Wikidata talk:Requests for comment/Improved model for document description

Genre, just physical books?

The word "genre" is often limited to writing that has value as art, such as works of fiction. It isn't always considered to include utilitarian non-fiction, such as scientific papers. Everything about this proposal seems to be oriented to physical books written on a medium such as paper. The proposal ignores electronic documents. But today the majority of documents are electronic, and many are only available in electronic form. One of the big uses of documents in the Wikipedia web sites is citing sources. Having two incompatible models of sources, one for electronic sources, and a different model for physical sources, does not seem desirable. Jc3s5h (talk) 14:47, 14 August 2019 (UTC)

@Jc3s5h: Nothing prevent you to be more constructive by explaining how the proposition can be changed: genre and form are not affected by the electronic format of a document, the support characteristic can be extended to "electronic" and the two last characteristics type writing and material can be set up to no value. So unless you have something to propose, it is difficult for me to see where this model is a problem to handle electronic documents.

And genre (P136) can be used to characterize more than fictional text: essay (Q35760), autobiography (Q4184),... Snipre (talk) 15:31, 14 August 2019 (UTC)

Edit conflict: I was transfering your comment and adding my answer at the same time, so sorry, I didn't want to delete your comment. Snipre (talk) 15:31, 14 August 2019 (UTC)

I see that you have, in the past, posted at Wikidata talk:WikiProject Books. I also am aware of a data model which seems to occupy most of the text at Wikidata:WikiProject Books. Could you describe how this proposal relates to the model at Wikidata:WikiProject Books? Jc3s5h (talk) 16:39, 14 August 2019 (UTC)

@Jc3s5h: This RfC aimed to find a solution for the description of a document and particularly of books. Currently 2 ways can be used to describe a book: instance of (P31) and genre (P136). genre (P136) can't be used to define all different characteristics of a text (characteristics about the form and about the content) and instance of (P31) is using mixed concepts without any systematic structure (we can have instance of written work (no information about the content or the form), or instance of novel (info about the form but not about the content) or instance of burlesque poetry (info about content and form))

I hope this RfC can improve the current data model presented in Wikidata:WikiProject Books by a better definition of the use of genre (P136), by reducing the diversity of values for instance of (P31) and merging the models for particular copy of book (we have currently two models for printed copies and for manuscripts). Snipre (talk) 20:56, 14 August 2019 (UTC)

Only works of fiction will have a specific genre. Works of non-fiction can be said to have a genre of nonfiction (Q213051), but really will need main subject (P921). The biography (Q36279) and its subgenre autobiography (Q4184) are among the few exceptions where non-fiction has a genre. --EncycloPetey (talk) 15:14, 15 August 2019 (UTC)

@EncycloPetey: What's about biography ? This is a genre or what else ? For sure not a format. And about essay, historical book,... ? And a travel guide about South America ? For this last example I would use genre to describe travel guide and subject to describe South America. Snipre (talk) 08:28, 20 August 2019 (UTC)

For example: here. Snipre (talk) 08:48, 20 August 2019 (UTC)

An essay might be considered a genre. I did say that biography was one of the few exceptions, though not the only one. --EncycloPetey (talk) 17:52, 20 August 2019 (UTC)

@Jc3s5h: No, "genre" is not even limited to narrative, it's also used to describe films, comics, videogames, ... In fact, all sort of mediums that support a fictional work can be described in terms of genres (western, noir, horror, etc), so maybe "fictional" is the key here, and we should characterize them as "fiction genres" and keep apart from other uses of "genre". --JavierCantero (talk) 09:00, 20 August 2019 (UTC)

Why this venue?

We don't use RfC's for creating new properties but generally create them in property proposal discussion. We even have rules that specify the necessity of the property proposal discussion. When it comes to discussing the datamodel it seems to me like the Wikiproject Source Metadata or the books project would be the more fitting place. ChristianKl ❪✉❫ 18:29, 14 August 2019 (UTC)

Perhaps did you read the text in discussion: the first idea of this RfC is to discuss about a data model. A property proposal doesn't allow the possibility to discuss the definition of an organized set of properties: some properties can be accepted some can be rejected and at the end missing one property can generate a bunch of useless properties.

Then a discussion in a project talk's page attracts in general 5-7 persons, a RfC can attract more people (or not, this a bet) especially people which are not used to take part to project discussion. This is important because the topic of this RfC was already discussed several times in the books project and we restart again and again the discussion so having new persons can help to reach a conclusion.

RfC conclusions can be better accepted than a simple discussion between 3-4 experts in a project discussion. especially when you have to revert again and again some edits. Snipre (talk) 20:32, 14 August 2019 (UTC)

I am not sure anyone but the same "3-4 experts" (closer to 10 in regards to books, I think) would comment on your RfC. This sort of discussion is exactly what projects are for. I do think the SourcMD/WikiCite bibliography experts are not typically active in the Books project, and moreover have most of their discussions in person or otherwise off-Wiki. - PKM (talk) 19:10, 15 August 2019 (UTC)

@PKM: RfC offers an better framework to take a decision than a discussion in a talk's page which will be archived and lost. So even if we have the usual experts, if we can reach a decision, the RfC will provide a clear starting point to modify a large set of items.

A RfC is a community call, so even if a imited number of experts is providing their comments, the decision can be considered as being a community decision. Snipre (talk) 20:24, 15 August 2019 (UTC)

Font?

We have typeface/font used (P2739), maybe that should be added to your list of descriptive properties? ArthurPSmith (talk) 15:34, 15 August 2019 (UTC)

Possibly, but that would apply only to specific editions (or individual copies), and will not apply to a work as a whole. --EncycloPetey (talk) 21:07, 15 August 2019 (UTC)

Actually I've read a few books where the author was very particular to mention their use of a certain font. Not that that might be respected in all editions of the work, but there may be a specific author intent associated with that. In any case, I think we are talking here about properties associated with specific editions (too). ArthurPSmith (talk) 13:48, 16 August 2019 (UTC)

Ok to add it for edition and manuscript (case where only one exemplar exists). But as optional and not mandatory properties. Snipre (talk) 08:24, 20 August 2019 (UTC)

Support

The proposed property "support" is pretty well covered for digital and physical formats by distribution format (P437) which includes values like softcover, quarto, etc. as well as CD-ROM, ebook, and so on. We might want to expand it to include "scroll" for historical works but really the rest of it is there. - PKM (talk) 19:05, 15 August 2019 (UTC)

Would palimpsest (Q274076) apply here as well? --EncycloPetey (talk) 21:08, 15 August 2019 (UTC)

@EncycloPetey: This would imply that the property distribution format (P437) will have several values. Snipre (talk) 08:35, 20 August 2019 (UTC)

If we can get a consensus to extend the use of that property, why not. But I am a little afraid that people working on manuscript will have a problem with distribution concept. But this is a supposition. Snipre (talk) 08:31, 20 August 2019 (UTC)