Language attributes- what are they?

[Peter Constable]

> From: Tex Texin [mailto:tex at xencraft.com]


> 3) Your test for locale vs language involving whether an API is needed
> seems
> suspect to me.
> If I am given a document with 3/4/5 and told it is English US, or even
if
> I
> assume the origin, then it tells me (generally) how I should read
(parse)
> it to
> understand it.

Sorry, there were some other comments on this I meant to add: 

Yes, knowing that the document is in US English will generally indicate
how you should interpret "3/4/05". But that is simply related to the
comment I made earlier that knowing the language variety/writing
system/spelling of content may be enough to select a corresponding
locale, especially when part of the language variety/spelling identity
involves a regional variant. 

(And note that it doesn't guarantee you know how to interpret date
stings, which an author might write manually in any number of ways. As a
Canadian living in the US, I'm somewhat inconsistent with date formats;
whether I tag my text as "en-US", "en-CA" or "en-GB" has little bearing
on how I write dates.)

But that's different from saying that the metadata element we use to
declare linguistic attributes of static content should also include a
declaration of how date strings are to be intrepreted. That would have
to lead to language tags along the lines of
"en-US-d-m-slash-d-slash-yy". Considering common scenarios, I don't
think it makes sense to go there with language tags cum language tags.
(If someone wants to create a specification for locale IDs that use RFC
3066bis language tags with various extensions, that's another matter;
but those would be something distinct from language tags, IMO, and I
wouldn't suggest that that spec be referenced for something like
xml:lang.)


Peter Constable