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

Re: missing (and incorrect) RDFS axioms

From: pat hayes <phayes@ai.uwf.edu>
Date: Fri, 8 Nov 2002 23:22:01 -0600
Message-Id: <p05111b35b9f24862052b@[65.217.30.130]>
To: Brian McBride <bwm@hplb.hpl.hp.com>
Cc: w3c-rdfcore-wg@w3.org

>At 14:47 08/11/2002 -0500, Dan Brickley wrote:
>
>>On 8 Nov 2002, Dan Connolly wrote:
>>
>>>  > > rdf:object        rdfs:range  rdfs:Resource .             *
>>>  >
>>>  > ...did we agree that all literals are resources?
>>>
>>>  regardless, it's redundant to say range Resource.
>>>  Please let's don't.
>>
>>If there were some er... 'things' that aren't resources (eg. literals),
>>then this wouldn't be redundant. I've lost track of our decisions on that
>>front, hence the prev. msg.
>
>My understanding is:
>
>   - Pat asked the question recently
>   - there has been no formal decision
>   - Jeremy argued there was in effect because
>
>a b "foo"
>
>entails
>
>a b _:l .
>
>"foo" must be a resource.
>
>I don't grok that one myself.

Heres why: because that conclusion says that something exists which 
is the b-value of a. Things that exist are resources, by definition. 
So, if literal values are not resources then they might not exist, so 
the truth of the first triple wouldn't establish the truth of the 
second one. In effect literals would be like fictional names, as far 
as RDF was concerned. They wouldn't be the names of 'things', because 
as far as RDF is concerned, resources are the only things there are.

>   - Pat got various bits of feedback
>   - whatever the MT says is what Pat is recommending
>   - we make the decision by endorsing or not the MT.
>   - I hope he recommends we say nothing on the subject this time 
>round otherwise we'll get a load of questions asking how to express 
>literals as subjects in rdf/xml.

??? I completely fail to see the connection there. Seems to me this 
has nothing to do with literals as subjects.

>I'd rather we said nothing, or maybe gave the implementors some 
>advice to allow for it happening in the future, but not now.
>
>And I'm too tired to go look up what Pat actually decided.

Pat is somewhat confused about this,  to tell you the truth. On the 
one hand, the MT has been assuming from day one that 'resource' just 
means 'entity', that the class rdfs:Resource is the total universe of 
all things that exist, that everything under the sun and indeed over 
it is a resource, etc. . Nobody has ever spoken against that idea. On 
the other hand, saying that literal values are resources seems to get 
some people all in a tizzy. But if Santa Clause and the number seven 
and my cat Henry are all resources, surely such everyday things as 
literal values are resources?

Less philosophically, if literal values are not resources then the 
interpolation lemma will need to be re-stated so as to exclude 
literal objects, and RDF entailment will need to treat literals 
differently from urirefs and bnodes.

If we want to be able to conclude from a triple using a literal that 
something exists which is the value of the property:

aaa ppp "whatever" .
-->
aaa ppp _:xxx .

then we have to say that being a literal value means being in the 
universe, which (see above) means being a resource.

If being in the universe does NOT mean being a resource, then RDF has 
no term for the RDF universe, which would be rather odd and would 
probably break something.

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
s.pam@ai.uwf.edu   for spam
Received on Saturday, 9 November 2002 00:21:43 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:54:01 EDT