W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > October 2002

Re: possible untidy route

From: Patrick Stickler <patrick.stickler@nokia.com>
Date: Thu, 10 Oct 2002 11:03:39 +0300
Message-ID: <003201c27033$8ad78ba0$544516ac@NOE.Nokia.com>
To: <w3c-rdfcore-wg@w3.org>, "ext pat hayes" <phayes@ai.uwf.edu>

This either seems non-monotonic, or suggests two types of entailments
which may provide different answers to the same question.

I.e., if we have at first

   Jenny age "10" .
   Movie title "10" .

and that entails

   Jenny age _:x .
   Movie title _:x .

but then later add property based datatyping assertions

   age rdfs:range xsd:integer .
   title rdfs:range xsd:string .

then either the above entailment no longer holds, which is
non-monotonic, or else we have both a yes and no answer to
the entailment depending on whether we are performing a
lexical or value based comparison -- i.e. Jenny's age both
is the same as and is not the same as the Movie's title.

To quote Aaron, "Here be dragons..."  

Patrick


[Patrick Stickler, Nokia/Finland, (+358 40) 801 9690, patrick.stickler@nokia.com]


----- Original Message ----- 
From: "ext pat hayes" <phayes@ai.uwf.edu>
To: <w3c-rdfcore-wg@w3.org>
Sent: 09 October, 2002 14:38
Subject: possible untidy route


> 
> I hesitate to tread into this yet again lest I fall through the 
> crust, but heres an idea which might just keep everyone happy. It is 
> a variant on the old idea of semantically untidy literals, but it 
> still supports the critical tidy-style entailment.
> 
> In our test-case style, here are the entailments you would get.
> 
> Jenny ex:age '10' .
> ex:movie ex:title '10' .
> 
> entail
> 
> Jenny ex:age _:x .
> ex:movie ex:title _:x .
> 
> BUT if you also say (ignore syntactic details)
> 
> ex:age dtyperange xsd:integer .
> ex:title dtyperange xsd:string .
> 
> then that is OK, and now you can infer
> 
> Jenny ex:age _:y .
> _:y xsd:integer '10' .
> 
> ex:movie ex:title _:z .
> _:z xsd:string '10' .
> 
> Obviously _:y isn't the same as _:z. The cost is, that _:x isn't the 
> same as either of them. In fact, _:x can't be a datatype value for 
> *any* datatype. Think of it as a kind of generic exemplar for the set 
> of all the possible datatype values, or something like that. Still, 
> it *exists*.
> 
> This could work with lexically tidy literals, but it would be classed 
> as semantically untidy, I guess. But it would be easy to tweak the MT 
> to allow this.
> 
> Any takers? Questions?
> 
> Pat
> 
> 
> -- 
> ---------------------------------------------------------------------
> IHMC (850)434 8903   home
> 40 South Alcaniz St. (850)202 4416   office
> Pensacola               (850)202 4440   fax
> FL 32501            (850)291 0667    cell
> phayes@ai.uwf.edu           http://www.coginst.uwf.edu/~phayes
> 
Received on Thursday, 10 October 2002 04:14:53 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:52:23 EDT