>>>Pete Johnston said:
> I said (regarding representation of xml:lang in RDF)
>
> > This appears to be a current issue for the RDF Core WG. See
> >
> > http://www.w3.org/2000/03/rdf-tracking/#rdfms-xmllang
>
> Dave Beckett said:
>
> > Yes, if the DC community has some feedback, let me/us know. We (RDF
> > Core WG) are likely to try to get this resolved soon.
and the delay in replying is because I've been away with other RDF
Core folk discussing this issue amongst others :)
> Could you provide any pointers to what is the current thinking within
> the RDF Core WG to how the language issue should be best addressed
> please? I'm afraid I found it rather difficult from scanning the RDF
> Core WG mailing list to get a clear picture of exactly what was being
> suggested. Or is it still under debate?
There is no public pointer to the current thinking yet; since at the
time you wrote this email (Wed), we had just discussed it.
<snip/>
> Is my reading of that correct? Will the RDF Core WG be making a
> recommendation suggesting a different approach please?
We are working on our recommendation, after discussion with the W3C
Internationalization (i18n) interest group as to the best approaches.
The approach is something like this (unofficial, not approved by WG yet)
In RDF, literals are pairs of unicode strings and a language tag -
any legal values of the xml:lang attribute. (They also record
whether the literal was a XML structure too; omitting that here for
simplicity)
So these are some RDF literals:
"chat"
"chat" with xml:lang="en"
"chat" with xml:lang="fr"
We've been using the term lang-strings to discuss these.
giving the pairs:
("chat")
("chat", "en")
("chat", "fr")
We discussed with I18N WG our requirement for
a mathematically rigorous and transitive equality.
and agreed to define it as:
1) exact matching of strings
and
2) exact (case insensitive) matching of lang-tags [when both present]
HOWEVER
NOTE: This definition of equality is appropriate when constructing
an RDF graph, when checking an RDF test case, and when
interpreting an RDF graph according to the RDF model theory. In
other contexts it is usually more appropriate to use the methods
described in RFC 3066 treating a missing language tag as "*".
-- Some proposed words by Jeremy Carroll
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Mar/0005.html
We are still working on this; consider the above a preview :)
Cheers
Dave
