Hi folks!
While reading the thread:
Use case: generation of short description
https://lists.wikimedia.org/pipermail/abstract-wikipedia/2020-July/000012.h…
I wondered about the reverse use case... generating long names if short
names or abbreviations are found in wiki text (sometimes very hard to
disambiguate and sometimes easy based on signals). Imagine a renderer that
was semi-smart enough to expand abbreviations for *demonym's* into their
full long form.
Example: *Brit* is the short name for *British* in American English and a
few other languages actually.
Would the use case potentially be fulfilled with a smart renderer that
could use the statement short name P1813
<https://www.wikidata.org/wiki/Property:P1813> on a form applied on a
Lexeme itself as I did on
https://www.wikidata.org/wiki/Lexeme:L156273 ?
or on the Item (location) itself as I did for *Brit* on United Kingdom Q145
<https://www.wikidata.org/wiki/Q145> ?
I ask because I am thinking of how to help developers in the future where
they code a smart renderer that best gets the signals it needs for this use
case. Once I know where to add those signals in a best practice, I plan to
populate the short name demonym's. (my hunch is that the Lexeme space is
probably the best, rather than the Item space and using P1549:P1813 as I
did on United Kingdom Q145 <https://www.wikidata.org/wiki/Q145> or both
Lexemes and Items for good measure, but it's not DRY ?)
In general, looking for best practices for dealing with abbreviations and
expanding them contextually. (yeah, rocket science in some respects, but
signals might help)
Thad
https://www.linkedin.com/in/thadguidry/