ISFDB:Moderator noticeboard/Archive 22

From ISFDB
< ISFDB:Moderator noticeboard
Revision as of 18:47, 26 September 2017 by Nihonjoe (talk | contribs) (archive July 2017)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

This is an archive page for the Moderator noticeboard. Please do not edit the contents. To start a new discussion, please click here.
This archive includes discussions from July 2017 - December 2017.

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


1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 · 9 · 10 · 11 · 12 · 13 · 14 · 15 · 16 · 17 · 18 · 19 · 20 · 21 · 22 · 23 · 24 · 25 · 26 · 27 · 28 · 29 · 30 · 31


Expanded archive listing



Sorry multiple pages in discussion

i created multiples of same query -- could you please remove the additonals? Thnaks —The preceding unsigned comment was added by Akua (talkcontribs) . on 23:10, 4 July 2017 (EDT)

Done and no worries.
PS: If you want to sign your messages, you can just add ~~~~ (that is 4 times the symbol ~ at the end of your messages. Welcome again :) Annie 23:33, 4 July 2017 (EDT)

Extending of input mask

Hello, can we (you) extend the input mask by several fields? E. g. "Translator", "Printing", "Edition", "Weight", "Dimensions". These data can I read often in the notes, but this is confusing. This would be better and more structured. Thanks. --Wolfram.winkler 03:46, 12 July 2017 (EDT)

There are plans to add more fields staring with catalog IDs and improved ISBN/price support. However, it's an involved process and will take time to implement. Ahasuerus 11:46, 12 July 2017 (EDT)
Hello Ahasuerus, please hurry up. --Wolfram.winkler 05:22, 13 July 2017 (EDT)
Alas, there are many competing priorities. I have installed 284 patches this year, but there are only so many hours in a day... Ahasuerus 09:26, 13 July 2017 (EDT)
Please do NOT hurry up. I would much rather see careful, accurate progress. Chavey 05:54, 16 July 2017 (EDT)
I agree with Chavey: Please choose accuracy over speed. ···日本穣 · 投稿 · Talk to Nihonjoe 00:03, 25 July 2017 (EDT)

ISBN-ASIN reconciliation: Part 1

Now that:

  • our software supports third party ISBNs, and
  • the vast majority of ASINs has been migrated to the new field

we can start reconciling ISBNs and ASINs.

As most of you probably know, Amazon typically doesn't display ISBNs for e-pubs. However, they are available via Amazon's back end API which Fixer has access to.

Fixer has identified 578 publications which:

  • have an ASIN external ID
  • do not have an ISBN in the publication record
  • have a matching ISBN in Fixer's database

I have tweaked Fixer's software to create 578 submissions to assign ISBNs to our ISBN-less pubs. You can see the first one here. It's currently on hold pending moderator feedback. If everything looks OK (including Note to Moderator), I will proceed with creating the other 577 submissions, probably in batches of 50 or so. Ahasuerus 20:20, 19 July 2017 (EDT)

When adding an ISBN to a pub, does the duplicate ISBN warning logic occur? -- JLaTondre (talk) 20:53, 19 July 2017 (EDT)
Yes, the submission display software will perform the normal checks, including the check for duplicate ISBNs. Ahasuerus 21:10, 19 July 2017 (EDT)
If so, the "Once approved, please check for possible duplicate publications" could be changed to state need to investigate if the warning is present. -- JLaTondre (talk) 20:53, 19 July 2017 (EDT)
Upon reflection, I should probably check for pre-existing duplicate ISBNs at submission creation time. It won't be completely bulletproof since Fixer's database is at least a few hours behind the main database, but it will make the approval process more straightforward. Also, I guess it's possible to have multiple ISBN-less publication records for the same e-book (from Amazon, B&N, etc.) It may be better to keep the moderator note sufficiently vague to account for various permutations. Ahasuerus 21:15, 19 July 2017 (EDT)
Should a pub note be added stating the ISBN source is per Amazon? In this example, the data is per Barnes & Noble. However, searching that ISBN on their site returns no results. -- JLaTondre (talk) 20:53, 19 July 2017 (EDT)
That's a good point, but I am not sure how much I can do at submission creation time. It's easy enough to create a submission to add a missing field value to a pub record. However, editing an existing field (in this case the Note field) is tricky because the value may have changed since Fixer's database was created. If that happens, it may overwrite previously entered data. Ahasuerus 21:19, 19 July 2017 (EDT)
I don't mind having to edit the note. The cases where that will be "necessary" (term used loosely) I'd think would be few. I wonder how many are not Amazon-sourced Fixer submissions in the first place, with an Amazon credit already in the notes. Maybe you could search the notes string Fixer has for Amazon and word the note to the moderator differently? --MartyD 21:25, 20 July 2017 (EDT)
An excellent idea! Ahasuerus 22:56, 20 July 2017 (EDT)
A new version of the same submission has been created for your viewing pleasure. The old one has been rejected. Ahasuerus 18:50, 23 July 2017 (EDT)
I like it. Annie 16:43, 24 July 2017 (EDT)
OK, Fixer's first 10 EditPub submissions have been queued up. It looks like we have quite a few ISBN/ASIN duplicates, so a fair amount of judicious pruning will be required. Please remember to put Fixer's submissions on hold before you begin working on them in order to avoid collisions. Ahasuerus 20:08, 24 July 2017 (EDT)
The first batch has been processed. (Hurray!) Batch #2 is on its way. Ahasuerus 23:32, 24 July 2017 (EDT)
Batch #2 has been processed. Batch #3 has been submitted. Approximately 500 ISBNs remain outstanding. Ahasuerus 10:40, 25 July 2017 (EDT)
I didn't realize that there were batches (just saw a funny new kind of fixer submission). I've processed them. Roughly one half added ISBN to a sole ISBN-less publication, the other half tried to add an ISBN to a duplicate (ISBN-less) of a publication that was already existing with (in most of the cases) exactly the same characteristics. I've rejected the submissions and deleted the extra publications. Hauck 10:59, 25 July 2017 (EDT)
Thanks for reminding me about the duplicate ISBN issue! I have modified Fixer's logic not to create EditPub submissions for ISBN-less pubs whose Fixer-identified ISBNs are currently associated with other pubs. Fixer is aware of 190 "truly new" ISBNs, so we will add them first. Once that's out of the way, we can clean up the remaining 300 pubs. Ahasuerus 11:41, 25 July 2017 (EDT)
And there went the fun - rejecting and deleting duplicates was the funniest part of that exercise. :)
More seriously now... you will need to have Fixer redo those lists a few more times - I am assigning ASINs to the ASIN-less no-ISBN publications. I am doing as much cleanup as I can while I am on it but it is possible that some duplicate pairs may escape. Annie 13:31, 25 July 2017 (EDT)
Sounds good! Fixer's database is typically refreshed on Saturday morning as part of the weekly backup process. With luck, we will have the bulk of the currently outstanding ISBNs processed over the next 3-4 days. Once I refresh Fixer's database, I will rerun the process against the fresh data and see what he finds. Ahasuerus 13:40, 25 July 2017 (EDT)

(unindent) Batch #4 has been submitted. Ahasuerus 00:12, 26 July 2017 (EDT)

Batch #5, which includes the remaining "easy" ISBNs, has been submitted. Once the data has been processed, I will start submitting batches of duplicate ISBNs which will require additional massaging. Ahasuerus 12:04, 26 July 2017 (EDT)
Fixer has identified and submitted ISBNs for a few oddball publications which can't be easily categorized. Once they have been processed, I'll start submitting "known duplicate ISBNs", which will require additional cleanup. Fixer is currently aware of 295 of them. Ahasuerus 14:57, 26 July 2017 (EDT)
These are done. Now can we have the fun ones again? :) Annie 18:05, 26 July 2017 (EDT)
Certainly! As soon as I finish refreshing Fixer's database, I'll ask him to submit a bunch. Ahasuerus 18:42, 26 July 2017 (EDT)
Batch #7 has been submitted. Enjoy! Ahasuerus 19:15, 26 July 2017 (EDT)
Put one on hold (D Is for Dinosaur) that is an error. It is trying to add the tp ISBN to the ebook. I confirmed that via Amazon Look Inside which shows a different ISBN for the ebook. Don't know if this is an Amazon screw-up or a bug. -- JLaTondre (talk) 19:54, 26 July 2017 (EDT)
Voices & Edge of Sundown are more cases. Seems something is wrong with this last run? -- JLaTondre (talk) 19:59, 26 July 2017 (EDT)
So is Altar. And then there are cases like Coexist (which is totally messed up and I am still looking at it - and finally remembering to put hold on the ones I am working on.) I deleted a couple of dozen so far that were real duplicate publications but most of the ones that are not real duplicates are weird Annie 20:01, 26 July 2017 (EDT)
Glory and Splendour is similar, but in that one's case, Amazon Look Inside shows the same copyright page for the physical & ebook (including ISBN) so assuming that one is on Amazon's side. -- JLaTondre (talk) 20:08, 26 July 2017 (EDT)
Fixer's submissions are based on what Amazon returns via its API. Occasionally the API does weird things, e.g. when Fixer sends it an ASIN, it returns both the e-book ISBN and the tp ISBN. I have added a certain amount of bulletproofing over the years, but sometimes things slip through. If you see consistent patterns, please let em know and I will try to add more bulletproofing to Fixer.
Also, it occurs to me that Fixer is using the ISBNs that Amazon's API returned back when the data was first captured. I should probably get Fixer to re-query the API in case some ISBN have changed. I am not 100% sure that the new data will be better than the original data, but it seems likely. Ahasuerus 20:47, 26 July 2017 (EDT)
So should we opt on the side of caution and reject anything that looks weird? We won't lose any existing data this way and if that ISBN comes up again, we can deal with it then and there... Thoughts? Annie 20:53, 26 July 2017 (EDT)
I guess there are a few ways to approach this issue since there are fewer than 200 outstanding ISBNs left. For example, if the number of questionable records is reasonably low, then the reviewing moderators could reject them. I would then re-run the "rejects" against the Amazon API to see if the ISBN has changed. Ahasuerus 21:41, 26 July 2017 (EDT)
If we go that way, you will need to look for rejects that are still in the DB only - I am rejecting updates that try to update a duplicate as well (followed by deleting the one Fixer tried to update after moving whatever relevant information is missing (usually none but an artist name now and then). If that is the agreed upon process, I can go and reject my holds from yesterday. Between me and JLaTondre, we seem to be holding 9 from yesterday's batch - the rest were cleared. Annie 13:17, 27 July 2017 (EDT)
Sounds like a plan. I'll go ahead and submit the 200 "new" ISBNs that Fixer is currently aware of. Ahasuerus 15:06, 27 July 2017 (EDT)

(unindent) After updating Fixer's database with the latest backup file, I examined our history of rejected Fixer-generated EditPub submissions. I then resubmitted the affected ASINs to the Amazon API to make sure that the ISBNs were still valid. Everything checked out except for one publication record, now updated.

Many thanks to the moderators who have worked on these submissions. I can now start working on Phase 2 of the project, which will add ASINs to ISBN-13-enabled e-books. Ahasuerus 18:42, 29 July 2017 (EDT)

Publishers to merge, 2017-07-22

Please merge Kristell Ink into Kristell Ink / Grimbold Books (keeping the latter). Thanks --Vasha 19:26, 22 July 2017 (EDT)

Done. --Ron ~ RtraceTalk 22:00, 22 July 2017 (EDT)

Suspect Untitled Awards

What is this report supposed to list? I ask because the three items there right now are not untitled awards as far as I can tell. ···日本穣 · 投稿 · Talk to Nihonjoe 22:13, 24 July 2017 (EDT)

The three listed awards were flagged because:
  • they were not linked to titles
  • their award categories contain the word "story", which suggests that they should probably be associated with title records
The first listed award was given to "the trilogy as a whole". We don't have an omnibus edition of the trilogy, so there is nothing to link to. If an omnibus edition exists in the wild, we can enter it into the database and then link the award to it. If not, then it will have to remain an "untitled award".
The second and the third awards were given to stories which we don't have on file yet, but we know where they appeared. Once the pubs have been entered, we can link the award records. Ahasuerus 22:48, 24 July 2017 (EDT)
Is there a current feature request to allow awards to be given to a series? This has happened in quite a number of cases (not just Japanese awards), so this might be good to have on the list. ···日本穣 · 投稿 · Talk to Nihonjoe 23:03, 24 July 2017 (EDT)
There are feature requests to allow linking awards to author records (FR 583) and to publisher records (FR 269), but not to series records. I agree that series awards are becoming more common -- even the Hugos have been changed to support series awards -- so it's something to consider. It will take some work, but it's definitely doable. It's actually harder to add support for author- and publisher-based awards because we'll need to decide how to handle record auto-deletion (which we don't do for series at this time.) Ahasuerus 23:31, 24 July 2017 (EDT)

A few extra mod eyes, please

I think OCLC has the wrong cover for a submission. Details at User talk:Zapp#Cover of Leaving Fishers. Thanks! ···日本穣 · 投稿 · Talk to Nihonjoe 16:28, 29 July 2017 (EDT)