model theory state of play

Sorry about the delay, I have been somewhat slowed down for medical 
reasons. There is a revised draft of the MT (dated 3 October) now 
visible at
http://www.coginst.uwf.edu/users/phayes/w3-rdf-mt-2.1_draft.html

It is not completely finished (or spell/link/html-checked); the 
appendices and references need fixing, but the main text is close to 
done and will give the flavor of the proposed changes, most of which 
are in the presentation rather than the content. Feedback welcomed. 
Sorry I havn't managed to style-mark all the changes, there were just 
too many.

I made minor improvements to the wording in several places, added a 
few explanatory sentences to cover cases that gave rise to 
misunderstandings, etc. .Major changes are:
rdf entailment re-christened 'simple entailment'; ('rdf entailment' 
now refers to the rdf mini-vocabulary in section 3)
0.2 RDF graph definition fixed, with some links to web explanations 
of technical terms (is this kind of outward linking considered kosher 
in W3C documents, by the way?)
1.2 rewritten to introduce notion of reserved vocabulary
old section 2 absorbed into subsections 1.5 and 1.6 (rationalization, 
no content changed) 1.6 is somewhat rewritten to refer to the DAML 
axiomatic semantics. The new figure should be bug-free.
2 (old 3) unchanged apart from terminology shift in introduction.
3. New
4. New, introduces idea of namespace entailment
5. and 6. Revised material from old document, but presented differently.
6.1 New (comments particularly solicited, this could be completely 
omitted and may be changed later in any case.)
7. Old section 6, rewritten to conform to new exposition style, but 
no real change in content. Makes points better, though. (This could 
now be rewritten now more formally in the same style as 3/4 and 
5/6,or incorporated into 3/4 as part of the rdf reserved vocabulary, 
if people think that would make more sense. I await input.)

My earlier goof regarding domain and range has been fixed, and so now 
rdf-entailment is a pretty piffling affair just involving rdf:type 
and rdf:Property, but it does serve to introduce the idea of 
namespace entailment and provides a contrast to the much more 
elaborate rdfs case. Ive tried to display the rdfs closure rules in a 
more organized fashion to show how they correspond to the semantic 
conditions. Jos, if you can find any bugs in this, I will buy you a 
beer.

The treatment of literals is still under discussion. At present the 
only substantial change from the published MT is the simple fix I 
mentioned earlier, where ICEXT(I(rdfs:Literal)) is required only to 
be a subset of LV.  Peter thinks that this isn't adequate, and that a 
somewhat larger fix will be required in which literal values are 
completely removed from all contact with the range/subclass 
conditions in rdfs, but I'm not fully convinced yet. :-)

Pat

-- 
---------------------------------------------------------------------
IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes

Received on Wednesday, 3 October 2001 21:58:49 UTC