[topicmapmail] Conceptual Graphs are Step 6

[Murray Altheim]

Jack Park wrote:
> Useful thread, this.
> 
> I suspect that the new RM is much closer to the right modeling tool for 
> doing CGs and having topic maps fall out simultaneously, than is XTM.
> 
> Jack

Peter Becker, myself, and a few others have been trying for several
years now to nudge the CG community into providing a reasonable
specification for Conceptual Graphs. In talking with various CG
developers, and in my own experience, there's no way to build a CG
tool that follows any specification (unless one is satisfied with
toys or creating from whole cloth), as the CG spec isn't complete
enough. All CG tools that I've seen are interpretations of CG. (If
anyone doesn't take my word for it, contact Finnegan Southey, the
developer of Notio.) In the past few days, Peter and I have again
been harping on about this subject on the CG list, so I find the
thread here a bit ironic. To put it bluntly, I don't think CGs are
any more ready for TMs than TMs are ready for CGs. And XTM is just
syntax. I don't see that anything is going to "fall out" of this,
as developing real ontology semantics is hard work, it's not just
a concept here and a relation there. There are whole conference
series on context alone, and years of research on what constitutes
identity. I've been looking into this for several years now, and I
still have only just touched the surface. I'd say the KR community
is still at least a decade away from solidifying a great deal of
this stuff, especially with a lot of valuable resources now being
sucked into the vortex of the Semantic Web.

I don't know how to evaluate the RM as being a necessary component
for ontology. If it provides anything, it would likely be only very
low level, such as some fundamental relation type or structure. I've
developed my own ontological framework based on some fundamental
concept and relation types, stealing some ideas from Cyc and Faceted
Classification. I used the graph semantics of Topic Maps and built
from there. I didn't need the RM for that, though I also don't know
that what I've done is in any way contrary to the RM. I could in
theory have done something similar using the graph structure of
either GXL or RDF, but I wouldn't have had the benefit of merging
(as defined in ISO 13250 and implemented in XTM 1.0), which has
proven invaluable. I also wanted a mapping metaphor, not simply a
graph (since my ontologies are meant to apply to organizing and
classifying resources). Topic Maps have proven a very good fit for
ontologies as mapped to resources, i.e., what they were designed
to do.

I'd like to think Conceptual Graphs might be part of this equation,
but I don't see this happening until there is a real CG standard,
not simply the continuing hand waving about it being FOL. The bones
need flesh. If and when there is either a Common Logic or Conceptual
Graphs standard, their PSIs (and I will do whatever I can to see
that there are URIs available for that purpose) can be used in
building interoperable Topic Map-based ontologies, interoperable
with any other CL or CG tools. Until that time, everything will be
custom, like Cyc, OWL, Notio, CharGer, CoGITaNT, etc. All good and
valuable work, but all non-standard and non-interoperable since
there is no standard core. We don't need a holy grail, but we do
need a standardized expression of FOL plus some higher level
semantics, such as in CG (and OWL is not it, despite the marketing).

Murray

......................................................................
Murray Altheim                    http://kmi.open.ac.uk/people/murray/
Knowledge Media Institute
The Open University, Milton Keynes, Bucks, MK7 6AA, UK               .

   The Resume of George W. Bush
   http://www.buzzflash.com/contributors/03/04/23_resume.html

                    The sleep of reason brings forth monsters -- Goya