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

Re: Using third-party vocabularies

From: pat hayes <phayes@ai.uwf.edu>
Date: Tue, 3 Dec 2002 23:44:57 -0600
Message-Id: <p05111b2aba1344cc721d@[10.0.100.247]>
To: Graham Klyne <GK@NineByNine.org>
Cc: w3c-rdfcore-wg@w3.org

>At 11:21 AM 11/21/02 +0100, Jeremy Carroll wrote:
>
>>I would like further discussion about social meaning, particularly the clown
>>example.
>>http://www.w3.org/TR/rdf-concepts/#section-InteractionExample
>
>(Good questions...)
>
>>Issues with that example include:
>>   does using a term from another namespace implicitly import all of the
>>triples used in the defining document?

No, but it doesn't have to explicitly import them in order for their 
meaning to be constrained by what is said about them by their owner. 
Thats part of the point of the example: the shreds of meaning 
attached to a name at its source still apply even if you only *use* 
the name. Thats what I meant by caveat lector: its up to the reader 
of what I publish to check that the meaning I attach to my terms is 
what he wants to inherit by his use of those terms.

>
>Ah... I think not in the sense that RDF entailment is a relationship 
>between RDF graphs, without reference to how those graphs are 
>constructed from documents.

Formal entailment, yes, but that the point of talking about these 
issues in a social setting.

>But then how are we to talk about the meaning of an RDF term as 
>defined by some other person?

Right, exactly.

>  If we can say that the meaning of the term is defined by its 
>'owner' to be constrained by the given RDF assertions (asserted 
>statements), then I think we can legitimately include those 
>statements in the resulting graph.

Well, at the least one cannot reasonably claim to be using the terms 
in a way that does not inherit that source meaning. Thats not quite 
the same thing as asserting it yourself, but its close enough.

>
>>   does using a term (legally) commit one to everything said in the defining
>>document?
>
>(Nit:  well, we're not defining legal matters here.  What we are 
>trying to convey is a sense that the meaning of some RDF depends on 
>the meaning of terms defined by some other party.)
>
>More importantly, I think RDF core is not specific about the 
>relationship between published documents and asserted RDF graphs: 
>there are also social considerations involved.
>
>>   relationship with owl:imports?

I would strongly recommend that we don't go there. Owl:imports is a 
minefield of unresolved issues and half-digested confusions. Webont 
would do well to abandon owl:imports and just use rdfs:seeAlso.

>
>RDFcore does not depend on owl.  owl:imports might be defined in 
>terms of RDF graph construction from referenced documents, which 
>would thus define a relationship
>
>>   if we don't have the example are we saying anything? or would any words be
>>so woolly as to have no real significance.
>
>It would be good to have words that would stand without the example, 
>but that may prove to be a tall order
>
>>A problem is that of how to find the defining triples and to discard other
>>triples from the defining document.
>
>I think that particular problem is out of scope for us...

I agree. And we should be careful about using the D-word.

>
>>The long lengthy and tedious discussion of owl:imports in the other group is
>>indicative of the difficulties.
>
>... and maybe that indicates why it should be so.
>
>>A particular problem with owl:imports is the significance of http errors and
>>whether any meaning can depend on the absence of http 500 errors. (I don't
>>care just trying to report the issues ...)
>>
>>This example started off life as a test case from Brian ("Would an expert
>>witness be justified in testifying ..."), and was partially developed by
>>Pat.
>
>Yes... and the conditions that would justify an expert witness 
>taking a particular view will depend on both the formal language 
>definition and the social circumstances of its publication.
>
>>I hope to have a editor's draft out later today. The example will still be
>>in, but labelled (informative), the enclosing section is normative.
>
>I agree that the example, if included, should be informative.
>
>...
>
>Related to all this, I think there is a legitimate caveat:  if one 
>makes legally binding statements using language terms defined by 
>some other party, then one would be well advised to confirm that 
>they define the terms persistently in a fashion that corresponds to 
>the expected meaning.   So, in important documents, one would 
>probably limit oneself to using vocabulary defined by some reputable 
>organization, or standards body, etc.

Good point to make. If you use terms from sources that you don't 
trust, they may have consequences that you don't intend. Caveat 
lector.

Pat

>#g
>
>
>-------------------
>Graham Klyne
><GK@NineByNine.org>


-- 
---------------------------------------------------------------------
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 Wednesday, 4 December 2002 00:44:46 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:54:48 EDT