SUO: Re: Focus and Volume
Hello John and all.
As a newcomer to this forum, I will be the shortest possible - but not
shorter - since I understand that focus and volume on this list are a deep
problem. As a practicioner at lower and middle level, I'm glad to read
things like :
[John Sowa]
> Bottom line: I agree with Pat that the upper level should be
> distilled from the lower level, but the process of creating
> the full lattice should be done by an automated method.
A quick description of my context, and a proposition of collaborative tool
and methodology, available for the community.
I'm working in the Topic Maps universe, with a software enabling
collaborative building of Topic Maps (see reference links at the end of
message). I precise what I mean by that, for people here not familiar with
Topic Maps concepts: The collaborative on-line tool allows editors to add
topics in a data base, defined either by a name or subject indicator
(external ressource) or both, and associations between topics. Topics play
roles in associations, and of course these roles are themselves defined as
topics of an upper level. The resulting data base has a graph form, in which
one can browse (through prepublished XML files), following the links defined
by roles in associations.
On such a basis, we are building a data base of ressources for the Semantic
Web, Topic Maps, Standards and correlative attractor, which I have called
the "semantopic universe". This data base is linking all levels, since we
have "individuals" (people, tools, projects, events, companies,
organizations) and classes of individuals, and middle-level "themes" or
"concepts" these individuals deal with, like "ontology" "knowledge
organization" "computer science". Of course this middle level can and has to
be structured by associations at an upper level, standardized or not ...
I would be very interested to see people here use this tool - the
"semantopic project" is open and collaborative - and add topics and
associations at the level they consider the most pertinent or central - BTW
I would buy more easily a "core" or "central" ontology than an "upper" one -
and link them by whatever associations they find pertinent.
What for? We are working with graph theorists on ways to extract from big
Topic Map Graphs the most pertinent or central topics and associations by
aggregation tools based on local density of nodes and edges. That looks to
me like an approach of the "distillation" John Sowa is speaking about.
If people here are interested by this approach, which looks to me more
constructive - not to say constructivist - than endless e-mail debate :) -
have a look at the following links and tell me what.
Thanks for your attention.
Bernard
The semantopic project is living at : www.universimmedia.com/semantopic.htm
Graph theory and Topic Maps:
http://www.gca.org/attend/2001_conferences/europe_2001/topicmaps.htm
-------------------------------------------------
bernard.vatant@mondeca.com
Mondeca - "Making Sense of Content"
www.mondeca.com
-------------------------------------------------