Re: [Proposal][SKOS-Core] skos:denotes

In message 
<350DC7048372D31197F200902773DF4C05E50C45@exchange11.rl.ac.uk> on Wed, 
29 Sep 2004, "Miles, AJ (Alistair)" <A.J.Miles@rl.ac.uk> wrote
>
>Hi Leonard,
>
>No, this is something a bit different.  Others will be able to explain the
>intention more clearly than I, but here's a brief go (I hope I don't do more
>harm than good) ...

Al

Thanks. Do tell me to shut up and go and read some basic documents if 
you think I'm asking for a tutorial in the underlying philosophy of 
these systems, but they do seem to be making things unnecessarily 
complicated. . .

>The 'ontology style' describes things in terms of their properties (i.e.
>with *logical* constructs).  For example, you can use OWL (and FOAF) to talk
>about 'the Person with email address <mailto:a.j.miles@rl.ac.uk>'.

i.e. the individual member (instance) of the class "persons" 
distinguished from other members of that class by having the specific 
email address "<mailto:a.j.miles@rl.ac.uk>". As someone else pointed out 
last week, an email address does not necessarily define a person 
uniquely. Presumably you could as an alternative say "the person with 
the social security number xx nnnnnnnnn" or "the person with the 
initials "AJM" who works at the Rutherford Laboratory", which might 
narrow the field sufficiently for the required purposes.

How is this style applied to non-specific concepts, though? You can't 
give URIs for concepts like "peace", "water", "God", "physics" or 
"teapots", unless it is to specific written definitions of these 
concepts which you have agreed to use as indicating the scope of the 
concept for the purposes of the ontology. In that case, as Stella 
suggested, it would seem much more straightforward to import the text of 
the definition as a scope note. If different ontologies point to 
different definitions, or different distinguishing properties, then you 
have no basis for saying that the concepts are the same.

>The 'thesaurus style' describes concepts by using their indicative lexical
>labels, explanatory prose, and links to semantically related concepts (i.e.
>with *linguistic* constructs).  For example, you can use SKOS Core to talk
>about 'what I have in mind when I use the term [Alistair Miles]'.

Yes, I understand and agree with that.

>The proposal here is to model 'the Person with email address
><mailto:a.j.miles@rl.ac.uk>' and 'what I have in mind when I use the term
>[Alistair Miles]' in RDF as *separate resources*.  A special predicate (i.e.
>skos:denotes or whatever) may then be used to link them.

You are saying that in some cases you can recognise (presumably on the 
basis of some external evidence) that concepts from two different 
ontologies or thesauri are near enough the same to be treated as 
identical, and what you are looking for is some way of labelling them as 
such. Is that right?

>It would be nice to have a firm philosophical basis for this proposal, but I
>think the primary motivation is pragmatic (i.e. how to cope in practise with
>co-existence of OWL and SKOS Core) so ... hmmm.

Indeed!

Leonard


-- 
Willpower Information       (Partners: Dr Leonard D Will, Sheena E Will)
Information Management Consultants              Tel: +44 (0)20 8372 0092
27 Calshot Way, Enfield, Middlesex EN2 7BQ, UK. Fax: +44 (0)870 051 7276
L.Will@Willpowerinfo.co.uk               Sheena.Will@Willpowerinfo.co.uk
---------------- <URL:http://www.willpowerinfo.co.uk/> -----------------

Received on Wednesday, 29 September 2004 16:16:30 UTC