ORB Visualization
3/17/2004 5:52 PM
Folks
As the author of the quoted paper and last one in the cc list, I take the freedom to jump in for a few remarks.
Irene
said
You may be interested to see how some vendors are
working on bringing
Topic Maps
and Semantic Web together by using OWL as a constraint
language for
Topic Maps:
http://www.idealliance.org/europe/04/call/xmlpapers/03-03-03.91/.03-03-03.html
Reply
... vendor*s* : maybe the plural is abusive so far. This is a path we have been exploring in Mondeca Inc. Whether other vendors are following the same track I'm not really aware of. But many folks seem to work on that borderline here and there, and I had already interesting feedback on this paper, mostly from outside the vendor's sphere, mostly positive ... and a quite critical one (on technical details, mainly) coming from another vendor's representative, the one quoted by Paul in his on-line comments at
http://www.bcngroup.org/beadgames/techInnovation/nine-three.htm.
That said, I fully agree with Prueitt that 'Bernard Vatant and Lars Marius Garshol are not representing the spirit of Topic Maps'. But not for the same reason as his, I'm afraid.
I agree with that sentence because I think there is nothing like 'the spirit of Topic Maps', so I don't see how Lars Marius, me or anyone else could represent it.
This leads me to make another point. The above paper is about Knowledge Engineering. When I first read this expression, a few years ago, I found it weird to begin with. Then I realized with time that it was really describing what I was about, and I ended up putting it on my VCard.
I'm not sure about what Knowledge is, but I have a few ideas about Engineering : taking tools, bits and pieces of interesting stuff here and there and put them together to make something that eventually works. This is what the paper is about. No more, no less.
Coming back to Irene's message - I would not go as far as saying that OWL (Ontology Web Language) is used as a 'constraint language'. That could be misleading to think that OWL can make for TMCL (Topic Map Control Language). It addresses only the 'ontological' constraints. There will be more than that in future TMCL, addressing specific TM (Topic Map) Data Model features.
Irene said
We believe
this to be a very positive development. It will allow people
who have
reasons to be in Topic Maps stay in Topic Maps and still be
part of the
Semantic Web. It will also give Topic Maps capabilities they
are currently missing.
Thanks. This is exactly what is sought by the author. This is also the objective he would like to follow up with in the framework of the new W3C Semantic Web Best Practices Group. See
http://www.w3.org/2001/sw/BestPractices/
Irene said
We, in
TopQuadrant, see this choice (Topic Maps vs. RDF) as a matter of
preference
for an organization. We are committed to helping companies
make sense
of technology options and make decisions that are appropriate
to their goals and existing infrastructure investments.
Yes. With the time and along projects in Mondeca, we have also pragmatically figured out where Topic Maps fit best and how they are best articulated with other representations. That's why we try to keep our sofware as open as possible to all semantic standards.
Irene said
Topic Map has
an intuitive meta model for talking about things. RDF is
more generic
which leads to it being less intuitive, nonetheless capable
to represent all the distinctions in Topic Maps.
I think we should stop trying to map and compare the two technologies, but show practically where each of them fits the best in an information architecture. There again, an engineering viewpoint ...
Irene said
We believe
that the key deciding factor for the adopters of semantic
technology
will be the availability of tools and their features. A few
years ago
Topic Map vendors were ahead of RDF tools. Today there is much
vendor
activity and excitement around RDF and OWL. As always, the market
drives what
happens in the technology adoption, perhaps the reason for
your e-mail.
To be honest to the TM community, it has to be said that TM standardization has been developed by a very small task force, mainly independent consultants and representatives from companies working on their spare time, and a few individuals from major companies that sometimes had hard time being supported for travel and meetings.
No support for any major institution, no paid permanent members, nor academics supported by public funds. Of course the market decides eventually which technology will survive. But it would be really sad to see the current hype around SW (Semantic Web) technologies let TM by the roadside. OTOH, TM folks should certainly not despise the opportunity offered by SW kick-off, and miss the boat for obscure and arcane debates like real nature of subjects, meaning of URIs, or relevancy of First Order Logic in Knowledge Representation.
Irene said
If we were
to look at this discussion only from the technology merit
prospective,
we would be describing Topic Maps as a computational
engine. We
would be examining OWL as enabling language for many
computational engines including Topic Maps. We would be describing
ways
of building
computational models on top of OWL. We see some elements of
this debate
going on in a number of mailing lists including the one on
topic maps.
I'm not sure what you mean by Topic Maps as 'computational engine'. I always figured TM to be more descriptive than computational, but certainly I miss your point. Could you expand ?
Bernard Vatant Senior Consultant Knowledge Engineering
Mondeca -