W3C home > Mailing lists > Public > public-esw-thes@w3.org > April 2004

RE: URIs for Concepts: Best Practices

From: Bernard Vatant <bernard.vatant@mondeca.com>
Date: Thu, 22 Apr 2004 15:25:15 +0200
To: "Kal Ahmed" <kal@techquila.com>, "Miles, AJ (Alistair)" <A.J.Miles@rl.ac.uk>
Cc: <public-esw-thes@w3.org>
Message-ID: <GOEIKOOAMJONEFCANOKCEEPLDNAA.bernard.vatant@mondeca.com>

Kal, Alistair

Just a quick note from a member and former chair of OASIS PubSubj TC

> I like the approach that is outlined by the OASIS Published Subjects TC
> [1]. Although this document is a draft and omits some pieces that I
> would like to see, I feel that the general approach is a good one.

This is not a draft, it's a TC recommendation. Should be the first one of a series,
unfortunately there has not been consensus yet in the TC for what next recommendations
should contain.

> To summarise, an HTTP identifier, when used to identify a concept *should*
> resolve to a human-readable resource that describes the concept. Despite
> coming from the topic maps community, I feel that this approach is
> applicable to the creation of any identifier scheme that uses HTTP for
> namespacing.

That was the primary idea of Published Subjects. Now I'm not sure it was not flawed from
the beginning, because identification always need "identifying properties" +
"identification context". Topic maps provide clear identification context allowing to make
unambiguous interpretation of HTTP URIs for subject identification. I'm not sure to which
contexts such an interpretation can be extended without ambiguity. Using HTTP namespace is
not a sufficient pre-condition, as endless debates on interpretation of URIs in RDF
clearly shows. I agree with folks saying that the same URI can be interpreted in various
ways to identify different things in different identification contexts. So, to the
question to know if HTTP URIS can be used without ambiguity as identifiers in SKOS, I
would answer that a pre-condition is to define what the identification context in a SKOS
application is or should be.

> The things that I would like to see the TC consider is recommendations
> for either embedding, linking to (e.g. using RDDL) or providing as
> parallel resource (via content negotiation), other machine-readable
> descriptions of the concept and related resource - so an RDF resource
> would be one example, the same information translated into XTM might be
> another and so on.

Before going in such technical details, the quoted above pre-condition has to be

> It would be good to get the Published Subjects work kick started again
> (the committee went quiet a long time ago) - perhaps we could work on
> putting together a technical report to pass either to the OASIS TC or
> just to publish as part of the SWAD work ?

As you are maybe not aware of, there are at least two reasons for quietness of PubSubj TC.
One is the lack of task force, the other is the lack of consensus on further deliverables.
And the latter is linked to the issue discussed here, among others.


Bernard Vatant
Senior Consultant
Knowledge Engineering
Mondeca - www.mondeca.com
Received on Thursday, 22 April 2004 09:25:43 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 13:32:03 UTC