[topicmaps-comment] Topic Map Designer and its ilk

[David Dodds <drdodds42@hotmail.com>]

Gen and Readers

I find a lot I liked in what Gen had to say. In the recent WROX book XML 
Meta Data I included a discussion on a topicmap constructor and how it might 
produce a topic map from a picture (which was represented in XML SVG and had 
embedded RDF metadata in the picture itself).

I can relate to what Gen says about hand creating maps. I too would like to 
see what in AI is called a knowledge engineer's editor. Such a suite of 
software provides an environment to the user such that(s)he can 
"encode"/enter knowledge and information, obtained from his own mind and 
from discussion with domain (topic) experts, and or expert level 
"users"/"doers" in that domain/practice.

In the UNIX world there is the Swiss Army knife called EMACS and it is 
programmable (if you know or will learn Lisp).

I think it would be swell if there was an Emacs-for-Topic-Maps program / 
environment. A visual GUI would be good, drag and drop graphically oriented 
graph-knowledgeable (graph-aware) visual metaphors
to visually construct a picture/diagram of one's intention and let the 
system build the topic map code to represent that picture/diagram. I 
discussed this notion in the book some. It would be useful if we could get 
some folks to volunteer some time to work on trying out "engines" which use 
one or more of the interchange DTD/Schemas for TM that have been proposed, 
and marry them with something like Graph Manipulation Language. (As I said 
in the book, once we have a visually operated (mouse) graph based "designer" 
then we would be in a position to morph graph structures like XTM and RDF 
back and forth!

Gen's point about a reference set of terms (dictionary) and the 
shading/modulation/CONTEXTUALIZATION of words [from that set] in 
conversation / communication is part of what Ive been trying to say about 
"situated context processing" and also about one way of selecting one (or 
several particular) piece of uncooked spaghetti from a handfull.
(my visual metaphor for the situation of a topic having multiple instances 
of subjectindicator refs (like English and Swahili example).

David Dodds



>From: Gennadi Bedjanian <gbedja@dsnet.net>
>Reply-To: Gennadi Bedjanian <gbedja@dsnet.net>
>To: topicmaps-comment@lists.oasis-open.org
>Subject: [topicmaps-comment] Topic Map Designer and its ilk
>Date: Thu, 18 Oct 2001 14:34:02 -0400
>
>Several threads were triggered by Ivan's "Can subjectIdentity elements
>guarantee topic identity?"
>Here's another viewpoint with a new bent and a hint at a solution.
>
>The discussion about "eternal/changing subject identities", "one
>notion - one topic subject", "situatedness" and on, reminds of the
>language - speech dichotomy. Dictionary meanings of words belong to
>the static language system, while in speech words acquire "situated"
>meanings. They influence each other but never fully coincide. In the
>dynamic speech realm individuals do things together, verbally
>communicate to do things more effectively, and arrive at accepted
>meanings.
>
>If a participant in a communication (= topic map) has a wrong
>understading of a meaning (= topic subject) then something happens
>that makes him rectify the fallacy. For a major misunderstanding it's
>communication breakdown, for a minor - some kind of built-in
>redundancy sustains the communication while the misconcept is
>corrected. Collective communication created the natural language with
>its dictionary and speech as a means to keep the dictionary up-to-date.
>
>As far as I know, TMs do not interact a lot nowadays. And a TM without
>interaction is like its author talking to himself. Assigning topics to
>subjects in unassailable solitude. Creating his own dictionary.
>
>Hence the importance of creating communities of TM producers who
>exchange TMs and arrive at consensuses on not-so-clear issues. Not
>through human negotiations but the TMs themselves filter out what is
>irrelevant. That's the "market principle" that blazes the trail for
>science. A sort of Japanese way with AI. Let's first put it into
>washing machines and gadgets, and eventually it may grow into
>something big. Meanwhile we'll have fun with "intelligent" cellphones,
>household appliances and so on.
>
>A boost to such communities would be a TM / knowledge editor. Creating
>a TM manually is too laborous for mass production of TMs. It must be
>done automatically in the background. The software must follow the
>user and do all the dirty work. The researcher / user should be able
>to draw conceptual (topical) structures, work with texts, tag-extract
>words-concepts right from the text and see visualizations of evolving
>models based on the TM that he is constructing (BTW without
>necessarily knowing TM specs).
>
>It may sound far-fetched but many of these features are already
>implemented in Topic Map Designer. It's not a finished product, but
>it's the right start. Ronald Heckel wrote it for his diploma. As
>usual, it takes a student.
>http://www.topicmap-design.com/en/topicmap-designer.htm
>
>Why there is no such software? Or maybe I don't know about it? Not
>demos of industrial systems, but something handy.
>
>To pick up the language analogy, it took a lot of talking before the
>first dictionary with static words appeared. And even its entries
>change, only gradually.
>
>I'm not sure if Topic Maps were intended as a tool for masses, but I
>have a strong suspicion that their full potential cannot be realized
>without the Internet collective thinking.
>
>Gen Bedjanian
>
>
>
>----------------------------------------------------------------
>To subscribe or unsubscribe from this elist use the subscription
>manager: <http://lists.oasis-open.org/ob/adm.pl>


_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp