FW: Global concept identification and reference

Fwded discussion with Leonard:


-----Original Message-----
From: Leonard Will [mailto:L.Will@willpowerinfo.co.uk]
Sent: 05 November 2004 12:20
To: Miles, AJ (Alistair)
Cc: Stella Dextre Clarke (E-mail)
Subject: Re: Global concept identification and reference


In message 
<350DC7048372D31197F200902773DF4C05E50CE3@exchange11.rl.ac.uk> on Thu, 4 
Nov 2004, "Miles, AJ (Alistair)" <A.J.Miles@rl.ac.uk> wrote
>Hi Stella, Leonard,
>
>I'm beginning to think that a major obstacle to using thesauri & similar
KOS
>types within a networked/semantic web environment is going to be the lack
of
>globally unique identifiers for concepts.
>
>Being able to uniquely reference a concept within a global information
space
>is the foundational requirement of using thesauri as part of the semantic
>web.
>
>I thought I'd write you something at this early stage of my thoughts, to
>hear your initial response.  The current part 2 draft of the BS standard,
as
>I recall, makes no mention of unique identifiers for concepts (beyond
>descriptors), and no mention of using URIs as globally unique identifiers
>for concepts.  Being a significant normative standard in this area, you
have
>the potential to greatly influence practise in this field.  But as yet I
>have no concrete suggestion about how to take this forward.
>
>What do you think about this?

Alistair -

The following are my personal views - Stella may have other ideas, but 
she is on holiday at the moment and may not have easy access or time to 
deal with email.

Perhaps I am misunderstanding what you are trying to do, so perhaps we 
should clarify first whether

(a) you are seeking a single definition for each concept that can be 
accepted and applied universally, so that you can build useful semantic 
networks incorporating that concept

or

(b) you just wish to have a way of pointing unambiguously to a specific 
definition of a concept within a specific controlled vocabulary.

If (a)

Thesauri do not generally attempt to provide definitions of concepts 
that will apply in "a global information space". In fact they generally 
explicitly disclaim any attempt to do this, by saying that the scope 
notes and relationships they provide apply only to the restricted 
definitions adopted for the purposes of organising information using 
that specific thesaurus (or more generally "structured vocabulary").

A concept defined for use within one thesaurus may overlap with or 
contain one or more concepts which are defined as separate concepts in 
another thesaurus. Neither thesaurus is wrong, but they may differ 
because they were designed for different applications or just because of 
the necessarily subjective decisions that were made differently by their 
creators. Are you saying that for semantic web purposes you want to 
select a single definition for each concept? That might work for 
individuals, as in FOAF, but it seems impossible to generalise; you 
cannot choose a single definition that everyone will accept for "drug", 
"peace", "God", "steel" or "tree", for example. All you can do is to 
point to the definition of these concepts as used in a particular 
thesaurus, as in (b).

If (b), you said in another message:

> The mechanism for unique identification of concepts within a print 
>environment is traditionally via the preferred term (or 'descriptor') 
>for that concept, which is a unique term within a thesaurus.  The 
>combination of the preferred term for a concept, and a URI identifying 
>the thesaurus, therefore provides a globally unique description of a 
>concept.

Yes, I agree with this, though it is not restricted to the print 
environment.

>reference by description for concepts as described above depends on at 
>least two properties (e.g. combination of skos:prefLabel and 
>skos:inScheme)

Yes, this follows.

>for which implementations would depend on the expression of identity 
>rules.

Sorry, I don't know what this means or what problem it causes. I thought 
from previous discussions that options were proposed for referring to 
the two elements above by giving a URL for the thesaurus followed by a # 
or a / and the descriptor. That seems a reasonable way of combining the 
two properties required.

You can use "concept numbers" or "term numbers" within a thesaurus as 
well as descriptors, though I have not personally found them to be 
necessary or helpful because of the difficulty in defining when a 
concept has become a new concept or has just changed in some way, as 
discussed by Stella earlier. These numbers are mentioned in the draft BS 
specification, paragraph 14.5.

>When describing best practise for creating RDF descriptions of thesauri 
>without official URIs, do we ...
>
> (a) attempt to remain neutral about whether people make up unofficial 
>URIs, and rely on the owl:sameAs machinery to cope with multiple 
>published URIs for the same concept,

Sorry again, but I don't know how a URI becomes "official" - can you 
enlighten me?

When you talk about "multiple published URIs for the same concept" do 
you mean "different ways of pointing to the same concept within the same 
thesaurus" or "pointers to concepts within different thesauri that 
appear to have more or less the same scope"?

Regards

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, 10 November 2004 12:15:12 UTC