JanZerebecki added a subscriber: JanZerebecki.
JanZerebecki added a comment.

I think Lydia actually answered your question. "United States" its "history of 
topic" (`P2184`)  is "History of the United States". But "History of the United 
States" is "facet of" "History" and "United States". "facet of" (`P1269`) is a 
"subproperty of" (`P1647`) "part of" (`P361`). I.e. everything that is a facet 
of X is also a part of X. "history of topic" is "subproperty of" "has part". 
"has part" is "inverse of" "part of".

For more links of properties to other properties: 
https://www.wikidata.org/w/index.php?title=Special%3AWhatLinksHere&target=Property%3AP361&namespac
 (following this recursively might be useful). Though that is likely incomplete 
further linking would be helpful.

Maybe Wikidata encodes the semantics you are asking for. Maybe it does not, 
then maybe your question could lead to a useful 
https://www.wikidata.org/wiki/Wikidata:Property_proposal . I would be 
interested in the outcome of your endeavor and also in the outcome of the 
mentioned research.

---

I not sure which of the following you already know or if it is useful so beware.

> main/sub-article relationships


This definition is not precise enough to say that "History of the United 
States" is not a "sub-article" of "History". Searching for "Main articles: " 
below a heading might be one definition, but even if that information was 
stored in Wikidata (it is not) it would be less reliable because these links 
are not generated from Wikidata.

> In fact, it once was!


That would be a more precise definition, but identity over time is a tricky 
concept, even more so for digital text in Mediawiki pages. Mediawiki records 
inserts and deletes, but no moves. It can not tie together modification on two 
articles as one changeset, so git style move detection is not possible. Anyway 
nothing Wikidata can help you with. You might use the history of changes in 
Wikidata to find the subdivision of concepts and the move of sitelinks from 
item to item, but that is not semantically the same thing.

> relationships between concepts cross-language where one language Wikipedia 
> would split a topic into multiple articles and another would only have a 
> single article


This portrayal of interwiki link conflicts is too simplistic. One encyclopedia 
can be sliced in arbitrary ways. Even saying slice is too simplistic as 
anything can be duplicated. I.e. there can be overlap. Even 
https://en.wikipedia.org/wiki/Venn_diagram#Extensions_to_higher_numbers_of_sets 
is probably too simple to capture all semantic relations between articles. Most 
human recognized concepts can be further divided into useful parts ( 
https://www.wikidata.org/wiki/Property:P361 ). Also there are errors, 
confusion, imprecision and so on, both in articles and on Wikidata.

https://www.wikidata.org/wiki/Wikidata:Interwiki_conflicts
https://meta.wikimedia.org/wiki/Interwiki_conflicts

One possible response to an interwiki conflict is subdividing and grouping the 
concepts on Wikidata so that each article can be linked to the correct concept. 
Another is changing articles. Any combination may be necessary.

> I'm working with a research team at the University of Minnesota that is 
> developing a strategy for robustly identifying main/sub-article relationships.


One correct answer is: everything is connected, the usual strategy for 
identifying "anything" is as robust as it gets. ;)


TASK DETAIL
  https://phabricator.wikimedia.org/T117875

EMAIL PREFERENCES
  https://phabricator.wikimedia.org/settings/panel/emailpreferences/

To: Lydia_Pintscher, JanZerebecki
Cc: JanZerebecki, Lydia_Pintscher, Aklapper, StudiesWorld, aude, Halfak, 
Wikidata-bugs, Mbch331



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

Reply via email to