Re: RDF Semantics: a partial review

>>...
>>
>>>
>>>>   >>In view of this, it seems better to assume that each
>>>>>>rdf(s)-interpretation satisfies all of rdfV (and
>>>>>>therefore satisfies all RDF axiomatic triples).
>>>>>
>>>>>Yes, of course (now you have pointed it out :-). I will make this
>>>>>change. Peter has previously expressed a dislike for the 'crdV'
>>>>>construction, which was introduced only to keep the closures finite
>>>>>in any case and is therefore now irrelevant.
>>>>
>>>>It seems that this is change is not consistently applied to the
>>>>document.
>>>>The definition of rdfs interpretation includes "which contain only
>>>>names form V union rdfV union rdfsV".
>>>>This phrase should be removed, and similarly for rdf interpretations.
>>>
>>>It seems harmless, since this is the vocabulary of the
>>>interpretation. But it may indeed be misleading, so I have deleted it
>>>as you request.
>>
>>
>>The phrase "for all names in (V union rdfV)" is not yet deleted from
>>the definition of rdf-interpretations.
>
>As applied to the semantic conditions, it should not be deleted. Of 
>course an interpretation need only satisfy the semantic conditions on 
>its own vocabulary, right? What would it even mean to require it to 
>satisfy conditions more broadly? This is in accordance with the 
>normal textbook definitions of satisfaction and entailment.

The table we are talking about, "RDF semantic conditions", has
three parts, none of which seems to need the additional phrase:
- part 1 deals with the universe, not the vocabulary
- parts 2 and 3 both make *explicit* that they talk about
  a certain name in V
Therefore, it seems confusing to add that this table holds "for
all names in (V union rdfV)".

>
>Pat
>
>>

Herman

Received on Friday, 7 November 2003 12:10:00 UTC