I18N - Re: test cases LCC

Jan wrote

[[
- Dave has also raised the fact that we're no longer normative on the
i18n stuff. However, there are some interesting test cases here. These
are hard to get right and involve huge swathes of unicode tables.
I'm open to suggestions as to what to do here: we can approve these but
mark them as non-normative, for instance. This is more of a
process/scope question than an editorial one, so I'd like some input
from the chair/staff contact.
]]

I think there's a misunderstanding here.

We are no longer normative on Charmod.

The I18N responsibilities were ours, and we have taken decisions informed by
Charmod and reflected in Concepts.

These were:

+ Literals are in NFC
+ RDF URI references are IRIs more or less.
+ RDF processors should use early unifrom normalization
+ Literals beginning with combinging chars are permitted but discouraged.

Of these, the test cases test the first two. The third has been dropped for
pragmatic reasons - Charmod is not baked, and the RDF WDs do not talk about
processing (except for the deleted part of concepts that imported the
concepts of processing from charmod).

Jeremy

Received on Thursday, 19 December 2002 06:17:04 UTC