Re: TDL

On 2002-02-03 20:53, "ext Seth Russell" <seth@robustai.net> wrote:

> From: "Patrick Stickler" <patrick.stickler@nokia.com>
> 
>> Well, I think that the MT should be able to make clear
>> the fact that if you have a TDL pairing (lexical form
>> and datatype context) then you have an unambiguous
>> denotation for a single mapping between a lexical form
>> and a value (between a member of the lexical space of
>> the datatype and its corresponding single member of
>> the value space of the same datatype). So the extra
>> triple would be superfluous insofar as the interpretation
>> is concerned.
> 
> Ok I can see now that the extra triple is irrelevant.  I'm sure I have
> missed some subtleties in  this controversy.  But it seems to me that the
> global idiom entails the local idiom and visa versa in a straight forward
> manner:

I believe you are right (though the MT is still being worked out).

What perhaps makes this less obvious is the lack of consistency
in the local vs. global graph representations (due to backwards
compatability and current usage considerations). There has been
discussion to possibly adopt a single consistent graph representation
for both global and local idioms, based on the current local idiom,
using a manditory bNode but making the rdf:type arc optional.

Thus, we'd have locally/explicit

   Bob ex:age _:1 .
   _:1 rdf:value "30" .
   _:1 rdf:type xsd:integer .

and globally/implicit

   Bob ex:age _:1 .
   _:1 rdf:value "30" .
   ex:age rdfs:range xsd:integer .

and in both cases, the bNode represents the
value. Whether the rdf:type arc is explicit
or implied by a range constraint is the only
difference between the idioms.

The serialization that equates to

   Bob ex:age "30" .

would simply be treated as a contracted form of the
bNode variant and handled by the parser when generating
ntriples (or some other representation of the graph).

Whether this approach will recieve acceptance from
a majority of the WG (or whether TDL itself will)
of course remains a topic of active debate ;-)


> So if its all just entailment of triples to triples what's the big problem
> with the MT?  Sorry, I keep trying to figure out what the MT gives us that
> formula like the above don't.

Well, to be humble and honest, I am not a mathematician and the
definition/debate of the MT is a little over my head (I follow
just enough of it to know to stay clear of most of it ;-)

Perhaps some of the other WG members with competence in that
area would care to comment, if following this thread...

Cheers,

Patrick

--
               
Patrick Stickler              Phone: +358 50 483 9453
Senior Research Scientist     Fax:   +358 7180 35409
Nokia Research Center         Email: patrick.stickler@nokia.com

Received on Monday, 4 February 2002 03:26:06 UTC