Consistency and well foundedness. It's acutally pretty confortable that the basic object we're classyfing are concrete stuffs. The diary of Ann Franck as a Work is more an abstract object. This ensure that we always indeed are trying to class concrete object at the base of the classification system. Classes and higher order classes are then always at some level : one for classes as they are collections of concrete objects or events, two for car models or chemical elements as they are classes of classes of objects or events, and so on.

This avoids to have some questions to be asked about "will my class will have instance" "oh shit, in fact we have articles on the concrete objects after all, what will we do ?????"

Let's do be right the first time on the basic principle we uses. https://www.wikidata.org/wiki/Help:Classification simplifies overall things because we use a regular and consistent scheme across the vasts domain of application Wikidata is a database for. Read the "Metaclass" article, follow the citations of the sources of it, you'll see that this is a widely accepted scheme actually. For a reason(s) :)

2015-10-19 19:58 GMT+02:00 Stas Malyshev <smalyshev@wikimedia.org>:
Hi!

> Similarly the "Diary of Anne Frank" is an instance of a memoir or a
> literary work but is a subclass of book (because there are lots of
> physical books with that name). Literary works have authors and
> publishers. Books have numbers of pages and printers and physical locations.

I'm not sure I understand this. What is the difference between "instance
of memoir" and "subclass of book"? You could literally argue with the
same words that it is also "subclass of memoir" and again since very
rarely any specific physical book is notable enough (maybe excluding
things like first Gutenberg Bible, etc.) we would have virtually no
instances of book at all. I do not think people think that way - if you
ask somebody is "Diary of Anne Frank" an example of a book or a class I
think most people would say it's an example of a book and not a class.
Unless we plan to seek out and record every printed physical copy of
that book, I don't see any practical reason to describe it as a class.
This class - and hundreds of thousands of other book titles, maybe with
rare exceptions of the Gutenberg Bible, etc. - would never have any
instances. So my question is - what is the use of modeling something as
a class if there won't be ever any instances of the class modeled?

--
Stas Malyshev
smalyshev@wikimedia.org

_______________________________________________
Wikidata mailing list
Wikidata@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata