Re: Vocab format

From: Norman Gray <norman-at-astro.gla.ac.uk>
Date: Sat, 15 Sep 2007 12:10:31 +0100

Doug, hello.

On 2007 Sep 15, at 03:55, Doug Tody wrote:

> But when we
> search for data for analysis we might be interested in all "Galaxy"
> type objects. If this is what is meant by broader/narrower type
> relationships, then it is a requirement for this to be useful.

It is. The broader/narrower relation in SKOS and similar contexts is specifically related to retrieval, with the relationship between the thus-annotated terms being that any item retrieved via the narrower term _should_ (as I understand it, this is not as strong as _must_) be one of the items retrieved via the broader term.

Thus 'bank' is a broader term than 'bank account', even though it is obviously not the case that a 'bank account' is a more specific type of 'bank' -- this is not the logical subClassOf relationship, and I think it's not even necessarily a partOf relationship.

We can be collectively reassured, therefore, that using SKOS would not deprive us of things to argue about.

Best wishes,

Norman

-- 
------------------------------------------------------------
Norman Gray  :  http://nxg.me.uk
eurovotech.org  :  University of Leicester, UK
Received on 2007-09-15Z13:11:01