Re: [SKOS]: [ISSUE 44] BroaderNarrowerSemantics

On Mon, 14 Jan 2008 at 22:24:14, Stephen Bounds <km@bounds.net.au> wrote
>Can I chime in with a brief practical example?  I'm currently doing some
>work with Keyword AAA (a thesaurus commonly used in the Government
>of Australia) and chose to use SKOS because it was simple and value-
>neutral.

Be careful with Keyword AAA, because although it calls itself a
thesaurus it is in fact a classification scheme. It uses the expressions
"broader term" and "narrower term", not in the thesaurus sense of
generic (or partitive or instantial) relationships, but just to indicate
which concepts may be combined or pre-coordinated to express complex
concepts. For example it has

COMMITTEES
    broader term
        COMPENSATION
        FLEET MANAGEMENT
        LEGAL SERVICES
        etc.
    narrower term
        Agenda
        Applications
        Archives
        Ethnic affairs
        etc.

This does not mean that the general concept of committees is a specific
type of fleet management, or that ethnic affairs is a specific type of
committees. It just means that within the scheme it is valid to
pre-coordinate these concepts into strings such as

FLEET MANAGEMENT : COMMITTEES

or

LEGAL SERVICES : COMMITTEES : Agenda

>To date, my experience using SKOS has been very positive -- the loose
>semantic rules around SKOS make it very quick and easy to do useful work
>with it.

As far as I know, SKOS has not yet been developed to represent
pre-coordinated strings of concepts of this type. It would be a gross
distortion of the relationships to implement Keyword AAA's use of BT/NT
as though these were valid thesaural relationships.

>In particular, the simplicity of the syntax makes it a viable translation target
>(I'm translating the XML export format from Tower Software's TRIM
>package into SKOS).
>
>As such, I am 100% in favour of trimming SKOS Core to the absolute
>minimum required (i.e. no transitivity).

I agree that we should not over-complicate SKOS by providing too many
options, and we have to consider who will be applying these. I'm not
convinced that many users will understand, or be concerned by, whether a
relationship is transitive or not. I have not seen a convincing
illustration of the need to represent intransitive BT/NT relationships,
or examples of intransitive relationships which still conform to
thesaurus standards. I would therefore prefer SKOS to assume that such
relationships are transitive until someone demonstrates the need for
more complexity.

The example of a mixed chain of relationships, using BTP + BTI and so
on, I agree is difficult, but again I doubt whether it would produce
results that would worry users. The underlying issue seems to me whether
we are trying to develop a scheme that is rigorously logical, for
machine use, or whether we are mainly interested in a format to encode
thesauri that exist and conform to standards.

Leonard Will

-- 
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, 14 January 2008 15:31:30 UTC