Re: What is the URI of Truth?

[...]
> And I stated that, no, the spec does not say URNs and only
> URNs (re-read my posting).
>
> I *do* read "designed with these goals in mind" as equating to
> more suitable than "can be managed [made to work]" but *if* you
> know how to manage them to be made to work that way, etc.

Yes, this is my problem - I think you're reading way too much into
that. Take all specifications at face value.

> Far more critical than the name versus location, URL as URN
> issue is the reliable derivation of URIs from QNames, [...]

Yes, I agree with this.

> [...] and that is what I see as threatening the inoperability
> of the SW.

I'm not sure I agree as much with this. On the one hand, we can't
represent stuff like xhtml:title element/attribute in RDF, simply
because there is no mapping, but on the other hand, I don't think that
this is a problem with http://abc#xyz as identifiers *in RDF*, because
we can use these abstract "terms" reliably. Your concern is actually
with importing non-RDF namespaces into RDF (and possibly vice versa),
i.e. namespaces that are designed for use only in thier own particular
systems, and I think that it is wrong of you to rain down on current
practices that are outside of this scope.

So I think that you have to disambiguate the "RDF-ness" and "XML-ness"
of namespaces right away. RDF uses concatenation, and that works fine
for any namespace *in RDF*, because it's just a simple alias. It does
not work when importing external "types", i.e. concepts, into RDF, and
it is a matter of contention whether or not they work when exporting
them out of RDF. Whatever, I don't particularly agree that it is a
critical apect of the SW to be able to define non-ambiguous interplay
between different namespacing mechanisms, simply because the Semantic
Web as it is is based on anything that is in RDF, or that can be
modelled in RDF, can modelling stuff it quite easy on a contextual
basis. I agree that broadening this is a problem that should be
addressed, but I don't think of it as being an immediately critical
thing. 99.99% of SW applications will have context (I mean that in the
abstract sense of the word).

My point is that we have many RDF-centric applications, and that these
should not be forced to adopt any sort of naming scheme which is a
restriction of the current standards and practices.

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
:Sean :hasHomepage <http://purl.org/net/sbp/> .

Received on Friday, 8 June 2001 12:31:07 UTC