This is message 1 of 3 from Martin Bryan, forwarded with permission.
I'm posting these because I think they inform the current conversation,
coming from one of the three original Topic Map designers. I note that
Martin currently has some sort of proposal for facets in front of the
ISO committee, I believe based on work he's doing for the European
Parliament, so it's no surprise he's enthusiastic about a discussion
of the subject here.
-------- Original Message --------
Subject: [topicmapmail] Two Models of Facets
Date: Tue, 2 Dec 2003 09:12:52 -0000
From: Martin Bryan <martin-akjmzRsnkgGnhEYp0SPQIA@xxxxxxxxxxxxxxxx>
To: m.altheim <m.altheim-aXtgmyahLfa1Qrn1Bg8BZw@xxxxxxxxxxxxxxxx>
CC: pbrown <pbrown-4UceUaxaNkJ14INtzM54zw@xxxxxxxxxxxxxxxx>, peter
<peter@xxxxxxxxxxxxx>
Murray
Peter pointed me to the fact that facets had raised their head on the
topicmapmail list again. I haven't time to register for the list at present:
I just popped in to look at the archive for a moment. But for the record,
I'd like to point out that neither Model A or Model T is a correct
interpretation of what facets were designed to do in ISO 13250.
Model A describes, more or less, the compromise that came out of the
discussion.
Model T describes, more or less, the starting point of what I wanted facets
to achieve.
How did we get to this position?
Facetted classification works by subsetting a topic so that those things
that need to be seen at the same time can be grouped together. Ideally the
facetting will be multilayered, but that requires a nesting mechanism, but
with SGML/XML you can't do nesting with attributes, only with elements. So
what's the alternative?
What we ended up with is a compromise: "multiple single level facets". You
can then say that if an occurrence (or a topic) has this characteristic AND
that characteristic it should be included in the set to be presented to the
user. So, for instance, if the book was published after 1/1/2002 and is in
English and is available on-line then it should be added to the list of
presentable occurrences on this topic. This is not a chain of nested
subsets, but the intersection of sets of properties.
Remember also that one of the design criteria of ISO 13250 was to allow
users to make an existing element set that fitted clear rules generate a
topic map on the fly. Therefore we wanted a way of utilizing attributes of
existing elements as properties within a topic map. This would allow us to
use a reference to an occurrence type such as <book published="1/1/2000"
xml:lang="EN">Topic Maps are fun</book> as input for a topic map, converting
selected attributes to facets.
It is the intersection of facets that is the key. Only recently has anyone
come up with a way of nesting facets effectively electronically (with OWL).
Back in the last century (yes, that's when Topic Maps were designed) the
best compromise we could manage was "intersecting facets". That is what ISO
13250 was seeking to achieve though, as it was one of the last features
added, we didn't explain it very clearly.
Martin Bryan
IS-Thought: Thinkers for the Information Society
29 Oldbury Orchard, Churchdown, Glos. GL3 2PU, UK
Phone: +44 1452 714029 Fax: +44 1452 859991
E-mail: martin-akjmzRsnkgGnhEYp0SPQIA@xxxxxxxxxxxxxxxx
--
Murray
......................................................................
Murray Altheim http://kmi.open.ac.uk/people/murray/
Knowledge Media Institute
The Open University, Milton Keynes, Bucks, MK7 6AA, UK .
[...] all matters of authority and responsibility are ultimately
matters of social practice, and never matters of ontology (that
is, never just a matter of how things in fact are in the nonhuman
world). [...] just as we should not look to ground our moral
judgments in the nonhuman authority of a god, so we should not
look to ground our empirical judgments in the nonhuman authority
of an external world. -- Robert Brandom
http://www.tilgher.it/brandom.html
|
