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

Re: SKOS properties

From: Sue Ellen Wright <sellenwright@gmail.com>
Date: Fri, 27 Apr 2007 09:22:58 -0400
Message-ID: <e35499310704270622j13af83e7i91f42244cd40d1a8@mail.gmail.com>
To: "L.Si@lboro.ac.uk" <L.Si@lboro.ac.uk>
Cc: public-esw-thes@w3.org
Hi, All,
Of course, I agree with the rdf discussion -- this works. But I'm also
interested in the idea of exploring antonymity on a cline of equivalence --
this is similar to some of what we do in terminology management. Libo Si
cites: exact match, broadmatch, narrowmatch, majormatch, minormatch -- as
degrees of equivalence. We (the terminologists in ISO 12620) have a rather
vague "degree of equivalence", plotted on a scale of 1-9, with 1 being exact
match and 9 probably being out to the right of "minormatch" -- maybe
"minuscule match" (I'm just making this up at this point). Conceivably one
could plot a somewhat artificial point where there is a break in the cline
and you revert to 0 (ostensibly nullmatch, but with the agreed assumption
that 0 implies antonymity to some degree). Of course "not-equivalent" is not
the same thing as antonymous. Black and magenta are not equivalent, but we
don't think of them as antonymous. So there'd have to be a clear distinction
between "not equivalent" and "antonymous". One of the problems with
cline-like scales is my poison cookie analogy (heh, everybody has to have
something people remember them for). I have a nice plate of cookies
(multilingual conceptual equivalents). They are very attractive and you are
invited to take your pick. Oh, and by the way, one of the cookies is
poisonous (i.e., it has an equivalence value of, let's say, an 8 as opposed
to 1 or 2). But I don't tell you what there is about the cookie that makes
it an 8, so you are in the dark as to how to recognize, or whether you could
go ahead and eat, that cookie (use that term) and not suffer unpleasant
consequences. This is where the scope note (we have something called
"transfer comment" that is specific to translation equivalents) comes in. A
human user can interpret it, and it could conceivably be configured in such
a way that a parser could also control that problematic entry as well,
provided that we came up with a good way to mark the discrepancies.
Bye for now
Sue Ellen

On 4/27/07, L.Si@lboro.ac.uk <L.Si@lboro.ac.uk> wrote:
> Hi everybody,
> Yes, in a lot of cases, in a particular thesaurus, we need to think about
> the antonym relationship between concepts, which refers to SKOS-Core.
> However, as Bernard Vatant indicate, we can treat the antonym relation as
> a
> sub-property of SKOS:related or SKOS:synonyms.
> what i want to mention is that for SKOS-Mapping, in most cases, we really
> need to declare antonym relations between different concepts from
> different
> vocabularies, for example, antiradar system Vs. radar. in this case,
> whether
> we need to develop a new relationship called "SKOS:antonyms" or donot
> create
> any conceptaul connections between them is worth considering. Because in
> SKOS-Mapping, we can find the mapping relationships defined are just based
> on equivalence between concepts, such as, exact match, broadmatch,
> narrowmatch, majormatch, minormatch. It is difficult to derive a antonym
> relation from these defined inter-mapping relations.
> Kind regards.
> Libo Si
> Departmant of Information Science
> Loughborough University
> UK

Sue Ellen Wright
Institute for Applied Linguistics
Kent State University
Kent OH 44242 USA
Received on Friday, 27 April 2007 13:23:10 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 13:32:08 UTC