W3C home > Mailing lists > Public > public-esw-thes@w3.org > December 2007

Re: Issue : unicity of prefLabel per language per concept scheme

From: Leonard Will <L.Will@willpowerinfo.co.uk>
Date: Mon, 3 Dec 2007 15:23:11 +0000
Message-ID: <9nYNM6Lf9BVHFALN@mail.willpowerinfo.co.uk>
To: SKOS <public-esw-thes@w3.org>

On Mon, 3 Dec 2007 at 15:12:43, Bernard Vatant 
<bernard.vatant@mondeca.com> wrote

>> It seems to me that "Children custody" represents the same concept in 
>>each of these cases - something like "care of and responsibility for 
>>children". The cases are different because the strings represent 
>>compound concepts, combining the concept of "children custody" with 
>>two other concepts in each case.

>Well, indeed it could be seen that way in this very case ... but in 
>theory only. Except that in this very case (keeping in mind those 
>examples are actually adapted from our customer vocabulary, which 
>happens moreover to be in French ...), the compound elements are not 
>defined independently in the vocabulary, e.g., "Children custody" is 
>*not* defined in the vocabulary as a separate context-independent concept.

Well, perhaps it should be ... Perhaps it is implicit, though not 
explicit.

Failing separate definitions of the components, then you have to live 
with the vocabulary containing cumbersome strings as labels for complex 
compound concepts.  You still need some rules for constructing these 
strings, which in effect specifies a citation order for combining their 
component concepts, and so recognises the existence of the components. 
Otherwise you could have one string that reads

Contentious divorce: Temporary arrangements: Children custody

and another that reads

Children custody: Definitive arrangements: Contentious divorce

or any other permutation.  That way madness lies.

>Identity issue is independent of the prefLabel unicity issue. Even if 
>the prefLabel is unique in the concept scheme, it's not used for 
>identification in SKOS framework (no more than in any RDF framework). 
>URIs, certainly constructed on unique codes/notations, are there for 
>identification purposes, at least for machines. In the original 
>thesaurus perspective, the unicity of prefLabel is a way to express 
>that it conveys identity, because there are no other identification 
>process in a thesaurus. But since identification is not conveyed by 
>prefLabel for machines, what are the other applications of this 
>unicity, beyond disambiguation for humans?

If concepts are identified by some other means, then the whole purpose 
of labels is to make the interface between humans and machines work 
effectively - to allow machines to understand queries that humans 
express in words. Do we need to look for any other reasons for having 
unique labels beyond ensuring that the labels are not ambiguous?

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 Monday, 3 December 2007 15:23:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:38:58 GMT