Wikidata talk:WikiProject Books/2021

From Wikidata
Jump to navigation Jump to search

Best way to indicate a cross reference subpage?

The listing of articles within an edition is always an interesting exercise. When I have done such the biographical and encyclopaedic compilation I generally have ignored that they are their own editions, and labelled them as articles of some sort.

Also in some of these works there is a 'see cross-reference/redirect. Where they have been reproduced in DNB they were made an instance DNB redirect page (Q19648608) (subclass of cross-reference (Q1302249) and Wikimedia redirect (Q21528878)). Now a bit stuck with how these would be represented where they are recreated at WikiSource, as there is no generic equivalent and looking to utilise something on O'Kelly, John Thomas (Q99196993) as both the subclass items mentioned throw up a violation. I don't wish to create a DNB equivalent in this case. @ChristianKl:  — billinghurst sDrewth 22:36, 28 December 2020 (UTC)

  • I am saying what currently exists within DNB and has existed for years, and was what I attempted to copy. So would prefer solutions of what we can do, rather than to be told what I shouldn't.  — billinghurst sDrewth 07:16, 3 January 2021 (UTC)

Instance of literary work vs book

I've created my first WD item for a book, which will be followed by matching, separate items for each of its four translated editions (all by different publishers, different dates). Stuck on the first Statement: Instance of - which is correct, literary work or book? The "book" has won numerous awards, so does that relate to "book" (as a publication) rather than the creative aspect of a "literary work"? -- Deborahjay (talk) 16:04, 3 February 2021 (UTC)

I think a "book" is an edition, and you'd want a literary work item to link together multiple editions. Ghouston (talk) 22:36, 3 February 2021 (UTC)
A book is not an edition in my camp, and I avoid the term book as it is too indeterminate IMNSHO. Use of literary work can never be wrong, then I always just use edition for each version. About the only time that I vary is when it is a speech, that is later published, so I will always identify the speech and the date it was given as a different criteria. @Deborahjay: Can I suggest that you utilise the WE framework gadget/tool as it allows you to do a work or an edition and populate all in one fell swoop. It is set up at English Wikisource as a gadget, though set mine to run globally through my global.js file.  — billinghurst sDrewth 04:55, 4 February 2021 (UTC)
We've decided before not to use "book" since it can refer to many different levels: a work, an edition, a physical copy, a format of manufacture, a section of a classical work, and many other things besides. --EncycloPetey (talk) 01:35, 3 March 2021 (UTC)

book edition (Q57933693)

Hi to all. We either need to build Q57933693 into our documentation, or we need to get it otherwise managed. It just becomes an extra level of confusion and management to have random unattached items out there.  — billinghurst sDrewth 01:12, 19 October 2020 (UTC)

This data item is confusing is it limited to editions of "books" that are themselves "books", and does it include e-books, audio-books and other formats, or only printed ones bound with a spine? The explanation isn't clear on this, and the linked Commons category includes "Bible" editions, which is sufficiently broad enough to render most definitions meaningless. In short, what is the data item's purpose and scope? --EncycloPetey (talk) 01:19, 9 March 2021 (UTC)

Conflicting information

The identifiers on Nightmare Japan: Contemporary Japanese Horror Cinema (2008 Rodopi Publishers ed.) (Q105713976) are giving conflicting information regarding publication date and number of pages. Can anyone help me sort it out? --Trade (talk) 00:56, 3 March 2021 (UTC)

For the date, I'd probably go with 2008. Amazon has a date on its page of November 19, 2007, but using the "look inside", the date printed in the book is 2008, and that's what other sources have. "1 January 2008" is probably just another way of saying 2008. If you also look inside at the table of contents, the numbered pages run from 1 to 217 plus the size of the index, but if you scroll to the end of the book, it finishes on page 219. I don't know whether the page count in Wikidata is supposed to include front matter, etc., though. Property talk:P1104 isn't entirely conclusive, to me. Ghouston (talk) 01:46, 3 March 2021 (UTC)
I can see that both OL8581335W and OL12805120M share the same ISBN numbers although i'm not sure if this is done on purpose. I know very little about OpenLibrary. --Trade (talk) 17:50, 6 March 2021 (UTC)
Identifiers ending in W are for works. Identifiers ending in M are for editions. --EncycloPetey (talk) 18:52, 6 March 2021 (UTC)
But OL8581335W are linking to Nightmare Japan: Contemporary Japanese Horror Cinema (2008 Rodopi Publishers ed.) (Q105713976) which is a book edition, not a work. --Trade (talk) 00:31, 9 March 2021 (UTC)
Open Library seems to mix in edition details when you view the work item. I'm not sure how it picks which edition to display, in cases where there's more than one. Ghouston (talk) 02:00, 9 March 2021 (UTC)
Trade, OpenLibrary often makes errors where work and edition information get mixed together incorrectly. It happens here on WikiData too. But OpenLibrary items ending in "W" are supposed to be works and those in "M" are supposed to be editions. But you will find situations where the data at OpenLibrary violates this standard. --EncycloPetey (talk) 07:18, 9 March 2021 (UTC)

Massive repurposing of items

WikiProject Books has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

Hi everyone. It seems that at least two contributors, Francesca Pisano and Annacarmen Plaitano, whom I strongly suggest to read this project page, are repurposing dozens of works with editions. I fixed a few, but that's a heavy workload. If anyone could lend a hand it would be greatly appreciated. --Jahl de Vautban (talk) 05:19, 13 April 2021 (UTC)

For the record, @Алексей Скрипник: and @Billinghurst: left a similar comment on User talk:Francesca Pisano. Cheers, VIGNERON (talk) 10:22, 20 April 2021 (UTC)
Oh, I only saw one from my maintenance, so cannot say I am around the contributions particularly. Typically I saw no merges or interwiki link additions so I didn't look deeper. <shrug>  — billinghurst sDrewth 10:29, 20 April 2021 (UTC)

Articles within our books

I cannot say that I have seen us have a discussion nor give guidance about how we itemise the articles within our published books.

Something like Abraham, Lieut.-Col. James Johnston (Q67202946) and s:en:Thom's Irish Who's Who/Abraham, Lieut.-Col. James Johnston.

I had the conversation Wikidata:Project_chat#Use_of_title_(P1476)_with_biographical_article_(Q19389637)_items? which is going to now have me go back and put titles on each of the 0000s of articles that I have created.

I typically don't mark then as editions, and use a form of the word article, as I have them specific to the work. I have always added the author, though usually left the language out as it is covered by the parent IMO. I also had someone come and put follows/followed by on this particular work, and now as a qualifiers.

So to me I think that it is within the remit of this project to look to what we considered mandatory, what is preferred, and what is optional or to be considered.

Do we do this, though others may disagree.  — billinghurst sDrewth 22:39, 20 April 2021 (UTC)

@billinghurst: I think there was discussions, at least I clearly remember the example of the DNB and items like Abbadie, Jacques (DNB00) (Q19027532) (which is a bit poorer than your item). Anyway, this item seems good to me. Indeed when we are this granular the work/edition model does not apply anymore, "article" is the right P31. P1476 could be added in bonus but not formally needed. Cheers, VIGNERON (talk) 06:54, 21 April 2021 (UTC)
@VIGNERON: Thanks for your feedback. Do you think that we can own setting up the basic guidance, or do you think that that someone else may own it? If not us, which project? If us, then I will scope up something for us to chew upon.  — billinghurst sDrewth 07:01, 21 April 2021 (UTC)
  • @billinghurst: It's my understanding that any analytic (part of a book or periodical, i.e., chapter, journal article, encyclopedia entry) is a FRBR work level entry, and should have P1476. This makes sense in that a chapter, journal article, or encyclopedia article could have a version, edition, or translation. In practice for your example, I'd focus less on P1476 and add inception and language. Prburley (talk)
    But an individual data item will be either a work or an edition/version. It should not be both. A journal article in a specific journal, with date, place of publication , issue info, etc. with be a version/edition, not a work. A data item for a work would have no information specific to where it was published, since that information applies only to editions. --EncycloPetey (talk) 16:17, 30 April 2021 (UTC)
    Either way, it seems that there is considered value in adding title whether it is explicit or implicitly applied; and there is the indication that explicitly adding language at an item level could be part of a schema, rather than relying on an implicit assumption of the parent item, so adding it is what I will do. Doing so is pretty easy using WEF, so I will just do it.  — billinghurst sDrewth 02:14, 1 May 2021 (UTC)

What to do about existing mixed work/edition items?

According to this query we have more than 16.000 items with ISBN-13 (P212) and publisher (P123) (indicating it is a book edition), not being instance of version, edition or translation (Q3331189), neither having an edition. I found two cases:

Could we make the first or both of these cases also instance of version, edition or translation (Q3331189)? -- JakobVoss (talk) 07:59, 14 December 2020 (UTC)

  • I made a suggestion at Wikidata_talk:WikiProject_Periodicals for an item class representing a conflation of literary work (Q7725634) and version, edition or translation (Q3331189). I'm assuming that many articles published in periodicals, including academic articles, will never have separate edition items, so it should be handled explicitly somehow. A problem, however, is what you do if you decide later to split the work and edition, since the item may be linked from elsewhere as either a work (such as from VIAF) or as an edition (such as a reference in Wikipedia). This problem would also occur with conflations as they are currently encoded. Perhaps you would just leave the item as a conflation of a work and one of its editions, with additional edition items created as required. Ghouston (talk) 01:04, 29 December 2020 (UTC)
    (general comment about what I have done around nexus). It is an imperfect model as there so many variations. My approach (not calling it a solution) has been to list biographical compilations as version, edition or translation (Q3331189) instance, and probably as a own work item (if I can be bothered). Each biographical component, has been listed as biographical article (Q19389637) and is linked to the edition with published in (P1433) as I think that either the whole work is reproduced as a new edition, never really as individual articles. Example Men of Kent and Kentishmen (Q101589898) containing William Prude (Q104636618). yes less than perfect.
    Follow similar process if I have a compilation of papers; or chapters of a book. Works, though imperfect.
    Poetry compilations, it doesn't work as individual pieces get a life of their own, in republishing by author, by other authors, or translations. So I always make each item Q3331189, and will look to complete a parent item.
    there are other variations and it defeats good rules, so to me my internal guidance is to whether it is referenced (article/chapter/...) or readily reproduced (separate edition). I think it falls outside of any guidance, but it is my working practice (until someone gives me something better).  — billinghurst sDrewth 13:30, 7 January 2021 (UTC)
  • The problem with that is that such conflations aren't a very good database model. Perhaps a better idea would be to accept that for many works it's only necessary to have an edition item. The "work" item would only be needed if you want either a) to link multiple edition items together b) include external links to "work" items in other databases. Ghouston (talk) 06:05, 3 January 2021 (UTC)
I agree that it's not a very good database model, and would go further to say that for the consistency of the data, each edition linking to a work via edition or translation of (P629) would be great: that's for instance a pre-requisite for a project reusing Wikidata books data such as inventaire.io (Q32193244), where we expect to find certain properties on the work and others on the edition. -- Maxlath (talk) 07:30, 2 May 2021 (UTC)

bypassing site redirects and their conflicts

I want to use s:en:The Bronze Ring at The Bronze Ring (Q7720315). At source, it is a redirect to the only version there The Bronze Ring (Q106714641), which should be a good way to manage it until another version arrives there (at source).

The only option I get, even with the "intentional redirect" from the site link menu is to merge, which is the wrong publication, language, etc. It is a version, edition, translation. The original will never be at that wiki, other versions might never be there.

I feel that I am being thwarted at the rule following....--RaboKarbakian (talk) 17:20, 5 May 2021 (UTC)

This is part of a discussion above. An edition of a text should not be added to the general data item for the work. Create a data item for the edition, then list it as an edition from the main data item for the work. --EncycloPetey (talk) 22:41, 5 May 2021 (UTC)

 Comment we shouldn't be using redirects as item targets. Until English Wikisource has a versions page for that work, we would not create that as an independent page. The logics of interwikis is set up so that it all links well from the article if you set up the "has edition" <-> "edition of" properly.  — billinghurst sDrewth 23:42, 5 May 2021 (UTC)

translated edition (Q39811647)

Hi. This item has it listed as a subclass of version, edition or translation (Q3331189). Do people think that it is the case? It adds a level of confusion to have something labelled "work" under the edition label. I know that we tend not to emphasise a translation and it has to be somewhat interpreted by the change of languages, so I am not adverse to a better indicator, but this item name is not one that I favour (and maybe it is just an English language issue.)  — billinghurst sDrewth 13:04, 10 May 2021 (UTC)

And if it is used, should we be qualifying it? Depending on thoughts, Cinderella, or the Glass Slipper (Q19112901) is what as I see as the conflicted item, and probably others.  — billinghurst sDrewth 13:08, 10 May 2021 (UTC)
The question is, if we do change the label, what would we change it to? Most options I can think of have their own problems with interpretation of what they mean. --EncycloPetey (talk) 20:25, 27 May 2021 (UTC)

Adding Price One Penny data about 19th-century cheap literature

WikiProject Books has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

I would like to contribute the data from my database Price One Penny to Wikidata. I've already created an element (Price One Penny (Q106923678)) that I've used for instance as a reference regarding the creator of Sweeney Todd (Q669485). I was wondering if you could please chime in on my 5 generic property proposals for external identifiers (for authors, publishers, works, periodicals, and libraries). Thanks! Marianika (talk) 02:10, 27 May 2021 (UTC)

Serial literature properties

The works in Price One Penny (see above topic) were either serialized in periodicals (and I'll alert Wikidata:WikiProject Periodicals about the incoming data) OR issued in weekly numbers to be bound together (or not if you didn't have the money).

  1. I would like to add a unit "numbers" (livraisons in French) to number of parts of this work (P2635) to capture this important property of serial literature. The number of numbers is as a crucial a bibliographic data point for this type of literature as the number of pages, even more so as it gives an indication of the starting date. I see through the Wikidata Query Service that it isn't a unit currently used. Should I just use livraison (Q106642729) as a unit?
  2. The edition item properties already provide for start time (P580) and end time (P582). The work item properties in contrast have inception (P571), which I assume should be the same as the start time (P580) for the first edition. The problem is that the most commonly associated date with works of serial literature is the end time (P582) because that's the one that appears on the title page or the preface (issued with the last number). Could we provide guidance about this on Wikidata:WikiProject Books page?

Thanks and I look forward to working with y'all! Marianika (talk) 02:28, 27 May 2021 (UTC)

As for the first question, we already have series ordinal (P1545) for numbering items in a sequence. It can be included as a modifier to do what you need. Or do you mean adding "parts" as in "5 parts"? That would be an issue because it is an edition that has that many parts, not the work. As the Property is currently named, it creates a confusion between "work" and "edition". We might want to have a separate property altogether. --EncycloPetey (talk) 20:22, 27 May 2021 (UTC)
@EncycloPetey: According to the WikiProject documentation, number of parts of this work (P2635) is an edition-level property. The label is therefore indeed misleading. (And yes, I mean "parts" as in "5 parts".) Looking at the units currently used, they're all dependent on the edition, so I guess it's just a matter of changing the label rather than splitting the property in two. Marianika (talk) 22:46, 27 May 2021 (UTC)
Are livraison (Q106642729) and livraison (Q97103453) identical, related, or two completely different concepts? --HHill (talk) 09:45, 1 June 2021 (UTC)
@HHill: I would think that they are identical. I notice that the Wikipedia.de page from which livraison (Q97103453) is drawn links to Wikimedia Commons Category:Book installment. I don't know if that should be linked to the Wikidata element too. Marianika (talk) 18:18, 1 June 2021 (UTC)
These items should probably be merged then. Commons category (P373) could be added too. HHill (talk) 18:51, 1 June 2021 (UTC)
done|merge done}}  — billinghurst sDrewth 12:00, 2 June 2021 (UTC)

The Navy List

Hi. I am trying to create an item to represent the Royal Navy's Navy List (Q6982656). This is so that I can reference it as a source for claims on related items. The list contained details of the Royal Navy's ships, facilities and key personnel. Replacement versions of the list (containing the most current information) were published several times a year. I am struggling to decide on how best to map this in Wikidata. Should it be a single work that has many editions, or is each new list an independent work with its own edition(s)? For example, are the January 1914 and February 1914 lists both version, edition or translation (Q3331189) of the single "The Naval List" written work (Q47461344)? From Hill To Shore (talk) 21:58, 31 May 2021 (UTC)

@From Hill To Shore: I would treat it like a serial. Then use published in (P1433) and utilise volume/page/... or whatever combination of identifying features it has.  — billinghurst sDrewth 08:17, 1 June 2021 (UTC)
@billinghurst: Thank you for the advice. I've not created a serial item before. Could you please point me to a good example of a serial item that I could use as a model? From Hill To Shore (talk) 18:24, 11 June 2021 (UTC)
@From Hill To Shore: Not something that I have needed to do either. If it has an ISBN maybe something like Jane's Fighting Ships (Q1343303) or Who's Who (Q2567271) or Encyclopædia Britannica (Q455). Does it have or an ISSN? What is the numbering like government gazette (Q2065227), like The London Gazette (Q1137581). I don't really know enough about the work itself to give good advice, or maybe it is more like a journal.  — billinghurst sDrewth 00:55, 12 June 2021 (UTC)
It doesn't have an issue number and most editions predate the introduction of ISBN or similar cataloguing systems. It is possible that some libraries have restrospectively assigned catalogue numbers to them, but that will require further research. As the publication took place over a few centuries, the format may have changed significantly. However, most of the copies I have seen simply stated "The Navy list corrected to x day, y month, z year." Here is an example from 1856 at internet archive,[1] and here is a later copy from 1921.[2] The closest equivalent I can think of is a phone book. The new phone book lists the names, addresses and phone numbers of the people currently resident in an area; some people will have been added compared to the previous edition and some will have been removed. From Hill To Shore (talk) 01:54, 12 June 2021 (UTC)
It looks like a government gazette to me. I'd just follow what is in place for London Gazette, that is close enough IMNSHO.  — billinghurst sDrewth 05:47, 12 June 2021 (UTC)

P747 and P629

Hi y'all,

has edition or translation (P747) and edition or translation of (P629) being inverse of each other and given the usual uses and the documentation of this project, it is my understand that these two properties are obviously incompatible.

I added the constraint but Jura1 reverted me (Special:Diff/1405103592).

I also started to clean all few items who wrongly use both property: 

SELECT DISTINCT ?q WHERE {
  ?q wdt:P747 ?a ; wdt:P629 ?b .
}
Try it!

WikiProject Books has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.: what do you think ?

Cheers, VIGNERON (talk) 10:20, 20 April 2021 (UTC)

I would think that they are mutually exclusive.  — billinghurst sDrewth 10:31, 20 April 2021 (UTC)
Can you not have A that is a translation of B that is a particular edition of C ?
Then both properties might appear on B. Jheald (talk) 10:46, 20 April 2021 (UTC)
So, in the case of different authors (as for translations), different illustrators (or lack of illustrator), or, different formats (as in serialized articles being published later as books) there is only one literary work for all? And Oliver Twist? Dickens serialized this in Bentley in 1837 and then again in Pickwick in 1839 and booked it in 1839 or 1840. Ingoldsby Tales, serialized articles (the articles being all over the place, individual and grouped and none called "Ingoldsby Tales"), and a book where the book has a further artists edition....
I have been treating publications as a literary work which is an edition of the first literary work if that pub has a different publisher, illustrator, or author, especially if there are different editions of that version of the work.
Meanwhile, there is the case of Mother Goose, which is a translation but that a bot (I think its a bot) always changes.--RaboKarbakian (talk) 13:29, 20 April 2021 (UTC)

Consider this, that the combination of "some work" (although, there has been no two person instigated attack upon "scientific work") with "version" is like a stairway with the flat part being the "work" and the vertical part being the "version of" leading to the real first -- whatever mess that was. Peter Pan in Kennsington (a few chapters from The Little White Bird) and really, there ie no way to make a singularity of Ingoldsby Tales, even if I get my staircase back from these two people. Help me figure that out! The problem is best illustrated at commons:Category:Richard Harris Barham here, the first is being attributed to the later Rackham version. It is a better project for the two of you. Let me know what you figure out!--RaboKarbakian (talk) 14:12, 20 April 2021 (UTC)

@RaboKarbakian: that's a lot of very good questions, thanks.
A work and an edition always has the same author in common (author (P50)). Others « authors » are not really the author ; on the edition, we either use other values as secondary authors or use other properties (editor (P98), translator (P655), illustrator (P110), contributor to the creative work or subject (P767), author of foreword (P2679), author of afterword (P2680), etc.).
Series can indeed be complex, but again we use other models and properties, like part of the series (P179) or the pair has part(s) (P527)/part of (P361) (same for part of work, see Genesis 1 (Q5532790)).
For a work deriving from an other work (both being works), then we also use other properties like based on (P144) and/or inspired by (P941). For instance: (that you added yourself).
All of this is documented on the main page Wikidata:WikiProject Books and has been discussed in previous discussion (see the archives since 2013).
« I have been treating publications as a literary work », well this is illogical, a work is by it's very definition unpublished. You can only encounter problems with this approach.
Anyway, I don't see a case where a concept can be both a work (general, intangible, unpublished) and an edition (specific, tangible and published). You can easily make a parallel with Wikisource, there is a page s:en:Romeo and Juliet (Shakespeare) (the work) and a page for each edition but there is no page containing both.
PS: could you give links to the items you are refering too?
Cheers, VIGNERON (talk) 18:18, 20 April 2021 (UTC)
@VIGNERON: The link to the commons was as good a link as any since I have been working there and here up to the Ingoldsby Legends here. commons:Category:Bentley's Miscellany is where the real first editions start. (even that, I am not completely sure). I am, due to my usual contribution and greatest experience, working mostly with illustrators and scans at the commons. Scans and illustrators usually also have authors with them. It is always a new edition when it has illustration (for the first time) or different illustrations, whether the book claims it or not. Then the illustrated editions get reversioned with more or less new illustrations, as what happens with most of Rackhams stuff, which is almost never (with exception of his Fairy Tales) never a first edition.
Tales from Shakespeare has maybe ten different illustrators, all in Public Domain!
I just was working on Flatland, in honor of billinghurst (who makes wonderful author data pages here, one single flat truly perfect item, each), where I am certain to find different scans of the second edition, with different ia values which can all be listed with the gutenberg edition as versions of the second edition. The second edition gets one entry at the First edition data point. See Flatland at Open Library for what happens when you don't do that.
I was feeling wrong about having wikisource links getting lost in the versions, but by having each scan linked to the index page and the Main linked to the Second version, and the Source versions page linked to the "literary work" all is good! s:en:Flatland you can get to wd from there. Please let me finish it before you start removing things. It will be a nice example of what I have worked out.
As for me providing links here, I need a computer, and, almost everthing I had mentioned is wrong -- I was working up to fix them from commons. I can do a lot with my mobile device, but everything takes so much more time....--RaboKarbakian (talk) 19:03, 20 April 2021 (UTC)
Ok, I'll let you work (10 illustrators is not a problem, The Constitution of India (Original Calligraphed and Illuminated Version) (Q21177783) has 13 of them and Flatland (Q728312) seems mostly good). I don't see the relation with an item needing both P747 and P629 at the same time. Cheers, VIGNERON (talk) 19:26, 20 April 2021 (UTC)
  • @Jheald: All translations necessarily come from an edition, they don't come from nowhere, so we already face that aspect. The previous discussion about translations, and later if there were subsequent editions of translations were editions of a translated linked from a new subsidiary literary work, or all originating from the primary work => determination was all link back to "primary work". So we have crossed that bridge and dismissed it. So my understanding is that we have a primary work and in all cases != edition. I would think though that we could consider an extension to based on (P144) for the use case you propose.  — billinghurst sDrewth 02:09, 21 April 2021 (UTC)

WikiProject Books has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. any more comment or remark? So far, I don't see any good reason for @Jura1: revert. Cheers, VIGNERON (talk) 11:39, 25 April 2021 (UTC)

They are not mutually incompatible. There are Greek editions of the Bible that have been translated and these translations have gone through multiple revisions and editions. Johan Huizinga's book Herfsttij der Middeleeuwen was translated into French, and the French translation was then translated into English as The Waning of the Middle Ages; but the original Dutch was translated into English as The Autumn of the Middle Ages. Anna Swanwick's translation of the plays of Aeschylus went through multiple editions, so it is a translation that has editions. Editions of translations are extremely common, so the two properties are not at all exclusive to each other, and Jura's revert is correct. --EncycloPetey (talk) 18:44, 25 April 2021 (UTC)
I agree with @VIGNERON - they are mutually incompatible... you mix up "translation" of a work, and the specific "edition" of a work that publishes the translation...
if you need to convey the idea that a specific translation was based on a specific "edition or translation", use based on (P144), or expression of (P6524)/manifestation of (P1557), not P747/P629--Hsarrazin (talk) 08:21, 30 April 2021 (UTC)
I did not "mix up" edition and translation. The property edition or translation of (P629) is for "edition or translation of". WikiProject Books has decided editions and translations are the same thing and should be treated the same. --EncycloPetey (talk) 16:19, 30 April 2021 (UTC)
Yes, treated the same as being derived from the **parent work**. I think that the indication from Hsarrazin is that if you wish to say that a specific translation is derived from a specific edition that the way to indicate that is using P144 or P6524 or P1557, but **do not use** P629. That being the case, then the logic is that the pair are mutually exclusive. We should be putting some good documentation/guidance on WS:Wikibooks and P629 that directs people how to manage these aspects, as it all derives from our decision that all editions and translations derive directly from the work.  — billinghurst sDrewth 02:23, 1 May 2021 (UTC)
+1 with @billinghurst:. Plus, @EncycloPetey: you can't say « should be treated the same » and then treat translations differently (by proposing to create regular edition sometime and Frankenstein-y edition of edition sometime), this is completely illogical (and obviously with this illogical approach, you can only end up with illogical items). Cheers, VIGNERON (talk) 12:27, 1 May 2021 (UTC)
@billinghurst:and @VIGNERON: That does not solve the problem of editions of published translations. Because property edition or translation of (P629) is for "edition or translation of", both editions and translations are handled by the same property and are listed together on the parent data item. If one of those translations has multiple editions, then those are listed on the translation as editions of that translation. I raised this issue as a problem years ago, but this community decided to treat editions and translations the same. --EncycloPetey (talk) 14:37, 1 May 2021 (UTC)
@EncycloPetey: « editions of published translations » or « editions of that translation. » make absolutely no sense, it means nothing and it's not even possible, no need to make up false problem. For « edition/translation based on other edition/translation » the solution is simply to use based on (P144) (like said many times already in the last years, and four times in this conversation only). « this community decided to treat editions and translations the same. » this community decide to follow what lirbarians and cataloguers are doing for centuries. Cdlt, VIGNERON (talk) 16:06, 1 May 2021 (UTC)
That is not what based on (P144) "based on, adapted from" is for. I have a copy of Swanick's translation that is labelled 5th edition. What is it the 5th edition of? --EncycloPetey (talk) 20:02, 1 May 2021 (UTC)
@EncycloPetey: why would it not be? This seems exactly how it's used for years now (along with other uses and other properties, cf. supra), as documented and discussed many times. « What is it the 5th edition of? » it's the 5th edition, more exactly the 5th edition of the work by an editor and based on a translation (same remark, with properties like edition number (P393)). Cheers, VIGNERON (talk) 22:11, 1 May 2021 (UTC)
And what happens when I have a 2nd edition of Storr: it's the second edition of a "work" by Storr? But the text of the work is not a translation but an edition of the original text by a Classical author in the original language? --EncycloPetey (talk) 22:24, 1 May 2021 (UTC)
I want to be sure we nail this down, because if we follow what you are saying, we have to change thousands of data item linkages. Right now Iliad (Q8275) has 30 instances of has edition or translation (P747). But you have told me these are not editions of Iliad (Q8275) because they are translations. This means they are translations of works by editor/translators, which would therefore have separate data items, and so none of the Wikisource listings of translations should be linked at Iliad (Q8275), since they are editions of translations. Your viewpoint is NOT what everyone at the Wikisource projects has been doing for the past five years. --EncycloPetey (talk) 22:34, 1 May 2021 (UTC)
What's wrong? It's seems you try to break the model on rare specific cases (which in itself is good to make sure the model is correct and resilient) but I don't see where there is any problem here and you don't seem to understand what we are all saying (« these are not editions of Iliade (Q8275) because they are translations » or « they are editions of translations » make no sense, editions are translations and translations are editions and there can't be edition of edition). If there is problem, then open a different section with clear cases. And to go back to the original question, I don't see how your examples involve having a Schrödinger item both published (edition) and unpuslished (work) at the same time. Cheers, VIGNERON (talk) 22:44, 1 May 2021 (UTC)
How is the Iliad a "rare specific case"? Every single one of the volumes listed under has edition or translation (P747) on Iliad (Q8275) is an edition; sometimes (but not always) the first edition of a translation. And sometimes the second or third edition of a translation. Please explain how they are an "edition or translation" of Iliad (Q8275) but the 1st, 2nd, or 3rd edition of a different work that they are based on. What you have been saying makes no sense. --EncycloPetey (talk) 00:42, 2 May 2021 (UTC)
@EncycloPetey: there is an estimated of 200 million books, 99 % of ttem are mono-edition, monolingual, untranslated, and/or old. The Iliad is clearly in the 1 % (like most classics and best-seller). The Iliad is one and only work and everything else is editions (of the one and same work). Then some these editions are based/inspired/derived from other editions (and not work, a work is intangible, you can't really derive anything from it). Anyway, I took a deeper look at Iliad (Q8275) and everything looks fine (there is some bits of unnecessary redundant information, but that's a detail). Cheers, VIGNERON (talk) 08:34, 2 May 2021 (UTC)

 Comment (outdent) @EncycloPetey: I hear you about the model here not being perfect, and that was a conversation that was had here years ago. I know, as I raised it. I accepted that it wasn't going to happen so in our consensus model, I moved on. The outcome of that discussion was as previously mentioned that ALL editions and ALL translations are of the WORK. As such under the existing model a NEW translation is itself not its OWN WORK despite it having its own copyright. Yes, Wikidata's model basically ignores that translations and different translations have editions, they are all subsidiary to the WORK. Shit happens.

If we accept that is the WD model then with the accepted basis for how we are working THEN these two discussed properties are mutually incompatible in this model. That is the entirety of this discussion "of this model". With the limitations of the model, the only means to show editions of translations is to use an alternate property, and three were proposed, so we either choose one and document it well; or we ask for a new property (though doubt that will happen). If we wish to prosecute the case that the model is wrong, as the basis for has edition or translation (P747) <=> edition or translation of (P629) then it should be a separate and higher discussion, trying to prosecute that case in this discussion doesn't make sense and just causes things to unnecessarily stall.  — billinghurst sDrewth 04:06, 2 May 2021 (UTC)

version, edition or translation (Q3331189) is a subclass of work (Q386724) though. Ghouston (talk) 01:53, 3 May 2021 (UTC)

@VIGNERON: Rather than deleting links to projects [3], please correct them like this and like this. Deleting links destroys data and does not help editors understand how to do things correctly in Wikidata. --EncycloPetey (talk) 15:37, 3 May 2021 (UTC)

@EncycloPetey: I'm maybe wrong but I prefer good data rather than bad data. If you're so keen and correcting, I started Homeri Opera (Q106691509) but there is still a lot of work to do, si please correct them if you want to (meanwhile, I'm checking that this book should not be deleted, the copyright is not clear here). Cheers, VIGNERON (talk) 16:04, 3 May 2021 (UTC)
Are you certain it was that edition of the Greek text that el.WS used for their copy? How did you determine that, since their listing has no bibliographic data? And yes, I am keen an correcting. That's why all the other Wikisource listing have been corrected for the works of Homer, Pindar, the Greek playwrights, and Shakespeare, among others. Only the el.WS listings are a problem, because they have no bibliographic data on their associated talk pages. Without knowing the edition that served as the source text, they cannot be fully corrected. --EncycloPetey (talk) 22:38, 5 May 2021 (UTC)
Sometimes I delete them, sometimes I fix them (which means creating 1, 3, 4, 5, ... edition items and crosslinking). Fixing them can be a huge amount of effort, and sometimes I don't have the time or here doing other tasks. I know that I have created a query for enWS where there are issues and posted to s:en:WS:S so everyone there can get in and fix/resolve. The best outcome is not to do them in the first place, and to go back and look at one's own works added. I know that any of mine have been fixed. Also educating users, and I know that I have left messages for people this week about their merges, their additions, and their changes of my additions.  — billinghurst sDrewth 10:32, 4 May 2021 (UTC)

Hi! A bit a food for thoughts/request of help. On Q19173337, I want to say that this translation is based on an older translation by Vincent Thuillier (Q3559858), which I plan to do by using based on (P144). Thing is, according to the BNF there is at least 6 differents editions of this older translation (none of which is yet on Wikidata), and I can't say for sure which one is used without a full philological study of the text, if it is even possible to tell. So I'd like to say that Q19173337 is based on this older translation, generally speaking without pinpointing to a specific edition, as if it was a work instead of an edition. For what I understand from the previous discussion, this is actually not desirable. What do I do? --Jahl de Vautban (talk) 08:40, 9 July 2021 (UTC)

has edition or translation and edition or translation of in the same item

I'm not sure if that should be considered exception to constraint (P2303) or something is not properly linked.

There is a constraint violation in Nomenclature of Organic Chemistry. IUPAC Recommendations and Preferred Names 2013 (Q54019023) because of the presence of edition or translation of (P629) and has edition or translation (P747) in the same item. Wostr (talk) 12:27, 7 July 2021 (UTC)

When this was raised at an earlier time the decision was that the translation is of the original concept item, not of the edition itself. So the linking should be Nomenclature of Organic Chemistry (Q3343212) <=> Nomenklatura związków organicznych. Rekomendacje IUPAC i nazwy preferowane 2013 (version 1) (Q55809949), and not be concerned about the English language 2nd edition.  — billinghurst sDrewth 13:05, 7 July 2021 (UTC)
There are translations of both 1st and 2nd edition, so both translations should be in Nomenclature of Organic Chemistry (Q3343212)? The differences between 1st and 2nd editions are huge (559 pages vs 1568 pages), so the lack of any link between Nomenclature of Organic Chemistry. IUPAC Recommendations and Preferred Names 2013 (Q54019023) and Nomenklatura związków organicznych. Rekomendacje IUPAC i nazwy preferowane 2013 (version 1) (Q55809949) is a significant loss of data... Wostr (talk) 14:16, 7 July 2021 (UTC)
@Wostr: I am telling you the outcome of the discussion and what the direction was when I raised the issue.The conceptual item is the conceptual item.

The only other way to manage that is to say that the 1568 pp work is a new conceptual item based on (P144) the first conceptual item. Then have respective editions/translations that link to the second conceptual item. As you say it essentially a different work.  — billinghurst sDrewth 01:15, 8 July 2021 (UTC)

Items for scans of editions of works?

Do we want to make items and statements like this? This links to The Ingoldsby Legends; or, Mirth & Marvels (Q107303815), which is a scan of The Ingoldsby Legends; or, Mirth & Marvels (Q107303799), which is an edition of The Ingoldsby Legends (Q7742022).

@RaboKarbakian: does this solve a specific problem? I don't see why the statements for the scan can't belong on the item for the edition. If there's a good case for the scan items to exist, maybe they should be linked by digital representation of (P6243) or depicts (P180), or any other property than has edition or translation (P747). --Azertus (talk) 22:32, 13 July 2021 (UTC)

@Azertus: While this might be a little backwards in the eyes of the project here, I have been working at the commons with the book template and the infoboxes there. The results are beautiful and it makes sense when the big picture is looked at, at least, it did for me. A very nice example of having the scan in its own separate edition can be seen at commons:Category:Handbook of meteorology. The "image" being the title page. When there is an "image" image (P18) and a scan document file on Wikimedia Commons (P996), things (the book template in particular) start breaking. As for the book template, after some chat with a bot operator there (at commons) who was in the process of slapping an inception date inception (P571) upon everything, we determined that using the book template for the images from that book would be the best solution. Really, I just wanted to keep the inception date open for if the image had a date associated with it that would naturally be different from publication date publication date (P577). Once the compromise was achieved, the results were so much more than I could have asked for!! See commons:File:Handbook of Meteorology-172.jpg, the book template is displaying the page with the original picture, the information is being pulled in from the scan item Handbook of Meteorology (Q55345046) the structured data contains the publication date and the inception date. The images go with the scan and the index at wsource (which is linked to by every image and the scan) and the main (that will contain editions Handbook of Meteorology (Q55345046)) goes with the category and with the transclusion at wikisource (as is displayed and linked to in the infobox).
I am unable to justify this using the language or the consensus, but the practice and use and the negociations and the right things pointing to the other right things--I was simply trying to make everything work, and in my opinion, I succeeded. Could we have a vote, maybe, on using methods that work? Especially when they work better than expected?
The Ingoldsby Legends are going to be a pain. They were originally published in Bentley's Miscellany (Vols 1-17<--YES that is a "seventeen"), and not even serialized, (meaning, on a regular schedule, more like whatever and whenever). I was up to Volume 6 collecting them and have since taken a (microfilm) break. So, there are 4 definably different publications of that before World War I. First, unserialized (and occasionally illustrated) in a journal, second, unillustrated in a book, third, illustrated (some of the illustrations appearing in the journal) in a book, fourth, illustrated by a single artist, in another book (can I call it version or is it an edition?, for sure, it is not a translation). Perhaps if you start juggling illustrations, the consensus might change for doing things this way.--RaboKarbakian (talk) 02:25, 14 July 2021 (UTC)

@Azertus: No, we don't want this and I have already told RaboKarbakian that this is the wrong way to progress. Creating tailored items to fit a template at Commons is ludicrous, and back to front and does not fit our model. In my opinion it should be stopped and the items created should be merged into the editions.  — billinghurst sDrewth 13:30, 14 July 2021 (UTC)

Formal discussion about translation of (P9745)

@billinghurst, EncycloPetey, Thadguidry, PKM, Tinker Bell, UWashPrincipalCataloger:

I'm starting a formal discussion about what to do about translation of (P9745), a newly created property proposed by me, relating to the translations of creative works.

It was suggested that I should have started a discussion here about creating it before it was proposed. I am starting that discussion right now so that everyone can be on the same page.

Links:

I currently am supportive of the property. I've just read the Functional Requirements for Bibliographic Records (FRBR) (see simple description here) which the books data model is based off of. It actually uses the translation of property in order to distinguish translation relationships between "expressions" (see page 70).

I see this particularly of use in a discussion brought up above by @Wostr: called "has edition or translation and edition or translation of in the same item" about establishing a documented relationship between the edition, Nomenclature of Organic Chemistry. IUPAC Recommendations and Preferred Names 2013 (Q54019023), and the translated version of it, Nomenklatura związków organicznych. Rekomendacje IUPAC i nazwy preferowane 2013 (version 1) (Q55809949).

In this example, according to FRBR:

And it works! Clearly the property is of use if we follow FRBR! Of course, some documentation and constraints will need to be established to limit the usage of translation of to version, edition or translation (Q3331189) instances.

This, and clearly all of the countless other examples of translations can utilize this property and no cleanup is required because edition or translation of (P629) will remain to establish the expression -> work relationship,

Please let me know your feedback! Lectrician1 (talk) 01:06, 25 July 2021 (UTC)

Wikidata would be better served it was aligned with bibliographic properties used in the cataloging world, where translation is a very different relationship from revision. In FRBR and IFLA LRM, both revision and translation create new derivative expressions. A revised edition of a textbook is quite a different thing to a translation of the textbook, and I think Wikidata ought to have separate properties to describe these relationships. Resource Description and Access (Q1519318), the cataloging standard used predominantly in North America, the U.K., Germany, and other places, has a variety of these properties to differentiate types of relationships between expressions: is translation of; is free translation of; is dubbed version of; is revision of. is translation of "relates an expression to an expression whose language is modified to create a new expression that is different from another expression of the same work." is revision of "relates an expression to an expression that is updated, corrected, or expanded to create a new expression of the same work." UWashPrincipalCataloger (talk) 23:02, 28 July 2021 (UTC)

To test the suitability of the proposal: (1) How would you handle the editions and translations of History of the Literature of Ancient Greece by Karl Otfried Müller (Q62113)? The work was written in German and translated into English, but the English translation was published first, and included chapters that were not yet written in the German. When the German edition came out (after the English edition) it included chapters back-translated from the English for the parts that had not originally been written in German. This is not that unusual a situation. (2) When Johan Huizinga (Q276280) had his The Autumn of the Middle Ages (Q2599259) published in English, it was not translated from the Dutch, from from a French translation that was heavily reduced and rewritten from the original, but which was never actually published. So the first English translation comes from an unpublished manuscript translation. A second "edition" of the translation was published decades later, but it was translated from the original Dutch. (3) With Classical literature, bitexts are frequent, in which the pages of the books contain both an edition of the original text in parallel with a translation of that text, typically with the original language text on the even numbered pages and the translation on the odd-numbered pages. The edition of the original has a history and sources, and the translation has a history and sources. How would the proposal handle situations such as these three? --EncycloPetey (talk) 19:20, 29 July 2021 (UTC)
@EncycloPetey:
As FRBR states:
"Similarly, abridgements or enlargements of an existing text, or the addition of parts or an accompaniment to a musical composition are considered to be different expressions of the same work. Translations from one language to another, musical transcriptions and arrangements, and dubbed or subtitled versions of a film are also considered simply as different expressions of the same original work."
(1)
Therefore:
  • The original German text would have an item which would be an "expression" (edition of) of the main work.
  • The expanded English text item would be a translation of (P9745) the original German text item and would also be a edition of the main work.
  • The German edition item would be a translation of (P9745) the English text and would also be a edition of the main work.
Structure:
  • History of the Literature of Ancient Greece work
    • History of the Literature of Ancient Greece (original German text) expression
    • History of the Literature of Ancient Greece - English Edition expression; translation of (P9745) above
    • History of the Literature of Ancient Greece - German Edition expression; translation of (P9745) above
(2)
(3)
With respect to the last item, no, it contains a translation, it is not a translation, and it contains both an edition of the text and a translation of the text. Now that we have specifics, I will allow others to comment. We have to consider more than FRBR; we must also consider the structure and limitations of Wikidata's data structure. --EncycloPetey (talk) 00:43, 2 August 2021 (UTC)
@EncycloPetey: For bitexts, another option could be qualifying the translation of (P9745) statement with nature of statement (P5102) partially (Q100349848). Lectrician1 (talk) 15:52, 2 August 2021 (UTC)
That makes it sound as if the translation is partially from that source (and partially from some other source). --EncycloPetey (talk) 00:17, 3 August 2021 (UTC)

@billinghurst: Any thoughts on the following discussed? --Lectrician1 (talk) 15:54, 2 August 2021 (UTC)

@EncycloPetey, UWashPrincipalCataloger: I'm going to start using the property since there hasn't been any further comment on it in a week. Lectrician1 (talk) 03:26, 11 August 2021 (UTC)

Clearly I am as dumb as fuck. I don't see that it clearly works, or that you demonstrate clear lines from the conceptual items through to physical representations. I cannot follow your examples above. I don't see how this improves what is in place, it just makes things more complex, and resolves nothing. I am not seeing any benefit in what is set out above.  — billinghurst sDrewth 06:50, 11 August 2021 (UTC)

@billinghurst: Let me attempt to explain :). Instead of just listing out all of the translations and editions using has edition or translation (P747) and leaving humans or qualifiers like language of work or name (P407) to distinguish which items are translations of each other, we can establish an actual relationship between editions that helps computers know what editions are translations of another that are under the same work.
For example, Q43361#P747 (Harry Potter and the Philosopher's Stone (Q43361)) uses qualifiers on has edition or translation (P747) and language of work or name (P407) on the translations themselves to identify what language they are in.
However, there is no relationship that identifies what edition the translation was originally based on.
That is what translation of (P9745) seeks to do: establish a relationship between editions.
This usage might seem really basic at first since we already have language of work or name (P407), but the property becomes especially useful for works where there might be translations based on translations, like as we see with example 2 (The Autumn of the Middle Ages (Q2599259)) above. Lectrician1 (talk) 17:41, 12 August 2021 (UTC)

Manifestations and ISBNs

I noticed when reading Wikidata:WikiProject Books/Book data model that there is structure for FRBR works (work (Q386724)), expressions (version, edition or translation (Q3331189)), and items (individual copy of a book (Q53731850)), but not manifestations (the class that is between expression and item).

I also noticed that we have the ISBN-13 (P212), which is a property that is supposed to be used on manifestations.

An ISBN number is created for every format a edition is in, so are multiple ISBN-13 (P212)s supposed to be present on an edition item for every format it's distributed in? Wouldn't it make more sense to make individual manifestation items?

We should probably support manifestations and have a property that creates a relationship between them and their parent expression (edition).

Thoughts? Lectrician1 (talk) 18:08, 12 August 2021 (UTC)

  • Wikidata seems to lump expressions with manifestations, and I agree that this is somewhat problematic. I ran into this just yesterday when editing the item Georgius Agricola De Re Metallica (Q51483341). This is the item for a 1912 English translation of Agricola's De Re Metallica. That item combines expression and manifestation information. There is an LC authority for the English expression of the work, but when I added that identifier to the item I got an error message, because it also contains the LC control number for the 1912 manifestation. The constraint warning was "An entity should not have statements for both Library of Congress authority ID and Library of Congress Control Number (LCCN) (bibliographic)." I would create a separate item for the English translation as an expression (which could and actually does have multiple manifestations published in different years), but it's not clear how then to link from work to expression and then to the individual manifestations. UWashPrincipalCataloger (talk) 20:43, 12 August 2021 (UTC)
    • @Lectrician1, UWashPrincipalCataloger: This is complicated but mostly yes, Wikidata decide to have a 3 level model (like BIBFRAME) and not a 4 level model (like FRBR). It's not ideal - some people complain it's too simple, some that it's too complicate - but I think it's a good trade off. Plus, ISBN number are even more comlicated, sometimes wrong (and printed wrong) or mis-used and a lot of "exceptionnal" cases, like one single unique manifestation with several ISBN number (I just created Dix mille saints (Q108369181) yesterday which is in this not-so-rare case). Soeven with a manifestation class, you would still have items with multiple ISBN. Finally, you don't need the class, you can just use the property to model the relationship and we do have the properties expression of (P6524) and manifestation of (P1557) since 2014 :P ). Cheers, VIGNERON (talk) 06:25, 2 September 2021 (UTC)

Chain of editioning

I've tried to get long chains of A "has edition" B "has edition" С and so on, but all I got is a long list of cycles: 

SELECT distinct ?p1 ?p1Label ?p2 ?p3 ?p4 ?p5 ?p6 ?p7 WHERE {
  ?p1 wdt:P629 ?p2.
  ?p2 wdt:P629 ?p3.
  ?p3 wdt:P629 ?p4.
  ?p4 wdt:P629 ?p5.
  ?p5 wdt:P629 ?p6.
  OPTIONAL {?p6 wdt:P629 ?p7.}
    SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
Try it!

Can anybody help to tangle them? --Infovarius (talk) 13:02, 22 September 2021 (UTC)

Chain of editioning

I've tried to get long chains of A "has edition" B "has edition" С and so on, but all I got is a long list of cycles: 

SELECT distinct ?p1 ?p1Label ?p2 ?p3 ?p4 ?p5 ?p6 ?p7 WHERE {
  ?p1 wdt:P629 ?p2.
  ?p2 wdt:P629 ?p3.
  ?p3 wdt:P629 ?p4.
  ?p4 wdt:P629 ?p5.
  ?p5 wdt:P629 ?p6.
  OPTIONAL {?p6 wdt:P629 ?p7.}
    SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
Try it!

Can anybody help to untangle them? --Infovarius (talk) 13:02, 22 September 2021 (UTC)

News (Q38926) does not accept author, date of publication, "published in", "full work available at URL (P953)"

@MartinPoulter: On 2021-10-08 I created Tunisian President Appoints Prime Minister Amid Protests Over Power Grab (Q108837495). As I'm writing this, that Wikidata item reports lots of errors or infelicities. I don't know what I should do about these. In fact, my contributions to Wikidata include many things like this. A few such issues have been corrected by others. However, I don't know if your edits to this page impact these issues or not. I don't know where to turn to resolve such issues. So I ignore them. Thanks, DavidMCEddy (talk) 14:18, 8 October 2021 (UTC)

@DavidMCEddy: you just used the wrong value for instance of (P31). I corrected it (pro-tip: constaint violation si often based on the value of instance of (P31) ; if you don't know where to start checking, start there). This project page is to talk about book, this item is not a book ; this question would be more appropriate on Wikidata talk:WikiProject Periodicals (see Wikidata:WikiProject_Periodicals#Article_item_properties for propety to use on item about articles). Cheers, VIGNERON (talk) 08:47, 10 October 2021 (UTC)

Jonas manuscripts authority

Hi,

@MartinPoulter, PKM: I'm still planning an import of manuscripts and I was pointed to this website https://jonas.irht.cnrs.fr the project Jonas by the Institut de recherche et d'histoire des textes (Q1664768) of the National Center for Scientific Research, France (Q280413). I was thinking about asking for creating a property for this database but I'm not sure how to proceed.

This query https://w.wiki/4DGR give 12 current uses (in described at URL (P973)) but it covers 5 differents sub-database :

Should I ask for one or 5 properties ? the second would be cleaner I guess (only the last digit would be the "true" identifier) but the first could be easier (only one property to maintain, tool can discriminate afterward, like for the IMDb property IMDb ID (P345)). Or else, I could also only create property for the insterresting info (mainly "manuscript" here). In bonus, there is the equivalent Wikibase property on Biblissima: https://data.biblissima.fr/w/Property:P140 (and either way we choose, we could probably use it).

What do you think? Do you have similar example that I can look at?

Cdlt, VIGNERON en résidence (talk) 15:28, 13 October 2021 (UTC)

@MartinPoulter, VIGNERON: I would prefer separate properties for manuscripts, works, persons, etc. For examples see the Wikidata properties at:

- PKM (talk) 22:17, 14 October 2021 (UTC)

Invitation to WikidataCon workshop on bibliographic data

Hi WikiProject Books has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.!

I hope you're getting excited about WikidataCon 2021 even if it's going to be virtual this time around. I'll be joining you from Topsham, England where User:Sic19, curator of the WikiCite track, will be hosting me!

I was wondering who was thinking of joining the workshop on pre-ISBN/ISSN bibliographic data that I'll be leading with another database builder/bibliographer, Demian Katz? On Sunday, 14:30-15:30 UTC (10:30 Eastern North America, 15:30 Central Europe, check for other time zones), we'll be looking for fields to add to better capture the challenges posed by translations, series of works, and works serialized in installments.

We also see it as a way of kickstarting what we hope will be an ongoing conversation between data nerds of all stripes regarding books and periodicals. You can check out the design of the workshop and let me know if you think it's missing something. I'm also looking for someone who could present the various communication channels used by the Wikidata community. Apart from the WikiProject Talk pages, Simon mentioned Telegram channels, also pointing out more assiduous users might be more apt to talk about them, =P

Looking forward to meeting the most of you possible on Sunday! Marianika (talk) 09:21, 26 October 2021 (UTC)

Book subclass of written work?

Hi all,

I started a discussion on Talk:Q571#Subclass of written work?. Could you all take a look (I think we had this discussion many times but the talk page is strangely short, we had a lot of discussion around ittho, see above).

Cheers, VIGNERON (talk) 06:46, 19 October 2021 (UTC)

One answer, I went ahead. Now, there is the related question of Property talk:P31#None-of_constraint:_book?... Cheers, VIGNERON (talk) 16:31, 31 October 2021 (UTC)

Use of literature (Q8242)

Hi everyone,

I've started a discussion on literature (Q8242) about the use of this entity, which seems pretty messy to me. That would be cool if the contributors of the WikiProject Books could give their opinion on the matter.

WikiProject Books has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

CaLéValab (talk) 15:37, 19 November 2021 (UTC)

Where is the discussion? --EncycloPetey (talk) 18:17, 19 November 2021 (UTC)
Sorry, on the discussion page of literature (Q8242). CaLéValab (talk) 18:36, 19 November 2021 (UTC)

Separate genres for audiobooks?

Do we need a separate audiobook genre metaclass with corresponding objects? Some music streaming services use such categories: Deezer, or existing directories like MnM YandexMusic where only audiobooks are left unassigned. Or is it better to put all the identifiers in literature genres? It's also worth saying that a separate branch of audio recording genres is already in use for poetry: audio content genre (Q108676140), as this sort of thing is very common on catalogue music sites, along with other non-music stuff. Solidest (talk) 17:23, 24 September 2021 (UTC)

Only just now noticed that Deezer link only works for certain countries. You can see its content if you set Germany/Austria/The Netherlands here in your account settings: https://www.deezer.com/account/country_selector
Regarding the topic itself. I'm now sticking to the idea that everything should probably be labeled with audio content genre (Q108676140). Separately dividing the subtypes into fiction narrative and non-fiction. Solidest (talk) 18:45, 23 November 2021 (UTC)

Cataloguing manuscripts

I've overhauled the manuscripts section to reflect the work I've been doing describing manuscripts from the Khalili Collections and from Oxford University. I've mostly added properties to the table, but have removed publication date (P577). A physical object can't be published. The content of a manuscript can be published in (P1433) a book, journal or web site and the publication date (P577) should apply to that publication. We have cases where manuscripts have been given a publication date (P577) which should be a inception (P571).

There are some questions on which I welcome community input:

Pinging PKM who has long been doing great work in this area. MartinPoulter (talk) 17:13, 7 October 2021 (UTC)
  • Re: Manuscripts split between multiple institutions, I'd prefer separate items for each section because the ownership history, inventory number, etc., will be different. I'd suggest these be "part of" an item for the manuscript as a whole. PKM (talk) 20:54, 7 October 2021 (UTC)
Just what I hoped you'd say! MartinPoulter (talk) 10:26, 8 October 2021 (UTC)
  • Or instead of "part of" a master item, we might just use "exemplar of" an item for the work. (Speaking of which, I wonder if we need a property "has exemplar" for a work which exists in multiple manuscript copies. I've fuddled around with using editions for manuscripts when there are distinct textual differences between sets of exemplars, but I am not entirely happy with my modeling of these.) PKM (talk) 21:30, 7 October 2021 (UTC)
  • Re: Images - if the folio number is in the inventory number for the image's item, as you did here, I think that's sufficient? It's what I did here. PKM (talk) 20:59, 7 October 2021 (UTC)
  • Re: Images as <instance of> "painting" - I don't see a problem with this. I made this map P31 = "pen and ink drawing" as well as "manuscript map". PKM (talk) 21:01, 7 October 2021 (UTC)
  • Re: Folio. Agree a folio isn't necessarily hand-written. PKM (talk) 21:03, 7 October 2021 (UTC)
Great. I've removed that subclass of (P279) statement from folio and put an explanatory statement on its Talk page. MartinPoulter (talk) 10:26, 8 October 2021 (UTC)
Agree. This has allowed me to pare down the list in the instructions. I'll add genre as well. MartinPoulter (talk) 10:26, 8 October 2021 (UTC)
Thanks for clarification. It sounds like writing system is a property I'd only rarely need to add for Arabic and Persian manuscripts since it'll be implicit from the language, but it's nice to have when a language has multiple writing systems. MartinPoulter (talk) 10:26, 8 October 2021 (UTC)
Done! Good to have, MartinPoulter (talk) 10:26, 8 October 2021 (UTC)
  • Re: Transcripts - I'd also like community feedback on modeling multiple transcriptions of the same manuscript. There are several examples here. I have used parentheticals in the labels, but I hate parentheticals in labels on principle, so I am conflicted. - PKM (talk) 21:42, 7 October 2021 (UTC)
    While I get what you say about brackets, bracketed parts of labels seem like the best solution for this case where we have lots of things which go by the same name but which have important differences which we need to be able to see at a glance. So long as what is represented in the brackets is also expressed in properties, I think this is good practice. MartinPoulter (talk) 10:31, 8 October 2021 (UTC)
Thanks a lot MartinPoulter. This is a lot to think about. First I realise that I wrongly use publication date (P577) on my last import, I totally agree « A physical object can't be published. »
Then you seems to focus only on manuscript but I guess we could benefit to extend it to most (if not all) non printed textual document. An obvious example is papyrus, scroll, codex, and maybe also incunable (as a transition between printed and non-printed). Not sure exactly how to deal with that, maybe we should focus on manuscript first and then see how it extend to other documents. We should also not confuse the instance of (P31) and the genre (P136) (as said by @PKM: some example obviously belong to the second).
For writing system (P282), the value is very general likely to already be on the value of language of work or name (P407). So I would never put it on the manuscript item itself.
For the part and relationship between parts, I need to think about it more and I don't have strong opinion (not yet at least), the most important thing is to have a link, whatever the property used.
For the images, maybe we could/should look at things like IIIF... and we definitely need to look at SDC and Wikimedia Commons Wikibase !
Cheers, VIGNERON en résidence (talk) 13:58, 8 October 2021 (UTC)
Very grateful for this input User:VIGNERON en résidence. Looks like we are all thinking a similar way. I agree that IIIF manifest URL (P6108) could be useful: I'll include that. I agree that incunable (Q216665) will have very similar properties to manuscript, but here I'm just focused on manuscripts. When an object in a collection is described as "a papyrus" I treat that as instance of (P31) manuscript (Q87167), made from material (P186) papyrus (Q125576). papyrus (Q125576) is the material, so while Wikidata has some items instance of (P31) papyrus (Q125576) I don't think they should do that. Cheers, MartinPoulter (talk) 13:39, 12 October 2021 (UTC)
@MartinPoulter: you're welcome. Of course "papyrus" should go as material but isn't it *also* the instance of concept? and can't be some pyapri not be made out of papyrus? (probably not but I wouldn't be surprise to stumble on some strange edge cases). That also raise the question for scroll/codex, should it go in the property distribution format (P437) and/or in instance of (P31) and maybe also for "manuscript" itself that could go on an other more specific property... We should probably look at what other databases do and what standards says (to avoid going on a wrong path on help interoperability in the futur), I'll try to ask my colleague for their input. Cheers, VIGNERON en résidence (talk) 15:06, 13 October 2021 (UTC)
list of ancient Egyptian papyri (Q638887) and list of New Testament papyri (Q1417837) list many examples where papyrus seems to be treated as instance of not just material (quite a few languages add the word manuscripts to list of New Testament uncials (Q1419177) though). Looking at list of papyrus collections (Q1863905), papyrus collections do not seem to be subsections of manuscript departments in libraries, at least not regularly.
Indeed, interoperability with e. g. Fragmentarium (Q40889231) for fragments would be nice to have. In Germany https://handschriftenportal.de/ will be the database to watch, starting to replace Manuscripta Mediaevalia (Q1891874) next year. HHill (talk) 21:07, 13 October 2021 (UTC)
The german cataloguing guidelines are linked at https://en.handschriftenzentren.de/materialien/?noredirect=en_GB, further down on the page are links to a selection of manuscript databases. HHill (talk) 18:28, 15 October 2021 (UTC) linkfix --HHill (talk) 03:47, 16 October 2021 (UTC)
@MartinPoulter, PKM: follow-up: we just created Clermont-Ferrand, Bibliothèque du Patrimoine. MS 84 (Q109602526) (to be used on the Artwokr template on the Commons file), all comments and remarks are obviously welcome and I have a specific question about the statements on Clermont-Ferrand, Bibliothèque du Patrimoine. MS 84 (Q109602526)genre (P136)book of hours (Q727715). IIRC, this is what we agreed on and what is written on Wikidata:WikiProject_Books#Manuscript_properties but it triggers a constraint violation. Should we fix the item, use an other value or change the constraint? (or something else). Cheers, VIGNERON en résidence (talk) 10:31, 19 November 2021 (UTC)
I am not sure why <genre> "book of hours" is causing a constraint violation. book of hours (Q727715) is <subclass of> devotional literature (Q90257546) which is both <instance of> "literary genre" and <subclass of> "religious literature" (which is also a literary genre). The subclass tree ought to be valid here. - PKM (talk) 22:42, 19 November 2021 (UTC)

poetry (Q482) vs poem (Q5185279)

Hi everyone,

I'm confused about poetry (Q482) and poem (Q5185279). I understand that the first one is a literary form, while the other one is the name given to a work with that form. Having both items would be like having two items for "novel", one to refer to the literary form, and the other to refer to a work with that form. We say "Lord of The Ring is a novel", and "Lord of the Ring belongs to the novel genre", while we say "Demain, dès l'aube" is a poem, but "Demain, dès l'aube" belongs to the poetry genre. I'm not against keeping both, since lots of languages seem to have two different words to express "poetry" and "a poem" (contrary to "novel" for example), but shouldn't we clarify a little bit more that one is a form, and should be used with form of creative work (P7937), while the other is a class, and should be used with instance of (P31) ? Am I missing something ? CaLéValab (talk) 22:40, 13 November 2021 (UTC)

@VIGNERON, EncycloPetey, MartinPoulter: I'm pinging some contributors that look active one this page, in hope to have an answer for this one. The poetry ontology is very messy, and I would like to correct it. But before that, I would like to be sure I make no mistake, which starts by making sure I understood subtleties like the one explained on my previous message. CaLéValab (talk) 12:53, 15 November 2021 (UTC)
The difference between poetry (Q482) and poem (Q5185279) is a difference between a style and a specific object. It is more like the difference between drama and play, but the WD items for that set of labels is also thoroughly muddled.
Consider that The Lord of the Rings contains poetry (Q482), but it is not a poem (Q5185279). The book Psalms (Q41064) is poetry (Q482), but it is not a poem (Q5185279). Most classical Greek dramas are written as poetry (Q482), but they are plays / dramas, not poem (Q5185279). There is a distinction in what these words mean. --EncycloPetey (talk) 15:56, 15 November 2021 (UTC)
poetry (Q482) is the counterpart of prose (Q676), and neither of these is the form for a work, but is a compositional style. Novels may be written in either compositional style, Epics can be written in either style, Plays can be written in either style, and even Poems can be written in either style. Neither poetry (Q482) nor prose (Q676) is a form, but both are a style of composition that can be applied to many different works, and the same work might be translated from one form to another, as I have seen translations of Greek works in verse (both plays and poems) translated into English prose. --EncycloPetey (talk) 16:01, 15 November 2021 (UTC)
Thank you EncycloPetey. So poetry (Q482) is a genre, not a form, right ? Or is it even something else, like a tone (Q1366327) ? Should it be used with the property genre (P136) or another property ? CaLéValab (talk) 16:25, 15 November 2021 (UTC)
No, poetry (Q482) is neither a genre nor a form. It is a compositional style independent of both form and genre. Any form or genre could be constructed in poetry (Q482). I do not know whether there is a property applicable, but note that a single work might be written in poetry, or prose, or both, and different editions of the same work might change the compositional style from poetry to prose, or vice versa. --EncycloPetey (talk) 23:49, 15 November 2021 (UTC)

Parallel question: what about using either of these as P31? Can we say that poetry (Q482) subclass of (P279) poem (Q5185279)? --Infovarius (talk) 20:30, 16 November 2021 (UTC)

No, poetry (Q482) is definitely NOT subclass of (P279) poem (Q5185279). ANY form or genre of work can be written as poetry (Q482), whether a play, an opera, an essay, a short story, a poem, etc. Style of composition is a separate factor from any form or genre. --EncycloPetey (talk) 22:24, 16 November 2021 (UTC)
And no, poetry (Q482) should not be used with instance of (P31), since an individual work can exist in editions that are composed as poetry (Q482) and the same work can exist in editions that are not.
Additionally, a work written in poetry (Q482) might be a literary work (Q7725634), a composed musical work (Q207628), or a play (Q25379) (dramatic work). These are not subclasses of poetry (Q482), nor is poetry (Q482) a subclass of those kinds of works. Compositional style is independent of both form and genre, and does not apply to any particular group of works, but can apply to all of them. --EncycloPetey (talk) 22:29, 16 November 2021 (UTC)
Sorry, I meant poem (Q5185279) subclass of (P279) poetry (Q482) which was true before CaLeValab's revolution. Also not true that poetry (Q482) has no subclasses even now, see e.g. narrative poetry (Q474090). --Infovarius (talk) 13:25, 19 November 2021 (UTC)
@Infovarius: not that either. Since poem (Q5185279) is a literary form (Q4263830) and poetry (Q482) could apply to something that is not literature. And thank you for calling me a revolutionary. As a French person, this is quite a compliment ;). CaLéValab (talk) 14:14, 19 November 2021 (UTC)
@Infovarius: We don't have novel (Q8261) subclass of (P279) prose (Q676), so why should we have poem (Q5185279) subclass of (P279) poetry (Q482)? Both prose (Q676) and poetry (Q482) are compositional styles that can be applied to ANY form of literature. As far as narrative poetry (Q474090), I agree that there is a lot of confusion, and a lot of cleanup is needed. narrative poetry (Q474090) is an intersection of the "poetry" compositional style with "narrative" intent. An essay can also be narrative, and so can a short story. "Narrative" is a storytelling style that can be used in many forms of literature. --EncycloPetey (talk) 18:15, 19 November 2021 (UTC)

Hi @Solidest:, I still try to understand what is poetry (I thought I understood, but after your comment, it seems like I didn't). Do you define poetry like I did, which is "literary style characterized by a strong expressiveness of words" (the description on the entity was added by me, and it seems like the wikipedia articles attached to the entity also define poetry this way). In which case, isn't poetry instance of literary style (Q109598321) better than poetry instance of literary form (Q4263830) ? CaLéValab (talk) 15:50, 19 November 2021 (UTC)

I was trying to find the proper metaclass for poetry (Q482). I couldn't find anything existing on WD in English, but instead I found pretty close meaning in Russian that covers it. It is "wordness" (Q4423833). It groups any types of creative works, as "object of thought" conveyed then through words. That is, it combines all written works (literature (Q8242), letters, diaries, etc), all oral works (rhetoric (Q81009)), and probably something in between, like folklore (Q36192). I'm not sure that there is a proper word for this term in English. As for metaclasses of that, I guess the closest one to use with it would be "narrative form" (at least in Russian). But "narrative form (Q6630149)" (looking at the linked categories and list) in English seems to be used in a different way - it's more of a narrowly literary category that partially duplicates the literary form (Q4263830). So I am no longer sure if I have changed that item (Q6630149) correctly.
I don't really like the term "literary style" because it essentially refers to the realm of literature and by and large to some light "descriptive forms" (styles/methods) within specific major literary forms by individual authors. I'm not sure if it's any different from writing style (Q2243831), as the only French article describes exactly that. So, as far as I see it, it belongs to the realm of narrative text, whereas poetry rather encompasses the realm of conveying thought through words, and it would be more appropriately described as "form" rather than "style". Solidest (talk) 16:23, 19 November 2021 (UTC)
I think of poetry (Q482) and prose (Q676) the same way that musicians think about instrumentation. A musical work can be "scored" or "arranged" for orchestra, or choir, or solo vocals with accompaniment. It is still the same musical work, but the manner is which it is presented has changed. The same can be true of writing: it may be composed in poetry or in prose, or a combination of the two. --EncycloPetey (talk) 18:15, 19 November 2021 (UTC)
You can probably play the same musical work/composition (Q105543609) with different instrumentation (P870). The musical composition (as the 'intellectual work') is still the same, but when it's an actual interpretation, work played with some specific instrumentation - the audio recording (Q3302947) is used if it's a record (Q107435521) or musical performance (Q6942562) if it's occurrence (Q1190554). I'm not sure that such an approach can be used in literature. You can't make poetry or prose from the same set of words (written or spoken). Poetry/prose is more than just a method of putting it down on paper, or a way of telling the same words. Solidest (talk) 18:51, 19 November 2021 (UTC)
You are forgetting translation. The same work can be written in different words, and those words may be prose or poetry. On en.WS we have translations of Pindar's Victory Odes (Q1347171) that include both poetic translations and prose translations. Whether they are prose or poetry, they are the same works, just in a different translation. And as for your other question, yes, the FRBR includes levels for specific editions, print runs, or even individual copies held in specific collections. --EncycloPetey (talk) 05:36, 20 November 2021 (UTC)
Yeah, I really forgot about that. So ideally we need a separate property like instrumentation (P870), something like "style of wording" or as you said "compositional style"? And while we don't have one, where should we put the poetry/prose? form of creative work (P7937) seems most appropriate, or at most has characteristic (P1552) (where in the music model we put everything excessive that has no its own property). Solidest (talk) 14:55, 20 November 2021 (UTC)
Hello to both of you. I'm not sure "poetry" could be reduced to "opposite of prose". We have poetic prose (Q15890340), which would be oxymoronic. From what I understood from the Wikipedia articles attached to "poetry", "poetry" is defined as a way of using words to make them mean more than what they mean in a prosaic way. "Poetry" would then be the opposite of "prosaic style" but not "prose". So, like EncycloPetey said, it is neither a genre nor a form, but more like a "style" i.e. "a way to use words to express meaning". So I'm not against creating another property for "style", where we could put "poetry", "formal", "informal", "vulgar", etc. CaLéValab (talk) 15:15, 20 November 2021 (UTC)
  • I tried to correlate the discussed properties for related areas of creative/art spheres and built such a table:
Sphere / Art Genre class Type class Form class
Music music genre (Q188451) "music by" metacategory musical form (Q862597)
musical work/composition (Q105543609)
type of musical work/composition (Q107487333)
  • song
  • sonata, ...
 song type (Q107356781) (for now it's a general category of different species that has yet to be divided) song form (Q1824109)
"Word-based art form" (Narration?) ? ? narrative form (Q6630149) ?

(poetry / prose)

"word-based work" / "narrated work"
?
  • folklore
  • "genre of folklore"
  • "folklore by country", ...
 ?
"Word-based written art form"

(Literature)

literary genre (Q223393) sub-set of literature (Q109551565) literary form (Q4263830)
literary work (Q7725634)
type of literary work (Q110169000)
  • novel
  • poem, ...
  • "novel by country", ...
 ?
"Word-based spoken art form"

(Rhetoric / Oral art)

? ? ?
"spoken entry"
?
  • speech
  • monologue
  • rap, ...
 ? ?
?
This is based on the most detailed model that we have on WD for creative arts = music. This would also be valid for film, dance, theatre. For many spheres this kind of model is already in the use, with slight deviations and many blank spots. I'm probably missing something, but it seems to me that in general we should workd to build such a correspondence between the spheres of creative works on these parameters, and we should develop the spheres like this (with a room for improvement). And speaking of poetry/prose, this seems to be the most undeveloped area at the moment. It also seems to me that introducing "Style" would only complicate things and add unnecessary overlap, because most of the time it is something completely subjective and when it's something more established, it's rather more properly to be called a genre. But "theme" would probably be a more useful addition. Such a thing is already used in the video game model: video game theme (Q42907216). Solidest (talk) 16:41, 20 November 2021 (UTC)
I'm completing your table with new items CaLéValab (talk) 02:11, 18 December 2021 (UTC)
I finally started Wikidata:WikiProject Arts and moved this table there. Feel free to join and expand it with existing items or with any future plans. Solidest (talk) 04:09, 18 December 2021 (UTC)

"Canonical" corpora

WikiProject Books has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

Hi all,

I'm asking for advice following a short discussion on Epìdosis' talk page, as the case isn't covered in the current Project's recommendations. I am currently working on the various works that are included in Moralia (Q651941), a collection of 70ish treaties on various subjects written by Plutarch (Q41523). The collection was assembled across the centuries and is found as such in manuscripts; in all probability however Plutarch didn't considered that his treaties were part of a unique collection. As Corpus Aristotelicum (Q1054239) or Hippocratic Corpus (Q1135155), Moralia (Q651941) is the result of the textual transmission way more than its author's will. Currently, the relation between a treaty and the collection is modelled with published in (P1433), which considering the way that the collection was assembled strikes me as misleading. I would personnaly gladly replace it with part of (P361), but perhaps other properties might be a better fit. --Jahl de Vautban (talk) 09:20, 5 December 2021 (UTC)

  • I agree that part of (P361) fits more than published in (P1433) in this case. Every work included by tradition in a canonical set (such as Corpus Aristotelicum (Q1054239) or Hippocratic Corpus (Q1135155)) is part of it. In my opinion, this case is very similar to that of a book (e.g. Book of Genesis (Q9184)) of the different biblical canons (such as the Tanakh (Q83367) or the Bible (Q1845)) and each book of the Bible is actually recorded in Wikidata as part of a section of the Bible (even if the exact section depends on the different canon to be considered and many books of the Bible are currently not correctly related). I would use published in (P1433) only for a work by an author published or to be published in a miscellaneous volume (such as conference papers, manuals with chapters by different authors, or papers in a journal). A different case is an anthology of texts collected by an editor. I think that it should be modelled with a work representing the anthology whose responsibility is related to the editor and to which the selected texts (e.g. one or more Plutarco's treaties, but not all the treaties of its canon and other texts by different authors) should be related by part of (P361). The item representig the antology at the work level should have also a has edition or translation (P747) for any of its editions. In the last case I feel that, instead of part of (P361), a property to express that the work is not natively or traditionally part of a canon, but that it was added to a newly created set by the explicit willingness of a specific person (the editor) would fit better. I do not know if it already exists. --Carlobia (talk) 14:19, 7 December 2021 (UTC)
  • I agree with @Carlobia:, the reasoning makes very much sense. --Epìdosis 19:22, 9 December 2021 (UTC)
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata_talk:WikiProject_Books/2021&oldid=1661309322"