Re: Summary of strings, markup, and language tagging in RDF (resend)

----- Original Message -----
From: "ext Brian McBride" <bwm@hplb.hpl.hp.com>
To: "pat hayes" <phayes@ihmc.us>
Cc: "rdf core" <w3c-rdfcore-wg@w3.org>; <jjc@hplb.hpl.hp.com>; "Martin
Duerst" <duerst@w3.org>
Sent: 03 July, 2003 14:24
Subject: Re: Summary of strings, markup, and language tagging in RDF
(resend)


>
> On Thu, 2003-07-03 at 02:52, pat hayes wrote:
>
> [...]
>
> >
> > Since this issue seems to be so centrally important, and since our
> > design now appears to people like Martin to be so completely
> > brain-damaged, let me propose that we re-open this issue and change
> > our design slightly,
>
> Pat, thank you for this proposal (in the same sense of thank you as in
> "thank you for slapping me about the face with a wet fish").
>
> Please could you hold that thought for now.

Perhaps in the form of a suppository?  ;-)

> My preferred approach to this discussion is first of all to determine
> whether the simpler, and I suspect less controversial, solution of
> reintroducing the wrapper

Now it's your turn Brian... (no clenching, now... ;-)

> would satisfy i18n so far as issue b) is
> concerned, which I'm hopeful it would.

I think that my recent posts,

http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jul/0045.html
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jul/0046.html

clarify issue (b) in that RDF users are not simple XML users and
should be fully aware of how the RDF/XML serialization maps to
the RDF graph and will be basing their applications on the abstract
syntax and not the RDF/XML.

Martins concerns, while very relevant for XML-centric consumption of
RDF/XML instances lose much of their weight when viewed from
the perspective of RDF-centric consumption of RDF statements
expressed in an RDF/XML instance. And I think it goes without
saying that RDF users should *never* base their applications (other
than parsers) on the RDF/XML.

It's a bit of an apples and oranges issue, I think.

As for issue (a), the present solution *does* provide a single way
to associate language with general textual content, as plain literals
are the only form of *general* textual content in the RDF graph. All
other content falls within the scope of a datatype, which is free
to provide its own mechanisms for language qualification, if any.

RDF users may encode marked up content in plain literals, with
the necessary escaping, and associate language tags to that. It's
simply that the RDF MT will ascribe no special semantics to
those literals with regards to the nature of the markup or operations
such as comparison, etc.

Thus, the only reasonable debate is whether it was reasonable to
model XML literals using an RDF datatype. I hope that my recent
posts, referenced above, provide sufficient justification to Martin
and others for that decision.

> RDFCore can then decide whether to accept I18N's arguments that the
> present design should be changed.  *If* they do, then they can decide
> how best to address those requirements.

I see no issue that actually needs to be resolved by the WG.

I am of the strong opinion that no change to the present design is
warranted, and none of the proposed alternatives would be better
than the current solution.

Of course, the rest of the WG is free to disagree...

Patrick

Received on Thursday, 3 July 2003 07:53:21 UTC