[sc34wg3] TR: comment - RDFTM: Survey of Interoperability Proposals

Murray Altheim sc34wg3@isotopicmaps.org
Thu, 10 Mar 2005 02:31:06 +0000

Bernard Vatant wrote:
> Folks
> Please find below an exchange with Peter F.Patel-Schneider,
> which maybe calls for some answer(s) from TM community.


One might note that the "formal semantic definition" of
RDF [1] was written by an outsider to the RDF community,
someone who was not involved in RDF until relatively
recently, and someone who has hardly been a fan of RDF or
the "Semantic Web" [2]. That  "formal semantic definition"
became a W3C Recommendation five years after RDF and
simply cannot be measured as part of any of RDF's success,
a success that can be attributed more to its embrace by a
part of the Description Logic community (as a syntax, not
a model for DAML) and the selling of the resulting product
to the US military and intelligence community, than to any
"formality" in the existing RDF specification, which was
flawed in many ways (which is not to say this is different
than any other first round of many specs, including Topic
Maps). But without the funding provided the W3C by DARPA
the "Semantic Web" would hardly have gotten off the ground.
What was it, something like US$70 million, just to get
started? It's not like Tim Berners-Lee is some kind of
expert in knowledge representation, any more than he is an
expert in SGML, but he's very good at marketing. Had the
Topic Map community been given $70 million I can tell you
we'd be at least be able to wave formal models under
people's noses by now.

So, speaking of noses, it seems kinda snotty to be touting
some kind of formal, logical definition for RDF when for
the vast majority of its history such a definition did not
exist, was not sought after by the W3C, and as such might
be seen as somewhat of an afterthought. Pat brought the
proposal to the W3C, not the other way around [4]. Pat's
name is now being used to lend weight to a project where a
majority of its participants are not even demonstrably
equipped to understand what he provided. But it's a nice
flag to wave under someone's nose. It'd be interesting to
see people tested on their understanding of Pat's model.

Nice try.

That someone managed to get Pat Hayes involved is to their
credit, but pretending that RDF's relative success is based
on that input, or that somehow "formalists" are paying more
attention to it now because of his input, is simply
disingenous. Pat thought RDF needed a model and he wrote one
for it (and to my understanding, somewhat begrudgingly). He
said, "Its all being done to help make the world a better
place, or something like that" [3], as part of an overall
effort to provide a unifying, underlying logical framework.
I think it laughable that someone might think that RDF
developers now understand Pat's model and are paying any
attention to it. If the "formalists" are paying any attention
to RDF (and that remains to be seen, other than as a graph
interchange syntax), it's because nowadays research money is
being poured into "Semantic Web" efforts, not because RDF is
itself thought highly of by those versed in logic. I think
the fact that a model didn't even exist until last year
proves that point. Or ask someone like John Sowa.

I see people here at KMi jumping all over SW efforts because
it's now perceived as the latest gravy train, that putting
"Semantic Web" in the title of a paper or a project is more
likely to garner funding. A project may have little or nothing
to do with RDF (and in KMi's case, the underlying technology
is still usually OCML, which was developed long before RDF was
a twinkle in anyone's eye), but dammit if we don't import and
export to OWL! How de rigueur! It's like owning a pink iPod.

Indeed, people like Pat, Chris Welty, Nicola Guarino and
others from the mathematics, logics, AI/KR, etc. worlds are
now involved in various pursuits related to "Semantic Web"
efforts, but I don't see much of anything of their work on
issues like identity or logic reflected in SW documents (if
anything, I see disagreement), and I don't see any integration
of "Semantic Web" concepts back into wider issues of knowledge
representation. Some in the KR community have been critical of
the SW sucking up money better spent on projects doing knowledge
representation outside of the narrow realm of DL. I'd be curious
to know whether or not the DL people and Pat even see eye-to-eye
on that formal model, other than as a compromise? I remain quite
skeptical that DL is even an appropriate logic for a "Semantic
Web" (why not some other more flexible form?) and I know there
are at least a few well-respected logicians and computational
linguists who feel the same way. I'm certain Doug Lenat would
have like to see more emphasis on Cyc, which is at least (as a
"common sense" ontology) more demonstrably related to the needs
of a Web for the Rest of Us, not just DARPA and large corporate
backers, though I note that RSS is working fine without Pat's

So I think we can relax in knowing that *at least* as much
thought has gone into modeling Topic Maps as was RDF, and we've
spent a great deal of time on issues like identity from a real
user-centered perspective (not particularly from logic, which
is not even demonstrably useful in a world that doesn't always
follow logical rules, i.e., the world we live in). Perhaps one
day we'll be so likewise fortunate to have a logician grace us
with a model so that we can strut our stuff too.

So what about all this?

Pat Hayes is certainly one of more well-respected logicians
around, and for any serious work I'm putting my money on the
Common Logic project, of which Pat, John Sowa, Chris Welty
and a few others are writing up for submission to ISO. Pat
does not consider things like SWRL serious contenders to CL [3]
so if people want to point to Pat as the expert, I suggest we
do the same and use what Pat would recommend, which would be
Common Logic. The CL work is being edited by Harry Delugach
and is expected to go out for ballot within SC32 this summer.

I'm doing my part in that effort in working with Pat to make
sure we have identifiers for all CL concepts so that we can
bring CL semantics into a Topic Map, just as we can likewise
bring in a Conceptual Graphs model or any other. As Bernard
is certain to know, it's also possible to bring in OWL
semantics into a Topic Map. This fallacy that some kind of
"formal model" needs exist before Topic Maps might be "taken
seriously by formalists" is a pretty silly argument. RDF was
able to tack on a model years after it was a W3C Recommenda-
tion, and I don't see as a consequence that it's being taken
any more seriously by the formalists I know.

But if we need any "answer" to this question, I see no reason
why Topic Maps can't do the same, but by summer we can choose
a better basis in Common Logic. There'll be mappings from CL
to most other things, like Pat's RDF Semantics, CGs and KIF
(probably even to PowerPoint and the iPod). At that point if
we map CL to TMs, we'll have a formal basis for doing RDF/OWL
conversions to and from Topic Maps. As formal as anyone else,
and it will eventually have the benefit of also being an ISO

At that point we can all begin carrying our own hankies.


[1] RDF Semantics, W3C Recommendation 10 Feb 2004, Pat Hayes, ed.
[2] Social Meaning and the Cult of Tim
[3] [SCL] Re: [CL] SWRL first-order logic, Pat Hayes
[4] KIF: Standard for Logical Foundations / "semantic web content
     transfer format" (included at bottom)
Murray Altheim                    http://kmi.open.ac.uk/people/murray/
Knowledge Media Institute
The Open University, Milton Keynes, Bucks, MK7 6AA, UK               .

  Sometimes things are so obvious that they merely need pointing out:

       Abstinence Clearinghouse