User:ErsatzCulture/Talk2019

From ISFDB
Jump to navigation Jump to search

Archived talk from 2019

Welcome!

Hello, ErsatzCulture/Talk2019, and welcome to the ISFDB Wiki! I hope you like the place and decide to stay. Here are some pages that you might find helpful:

Note: Image uploading isn't entirely automated. You're uploading the files to the wiki which will then have to be linked to the database by editing the publication record.

Please be careful in editing publications that have been primary verified by other editors. See Help:How to verify data#Making changes to verified pubs. But if you have a copy of an unverified publication, verifying it can be quite helpful. See Help:How to verify data for detailed information.

I hope you enjoy editing here! Please sign your name on talk pages using four tildes (~~~~); this will insert your name and the date. If you need help, check out the community portal, or ask me on my talk page. Again, welcome! Stonecreek 14:00, 20 March 2019 (EDT)

ISBN 0684138964 used twice

Hi, you were quite right with the submission, as I could check on WorldCat.--Dirk P Broer 18:40, 21 May 2019 (EDT)

Signing posts

Please sign your name on talk pages and community discussion pages using four tildes (~~~~). This will insert your name and the date. Thanks! ···日本穣 · 投稿 · Talk to Nihonjoe 20:02, 18 September 2019 (EDT)

Yes, sorry - I realized after I posted that I hadn't been signing my comments, and that the autosigner thing hadn't done it for me on the last one I posted. I did consider editing my final comment to add a signature, but given that there'd already been a couple of edit conflicts during that conversation, I didn't want to add another.
One thing I did realize in retrospect - on Wikipedia there's a toolbar of buttons under the edit textarea, which includes a "Sign your posts on talk pages: ErsatzCulture 03:52, 19 September 2019 (EDT)" reminder/button, but this doesn't appear in this Wiki. Do you know if that's a configurable thing (either per account, or site-wide), or if it's a MediaWiki version or plugin thing?
Regards ErsatzCulture 03:52, 19 September 2019 (EDT)
EDIT: I see there's a "Show Edit Toolbar (JavaScript)" option in the user prefs. However, that was already checked, and even deselecting and rechecking it makes no difference (that I can see). I don't see any JavaScript errors in the console that might indicate a failure to load the toolbar. Hmmm.... ErsatzCulture 04:11, 19 September 2019 (EDT)
Yeah, that's a bit of a pain. We use a really old version of MediaWiki software (1.12) instead of a more current version (1.34, I believe). There are many reasons for this, not the least of which is that the developers here are not familiar with MediaWiki software and don't want to break the site by updating it. At some point, it may be updated, but it won't be for a while. ···日本穣 · 投稿 · Talk to Nihonjoe 13:57, 19 September 2019 (EDT)

Kim Foster

Thanks for submitting the correct links. In theory, these should have been checked both when added and when approved (or if it was a moderator, it is one time but still links by default should be clicked before approving) but things do slip occasionally. We all are only humans after all! Thanks for all of your updates in the last days :) Annie 18:42, 27 September 2019 (EDT)

No worries, I know there's been at least one wrong link that I submitted (copypaste error), and I wouldn't expect anyone to click on all of my repetitive submissions these past few weeks. (Hope all these submissions aren't boring you and the other mods - they are primarily to aid in one of my personal projects, but hopefully having more links to Wikipedia and Twitter will be potentially useful to other people?) ErsatzCulture 18:49, 27 September 2019 (EDT)
If I am moderating you, I click on all links as a rule (and most moderators will do as well) - especially if the link is not obvious - the Wiki ones tend to be very obvious - if I miss one, it is because my finger slipped while opening them - or I did open it and misread it or my brain did not scream "but this is the wrong guy/gal" :) The Wiki links tend to be stable so it is good to have them. Same with Twitter - mostly anyway - until someone deletes their Twitter during an angry spat or another calamity. So no worries at all. Just do not get upset if you need to wait a bit when the queue get big and/or not enough people are around. Plus these are easy to moderate - click on the link, make sure it is the right person, done. They serve very well for cleaning one's palate after a particularly heavy one:) Annie 19:25, 27 September 2019 (EDT)

Duplicate tags display bug

Thanks for reporting Bug 738! You were right, the SQL query just needed to have DISTINCT added. Should be fixed now. Ahasuerus 18:18, 1 October 2019 (EDT)

Tsk - I was hoping to fix this one myself to be as an easy first contribution to the source code ;-) ErsatzCulture 18:40, 1 October 2019 (EDT)
Well, technically one might want to check the code of "adding/editing" tags and see what can be done so we do not save the duplicate tags at all (after all, fixing on the front end is always good but as it is a database, fixing the underlying data is an even better solution). If you are looking for something that is distinct enough to look at that is :) Annie 20:40, 1 October 2019 (EDT)
As I recall, there is an obscure bug in the "create/change variant title" code. Normally, variant titles (VTs) are not supposed to have tags associated with them. When a title is turned into a variant of another title, its tags are supposed to be transferred to the canonical title. Duplicate tags, i.e. identical tags owned by the same user, are supposed to be flattened into a single tag. Occasionally something goes wrong and the duplicate tag gets copied instead of deleted.
That said, I am not 100% sure that the display issue that ErsatzCulture found was due to this bug. Ahasuerus 11:44, 4 October 2019 (EDT)
I was wondering about that actually - I know we cannot add them twice technically so either we have dirty old data (which needs to be cleaned) or new doubled tags show up via varianting or merging (or both?) somehow... Annie 12:09, 4 October 2019 (EDT)

Your pending alternative name

Looking at the record that has "AE (George William Russell)" (your submission), it seems like an inexperienced editor doing something they should not (our pseudonyms can be a bit challenging) and a moderator letting it pass by mistake. Let me see if I can find out what is the exact credit in the fanzine (I suspect just AE - in which case it should have been recorded differently). I will keep the record on hold while digging around. Annie 12:18, 4 October 2019 (EDT)

No probs - I added the mod note because it felt like something that had a bit more to it than I had expertise in. ErsatzCulture 12:22, 4 October 2019 (EDT)
:) It was the correct thing to do. By the way - when you respond, instead of pressing Edit at the top of the page, press Edit on the item itself (and when you start a new topic, use the + and not Edit). See your history on how it differs (I see the latest in my Watch list - and if you have more topics, I need to chase down which you responded to instead of having the -> to get to the exact item directly) :) Annie 12:39, 4 October 2019 (EDT)

Title page

The definitive source of information is the title page. The Look Inside on the printed one does not show a title page but the kindle one does and it agrees with the cover of the printed one so we take that as a base. If a PV ever sees differently, it is easy to change. Annie 18:48, 6 October 2019 (EDT)

Thanks - I did look at that Look Inside for the printed edition, and as you say it's no help, but I didn't think of looking at the Kindle version. Will remember that for next time :-) ErsatzCulture 19:26, 6 October 2019 (EDT)
Sometimes it is misleading (Some books really have separate title pages for e- and print editions) but when it corroborates the cover, it is better than nothing. Annie 19:52, 6 October 2019 (EDT)

Title Tags on the Advanced Search Results page

Thanks for looking into this FR! I have posted your proposed solution on the Community Portal. Ahasuerus 10:37, 7 October 2019 (EDT)

Legal name

Legal names are always recorded as "Last, First Middles" :) Annie 14:18, 16 October 2019 (EDT)

Doh - I knew that, so I have no excuse for getting it wrong. Apologies! ErsatzCulture 14:57, 16 October 2019 (EDT)
No worries. I fixed it and stopped by to remind you. :) Annie 14:59, 16 October 2019 (EDT)

canonical_author fix

Thanks for submitting the "canonical_author" fix! Just to let you know where it is in the development queue at the moment: I am currently working on prioritizing newly identified ISBNs; then I'll need to take care of some software changes that are required to keep our access to the Amazon API. I hope to be able to get to your fix in a couple of days. Thanks again! Ahasuerus 14:49, 19 October 2019 (EDT)

Thanks for fixing the "as" problem! The latest version of the script worked fine on the development server, so I added a copyright statement and other administrivia to make it similar to other modules. I then committed the script and ran it on the ISFDB server without any issues. I have also added the "ISFDB developer" role to your SourceForge account -- it doesn't really give you any additional privileges at this time, but it lets me assign tickets to you. Besides, it looks spiffy! :-)
My take on this little adventure is that it was a useful exercise. We discovered that:
  • the current ISFDB software apparently runs fine under Python 2.7.15
  • trying to do development work using Python 2.7.15 can be chancy since it's not immediately obvious which Python 2.6 and/or Python 2.7 features will fail under Python 2.5.4
I guess the second bullet point is kind of obvious in retrospect, but that's the nature of retrospect -- it makes lot of things obvious retroactively.
Given the difficulties that you ran into getting Python 2.5.4 to run on your development system, I think it's fair to say that we need to upgrade the server to 2.7.x if we are to make it easy for other developers to contribute. I have bumped up the priority of SR 142, "Upgrade Python to 2.7", and reshuffled things on my internal development schedule. Thanks again! Ahasuerus 15:00, 20 October 2019 (EDT)
Thanks. I'll see if I can pick up an existing ticket - ideally a CGI script rather than a standalone - to see what Python version issues might crop up doing something more involved than a simple wrapper for a couple of SQL queries.
At the risk of raising a touchy subject (it caused some bloody culture wars at a previous place I worked at) I had some ideas about how it might be possible to use automated tests running against different versions of Python to provide some level of comfort about doing a Python upgrade without necessarily manually testing every single page or piece of functionality. This would in part (ab)use the fact that if you convert the tables to InnoDB, it seems that any database update effectively becomes a transaction that's never committed, so you can repeatedly run tests against the existing code with no ill effect (other than auto increment values increasing). At present this isn't any more than vague ideas in my head though - I'll have a play around to see if I can come up with some sort of proof-of-concept. ErsatzCulture 15:57, 20 October 2019 (EDT)
I know that test automation has been getting increasingly popular, but I don't have much first hand experience with it. Fixer, out data acquisition robot, has an iMacros adapter, but it accounts for a small part of what Fixer does and isn't used for automated testing. Also, it uses an ancient freeware version of iMacros, so it's not very robust. Something like Selenium may be a better fit, but I haven't dealt with it. I don't really have a preference as long as it gets the job done.
Re: using InnoDB without committing transations, I am not sure how well it would work with our workflow. Many, if not most, ISFDB filers update certain tables, then retrieve the newly created values and use them to build subsequent queries. Would that affect your approach? Ahasuerus 18:09, 20 October 2019 (EDT)
I've used Selenium, and it's good for what it is, but it's relatively expensive in terms of dev time, especially if your front end uses JS libraries that are a bit unpredictable in terms of what they do to the DOM. (Which was the case for me.) For ISFDB testing, I think having tests that call the CGI scripts/functions, read the HTML generated, and use something like BeautifulSoup to parse it and check it contains the expected content, would be more than good enough.
Re. InnoDB/transactions in the context of tests, I think it should be OK to have code doing lots of stuff - assuming the db connections aren't being closed and reopened. (I haven't looked at the code to see if that's the case or not.)
Anyway, all this talk of mine is cheap - I need to actually start putting something together to see if in practice the ideas I have for testing stuff would be more trouble than they're worth. ErsatzCulture 18:49, 20 October 2019 (EDT)

World Engines: Destroyer

I approved your submission for this pub. Nice job! I wish all new contributors did as well. Bob 19:42, 25 October 2019 (EDT)

However, please update your verified record (amazon is no longer the source for the data). Also, the title was changed, since it is the first in a new series. Stonecreek 03:30, 30 October 2019 (EDT)
Moved the number of pages from the Pages field to the notes in this one as per the policy for ebooks. Thanks! Annie 04:38, 30 October 2019 (EDT)

Macmillan

This is one of the publishers where we make a difference between the UK and the US operations. So even though they all say Macmillan on their title pages, when the book is a UK one, we use Macmillan UK. Changed it here as it is specifically a UK book so published there. You can follow the link to the publisher to see the notes and some links to the rest. :) Annie 03:17, 31 October 2019 (EDT)

OK, thanks. One query though - when adding this pub, I was also looking at what values the existing publications for this title had, and the export trade paperback uses "Macmillan". Is there a special case for XTPBs, or should that one be similarly updated? (This title is published by Del Rey in the US, and has a different cover, so it definitely originated from "Macmillan UK" - and not "Macmillan (US)" - even if it's - theoretically at least - not on sale in the UK.) ErsatzCulture 03:24, 31 October 2019 (EDT)
Nope - someone messed up when adding. Fixed now. If you spot something like that on a non-verified pub, just edit it. Worst case, someone will reject it. If it is verified, talk to the verifier. I am often checking our paired publishers for things like that... And yes - the airport editions are UK editions (I used to have a lot of them when I was traveling around Europe a lot) :) Annie 03:33, 31 October 2019 (EDT)
And looking at the rest, the Pan is not a pb either on the UK side - it is too big (13 x 19.7 cm) for that so fixed that as well (in case you are looking at all the editions). I think all editions look fine now. Annie 03:36, 31 October 2019 (EDT)

Collection vs Omnibus

The difference is in the types of the contents - an omnibus requires container-level titles (usually novels but sometimes multiple collections - although I would call these still collections). Binti as a series has a novel ("Binti: The Night Masquerade" is long enough for that); The Murderbot Diaries as a series contains only short fiction - none of the parts is over 40K words - so any combination of them are always collections. Hope that makes sense. Annie 12:25, 8 November 2019 (EST)

Or is "Binti: The Night Masquerade" actually long enough? Let me do some digging - because if it is not, we will have reorganization of the pubs in this series. Annie 12:47, 8 November 2019 (EST)
Thanks for the clarification.
I have a vague recollection that The Night Masquerade is a title that qualifies for both Hugo Novel and Novella categories due to the word count falling in the range where there's an overlap/fuzzy factor - but I wasn't able to find any confirmation of that after a quick look at places where similar topics have been been discussed before. ErsatzCulture 13:00, 8 November 2019 (EST)
That what I remember as well and it ended up on the novella's ballot. I need to track down the details and add some notes.
Think of the omnibus as a "nothing else fits, so it must be this" kind of container. Multiple novels always make an omnibus. One novel with stories can be a novel (if there is only 1 story), a collection (I rarely do that but it is a valid usage in some cases) or an omnibus if there are enough of them. One of the grey(ish) areas of the DB. See this for a longer explanation and some examples on the ambiguity. Annie 13:52, 8 November 2019 (EST)

ASINs

For new printed books, Amazon uses the ISBN-10 as the ASIN of the book. When that happens, we do not record them as ASINs - as we have a link to all the Amazons based on the ISBNs from the left menu anyway. I pulled it out from here. As a rule, if the ASIN does not start with B, make sure it is not ISBN-10 (and the book you are taking it from is the correct one if it is a number but not the ASIN-10 of that specific book) :) Annie 13:26, 21 November 2019 (EST)

Ah - I noticed that that ASIN was comprised only of numeric digits, which made me go back to the Amazon product page and double-check it, but I didn't think/realize that it could have been ISBN-derived. Will remember for next time... ErsatzCulture 13:44, 21 November 2019 (EST)
If the book has an ISBN and the ASIN for an e-book or a printed book is not a B-ASIN, I would generally just ignore it. Audible books are a bit different - I am still trying to figure something about some of their ASINs but for kindle/paper ones, that is as good of a differentiation as any. Either it is not your record or Amazon messed up - in both cases the ISBN will carry the day anyway. :)

The Crying Machine ebook ISBN

Pre-empting any mod question: I've just noticed that this pending submission http://www.isfdb.org/cgi-bin/view_submission.cgi?4494937 has the wrong ISBN in the ISBN field. The correct one for the ebook edition (which is in the info in the notes field) is 9780008308797.

If no mod cares to edit my submission, I'll do a follow-up submission to correct this error. (I don't see that I can edit a pending submission, and I'm too lazy to delete and resubmit it.)

Nope, no edit on already submitted submissions. Approved and fixed. Annie 21:45, 8 December 2019 (EST)
Thanks :-) ErsatzCulture 05:35, 9 December 2019 (EST)
Retrieved from "https://isfdb.org/wiki/index.php?title=User:ErsatzCulture/Talk2019&oldid=587865"