Re: recent tweaks to RDF semantics

A nit re. section 4.1 -- you say:
[[
1. Add all of the following triples *which contain a name in V*. These are 
true in any rdf-interpretation.
]]

Should that be more like:
[[
1. Add all of the following triples *which contain only names in V*. These 
are true in any rdf-interpretation.
]]

Otherwise you'd potentially end up introducing new names into the graph 
that are not in V?

#g
--

At 20:50 10/06/03 -0500, pat hayes wrote:

>  http://www.ihmc.us/users/phayes/RDF_Semantics_Edit_nT.html
>
>now has some tweaks which fix some bugs noted in recent emails.
>
>The 'se' rules (now moved into section 4 which now has all the rule stuff 
>in one place) have been modified so that they only generate a single bnode 
>from each URiref or literal in the original graph. This is enough for the 
>closure lemmas and keeps them from generating all those redundant 
>blank-node copies.
>
>The definitions of rdf and rdfs interpretations have been slightly tweaked 
>so that the relevant vocabularies contain the rdfV and rdfsV vocabularies 
>(previously, an rdf-interp of V was an interpretation of V+rdfV; now it is 
>just of V, so one has to be explicit about V including the 'special' 
>vocabulary; similarly for rdfsV).  The only effect of this tweak 
>(suggested by Dan C) is to make it possible for an rdf-interpretation to 
>only interpret the part of the infinite container vocabulary that it 
>actually needs; so the definitions of closure are now relativized to a 
>vocabulary, and now refer to those axioms which contain names from that 
>vocabulary; and the closure lemmas now refer to closures relative to the 
>vocabulary of (S+E).
>
>I have rewritten rule rdf2 so that it replaces an XML literal by its 
>canonical form, rather than allowing arbitrary equality substitutions. The 
>exact wording may need massaging, but the idea seems sound.
>
>Taken together, all this enables us to completely avoid the 'infinite' 
>cases in closures, so all closures (from finite graphs wrt a finite 
>vocabulary) are finite, without any need of weasel-words.
>
>The weird rule rdfs12 has been further modified. The elegant version 
>suggested by Jos unfortunately turned out to not be quite valid. This rule 
>may still get some further tweaking but this will not change any other 
>part of the MT.
>
>The Lbase appendix is now in line with the rest of the document; the 
>datatyping in particular has been rationalized.
>
>Apart from getting the XML literal terminology exactly aligned with 
>Concepts, and writing out the proofs properly, I believe this is now 
>pretty much done.  On the other hand, Ive said that before....
>
>Pat
>
>--
>---------------------------------------------------------------------
>IHMC    (850)434 8903 or (650)494 3973   home
>40 South Alcaniz St.    (850)202 4416   office
>Pensacola                       (850)202 4440   fax
>FL 32501                        (850)291 0667    cell
>phayes@ihmc.us       http://www.ihmc.us/users/phayes

-------------------
Graham Klyne
<GK@NineByNine.org>
PGP: 0FAA 69FF C083 000B A2E9  A131 01B9 1C7A DBCA CB5E

Received on Wednesday, 11 June 2003 03:40:24 UTC