Rolling back on lang tags for literals

Considering the difficulty and trouble that including lang tags
as part of literal nodes is causing, we did some brainstorming
about how we might live without them without huge modifications
to our architecture and came up with an approach whereby we partition
our schemas by language and maintain distinct graphs in the KB which
can be virtually merged as required for querying to achieve a form 
of language sentitive scoping of assertions.

It's a bit of work, but reasonable, and therefore lang tags
disappearing from literal nodes will not be as big a show stopper
as previously thought.

We therefore retract our dissent to the decision to remove lang
tags from literals, and would be able to live with their removal.

That would seem to make both the abstract syntax and the semantics
of literals easier for alot of folks.

Patrick

[Patrick Stickler, Nokia/Finland, (+358 40) 801 9690, patrick.stickler@nokia.com]


P.S. We're still royally pissed off about the string-based semantics ;-)


----- Original Message ----- 
From: "ext Jeremy Carroll" <jjc@hpl.hp.com>
To: <w3c-rdfcore-wg@w3.org>
Sent: 30 October, 2002 21:00
Subject: RE: Datatyping literals: question and test cases


> 
> >I would suggest that we let the model theory merge the cases back
> >together, by ignoring the lang tag for non-XML literals.
> 
> I am happy to do it in the (abstract) syntax if I've lost that argument.
> (Dan and Pat speaking against me so far - unless anyone else speaks up; I
> will concede)
> 
> it feels more of a syntactic issue to me - the model theory can then leave
> untyped literals as self denoting, which was definitely the WG preference.
> 
> Jeremy
>

Received on Thursday, 31 October 2002 03:13:28 UTC