Introduction: Bernard Vatant

I've been working more or less in the Semantic Web area from the Topic Maps
side of the mirror since 2000, as a consultant for the French software
company Mondeca.

Before that, my background activity has been in Mathematics and Astronomy
(20 years of teaching and popularization) and I've always had interest in
identification issues. What is the subject, how humans (and now systems)
can know they speak about the same thing or not is a critical question I
have met at every corner of learning process and scientific discovery as
well, and I was not completely surprised to find it as a corner issue also
in the IT universe in general, and Semantic Web in particular.

Meaning of URIs is actually just one place where the question is emerging
and hurts. I have worked on this issue in the Topic Maps context,
particularly through my activity in the OASIS Published Subjects Technical
Committee. I tend to consider now this work on subject indicators and
subject identifiers should provide an input to a more general work on the
notions of identification process, identifying properties and
identification context.

I have also followed closely the WebOnt work, and have been accepted
formally as a member lately, though I've not participated actually in the
meetings so far - found it difficult to jump in the process at this stage.

I've previously pushed the notion that any URI is in fact implicitly or
explicitly used to identify different things in different identification
contexts (process or protocols). And that is not a bad thing, provided the
context is explicit and applications are aware of it. Examples are the use
made by Topic Maps of URIs either as subject addresses or subject
indicators (or identifiers), and the efficiency of http protocol (most of
the time). There are certainly other implicit uses to be explicited, either
to recommend or forbid them. So my interest is to see sorted out here the
following questions :

- What are the various identification process/contexts/protocols that
could/should/might use URIs as identifiers.

- In each of those, what is the "thing" identified by the URI.

Hoping this is of course "what we are about" here :)

Bernard Vatant
Senior Consultant
Knowledge Engineering
Mondeca - www.mondeca.com
bernard.vatant@mondeca.com

Received on Tuesday, 16 September 2003 05:25:40 UTC