dump zip based solely on data obtained from the Amazon US API? I've never examined those XML files before, but I see they have a load of interesting data that's not in the .txt files. However, when I tried grepping for "Flames of Mira" or the tp ISBN 9781786189615, I couldn't find either of them, which makes me suspect they are from your website scraping?

(The angle I'm coming at is that if I had a list of UK ASINs and/or ISBNs that Fixer was aware of, but wasn't able to progress due to lack of info from the API, then I could plug them into my workflows.)

Thanks ErsatzCulture (talk) 17:51, 8 June 2023 (EDT)

"Fixer_dump_*.xml" contain API-provided data about outstanding ISBNs/ASINs. They do not include ISBNs/ASINs without API-provided data. "ASINs*" and "ISBNs*" list all known ISBNs/ASINs, including rejected (priority 9) and submitted (priority 8) ISBNs/ASINs as well as all ISBNs/ASINs which were not recognized by the Amazon API. I'll try to clarify User:Fixer#Fixer_Dumps_and_Lists when I am feeling better. Ahasuerus (talk) 19:43, 8 June 2023 (EDT)

FR 1571 has been created. Ahasuerus (talk) 14:22, 10 June 2023 (EDT)

Russian Science Fiction

https://www.isfdb.org/cgi-bin/view_submission.cgi?5689904; https://www.isfdb.org/cgi-bin/view_submission.cgi?5689906; The '68 cover image is now unstable; essay "note" in '68 should be "notes" and book title added in parentheses to differ it from same-titled essay in '69. Also, since title is written on covers and inside books with the year on a separate line ISFDB titles should probably have a colon before the year. --Username (talk) 13:22, 11 June 2023 (EDT)

ISBN things

Hi. While working on a submission involving a 1965 pub and an entry we had dated 1965 yet with an ISBN-10, I discovered that the SBN system was formulated in 1966 and went into use in 1967 (see here). So we couldn't have even an SNB-converted-to-ISBN on a pre-1967 publication (or pre-1966 if we want to be double super extra cautious). Such a combination must indicate a data error. Perhaps (a) worth a clean-up report and (b) a different/stronger warning than the ISBN-on-a-pre-1970-publication that comes up now.

And if that's not enough, while on the subject of submission warnings, I noticed I can change the publication date on an entry with an ISBN to something pre-1970, and there is no warning that the date conflicts with the presence of an ISBN. Don't know how often an occasion for such a warning might come up, but I figured I'd mention it. --MartyD (talk) 16:19, 11 June 2023 (EDT)