W3C home > Mailing lists > Public > public-sws-ig@w3.org > November 2005

Re: Where are the semantics in the semantic Web?

From: Harry Halpin <hhalpin@ibiblio.org>
Date: Fri, 25 Nov 2005 13:12:46 -1000
Message-ID: <43879A6E.7010500@ibiblio.org>
To: Bijan Parsia <bparsia@isr.umd.edu>
Cc: public-sws-ig@w3.org

Let's keep snarkiness down please, can we :)

I would say one good use I could see for RDF/OWL would be to allow
people to define a XML\XSD document neutral format for combining (graph
merge/OWL subsumption and identifying data (using URIs) from documents
with differing XML/XSDs. Then instead of having to make one XSLT
conversion to every other XSD you might want to integrate with, you have
one GRDDL to RDF/OWL/NextW3CRuleThing and back again.  Seems to save
time. I could even imagine making a business case for it.

 To connect the threads,I do think RDF/OWL/NextModelThing bindings make
sense since they give the user the ability to keep their data in their
custom XML format (whose contraints and data types are provided by XSD
which, thanks to the formal operational semantics of XQuery, we can now
do static type checking on and such), while allowing them model their
knowledge using a KR grounded in a well-defined formal semantic model
made for modelling things on a more abstract level and providing
inference capabilities, and *connecting* the two. The other case would
require re-engineering of XML (not going to happen) or using
non-standardized things possibly with no semantics. Therefore the best
of both worlds, at the level of forcing the user to use more technology
(XML+RDF/OWL) though. Is that cost worth it?

                                  cheers and good night,
                                          harry

P.S.     I apologize if I was using terms "denotational" and
"operational" in the manner they are usually used in programming
language semantics theory as opposed to informal  conversation. Again,
with list-servs this sort of "talking over each other" is easy to do.
However, I do think that  my terminology and distinction are correct, as
a cursory glance at the literature and Wikipedia show. As Bijan did
correctly note, there are differences in complexity of models - this is
reflected usually in differences in how one and with what one defines
semantics. Let's be less blustery and more gracious and look for the
differences between RDF/OWL and XML mappings, and if there is any use
for RDF/OWL, okay? I believe
the jury is still out, but I'm actually warming up to RDF/OWL.

P.S.S. I do think the connections between programming language types and
ontologies are interesting. I have yet to see anyone really give a good
run-down of the differences and similarities - has anyone?? - that would
be really useful in helping focus and going beyond the permathread level
to actual interesting use cases and solutions. I'll see if I can get a
good one running asap.

>>> Operational semantica (from Plotkin) has evaluation and
>>> execution relations specified by rules directed by syntax, the
>>> denotational semantics (from Scott) uses rich mathematical models ala
>>> partial orders, least fixed points
>>
>>
>> Sigh. Doesn't it depend on the complexity of the language? The main
>> point of difference is that operational semantics explicate meaning
>> in terms of the operations of an "abstract machine", whereas
>> denotational semantics provide an interpretation function (as with
>> "normal" model theory). Note that denotational semantics is quite
>> syntax driven (valuations for expressions depend on the valuations of
>> their parts), so your contrast is a bit confused.
>

> Of course, this doesn't matter since originally we were just looking
> for a "formal semantics". If you wanted a denotational semantics
> (rather than any sort of formal semantics), ask for it. I don't see it
> helps you, though.


> I notice that you didn't address the main point, either. Regardless of
> the *style* of semantics given, the expressiveness of the language
> either lets you do certain things or not.
Received on Friday, 25 November 2005 23:12:55 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 16 March 2008 00:11:02 GMT