W3C home > Mailing lists > Public > semantic-web@w3.org > August 2006

Re: In defence of 404 ...

From: Eric van der Vlist <vdv@dyomedea.com>
Date: Mon, 07 Aug 2006 17:33:21 +0200
To: Bernard Vatant <bernard.vatant@mondeca.com>
Cc: Dan Connolly <connolly@w3.org>, semantic-web@w3.org, Franck Cotton <franck.cotton@insee.fr>
Message-Id: <1154964801.17934.2.camel@localhost>
Bernard and Dan,

Le lundi 07 août 2006 à 16:59 +0200, Bernard Vatant a écrit : 
> Dan
> > Ah... I misread your point. Indeed, in the general case,
> > lots of documents may discuss the same resource, and none
> > of them is authoritative. If DOC1#T1 and DOC2#T1 both
> > refer to France, there is no web architecture mechanism
> > for determining which is authoritative.
> >   
> OK. But I don't want to have DOC1#T1 *and* DOC2#T1 as two distinct URIs 
> defining France. I want one URI to define France. 

[...]

> Sure it is. But the point is that I don't want to change the file. I 
> want to publish a new one about the same entities, without deprecating 
> the old one.

I don't know if that's what Bernard would call elegant :), but what
about defining T1 in both DOC1 and DOC2 with a xml:base equal to DOC
(which means that in both DOC1 and DOC2, the identifier for #T1 is
neither DOC1#T1 nor DOC2#T1 but DOC#T1) and using a placeholder at the
address DOC#T1 with links (using rdfs:seeAlso or whatever) toward DOC1
and DOC2?

(Note that the mechanism can be adapted to slash URIs.)

That keeps DOC1 and DOC2 "equally normative". Applications which want to
use them directly can do it and applications which don't have a clue
where they can find information about DOC#T1 can dereference this URI to
get a first idea where that can look. 

I see that as the equivalent of RDDL for namespace URIs: you use the URI
to publish a hint for applications that have no idea what the namespace
is about.

Would that be acceptable for both of you?

The last solution if this doesn't work for you could be to use either
non HTTP URIs or even anonymous RDF nodes and a
owl:InverseFunctionalProperty. Seems to be an overkill to me since in
this case we have a central authority which can define URIs but the
INSEE code is a perfect candidate for being an
owl:InverseFunctionalProperty...

Eric
-- 
GPG-PGP: 2A528005
Did you know it? Python has now a Relax NG (partial) implementation.
                                          http://advogato.org/proj/xvif/
------------------------------------------------------------------------
Eric van der Vlist       http://xmlfr.org            http://dyomedea.com
(ISO) RELAX NG   ISBN:0-596-00421-4 http://oreilly.com/catalog/relax
(W3C) XML Schema ISBN:0-596-00252-1 http://oreilly.com/catalog/xmlschema
------------------------------------------------------------------------

Received on Monday, 7 August 2006 15:33:44 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:47:17 UTC