Re: Discussion issue: should semantic relation statements be members of concept schemes?

I've been away for a few days, so was not able to participate in this 
discussion earlier, but some points occur to me in connection with some 
of the issues mentioned.

1. There is a risk of ambiguity if you use the expression 
"hasTopConcept", because it seems that you intend it to be an attribute 
of a _scheme_, equivalent to "contains as a top concept".

In traditional thesaurus terminology, "has top term" (symbol TT) is used 
as an attribute of a _term_ (i.e. of a concept)

term 1
TT: term 2

means "term 1 occurs in a hierarchical tree of which term 2 is the top"

2. In principle, semantic relationships should be independent of concept 
schemes, so long as the concepts are defined precisely and identically 
in all schemes, and so long as the only relationships specified are 
those which are generally valid and not context-dependent. Some steps in 
a hierarchy in one scheme may be missing from another scheme, but the 
direct or indirect BT/NT relationships should still be valid.

This is a counsel of perfection, however, and in reality concepts are 
often ill-defined and their definition has to be inferred from their 
relationships. There is thus a chicken and egg situation. To avoid 
ambiguity we have in general to say which scheme a relationship comes 
from (except perhaps in the case where all schemes under consideration 
specify the relationship identically).

3. In message 
<350DC7048372D31197F200902773DF4C05E50B79@exchange11.rl.ac.uk> on Wed, 4 
Aug 2004, "Miles, AJ (Alistair)" <A.J.Miles@rl.ac.uk> wrote
>
>If we made skos:hasTopConcept functional, each scheme would have to be
>defined with a single root concept ... do you think it's worth doing it that
>way?

If you constructed hierarchies with true generic relationships, any 
single root concept would be so general as to be useless. If you work up 
a hierarchical tree and group concepts as far as possible, you generally 
end up with top terms which effectively define facets, such as 
"organisms", "events", "abstract concepts", "places", "objects", and so 
on. The only broader term common to all of these is something like 
"concepts", which is not helpful.

4. There are different interpretations of "microthesaurus", as Stella 
says, but I think of them as selections of terms relevant to a 
particular subject field, drawn from a more general thesaurus. They are 
helpful to people wanting to browse terms in that field, but I don't 
think it is necessary or helpful to bother about top terms within a 
microthesaurus. You can have a microthesaurus for clothing, and many of 
the terms within it, like "shirts" may have broader terms like "textile 
products" that do not fall within that microthesaurus. The judgement of 
what is relevant to a subject field is necessarily subjective, and users 
should be able to navigate up a tree even if it takes them outside that 
selection.

In fact there is some inconsistency in the UKAT use of microthesauri, 
and in some cases they have drifted into defining them, or some of their 
contents, more in line with facet definitions, so that, e.g.,  they have 
"events" now in just one microthesaurus, rather than each event being in 
the microthesaurus covering the subject to which it relates. It's better 
to stick to one approach or the other (or indeed both, but separately).

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 Saturday, 7 August 2004 22:14:55 UTC