> A bug was reopened to "allow collation to be
specified per category"
> in Wiktionary (all languages are concerned). This would be a much
> better solution than to create sort keys in every article and for
> every language.
I'm not sure I understand what the point of
this feature [is], but
couldn't it be implemented through relation set in wikidata ?
To
be clear, collation is simply the way words are sorted and displayed in
categories. By default Mediawiki uses the Unicode default order, which
is suited for no language.
There is already a way to use a default
collation on a given project, but not for different categories in one
project. That's what this feature is about.
In short, this is a technical issue and not a matter of data sharing, so Wikidata should
not be used for this.
Please consider looking at [1] and if relevant,
add your suggestion, so
we have a wider view of what we need for wiktionnaries.
This
page is about a faraway future (unless I am mistaken). As you know, all
Wiktionaries are quite different right now and sharing anything semantic
or linguistic with Wikidata is probably out of question in the short
term. For now I do think we can use Wikidata for some shared data that
would be universal (not dependent of the community language and
conventions). Some examples :
- Translitterations & transcriptions. Although Lua may be best suited for this.
-
"See also" links at the top of typographically similar articles (mere,
mère, Mère...), cf {{see also}} on en.wikt, {{voir}} on fr.wikt, {{Siehe
auch}} on de.wikt, etc.
I started a (/yet another) discussion on
https://en.wiktionary.org/wiki/Wiktionary:Beer_parlour/2013/March#Wiktionar…
about this. As you can see I actually proposed to use Wikidata for
collation as you suggested, but CodeCat reopened the bug which is a much
better solution for this.
There may be other "small" things
that can be shared with Wikidata. But if we want to use it for actual
linguistic data (pronunciations, meanings, etymologies...) then we'll
probably have to rebuild the structure of all Wiktionaries. Also we
would need something more suited for structured data than the current
plain wikitext (it is already quite painful to edit). That is certainly
not straightforward.
Darkdadaah