Date: Mon, 3 Apr 2000 10:24:19 -0700
Mr. Rod Welch
rowelch@attglobal.net
The Welch Company
440 Davis Court #1602
San Francisco, CA 94111 2496
Subject: | Advancing KM Toward Intelligence |
From: "Jack Park"
Rod,
...of that discussion is an absolute requirement. I further believe that the
topic map should be both textual (like an index), and graphical in
presentation. One thrust of my personal research has been to find ways of
automating the construction of topic maps.
Presently, I have turned to the development of a Literate Programming
environment that allows one to jump gracefully between several forms of
representation. For instance, things which can easily be created
graphically, such as concept maps and so forth, are then mapped -- through
xml tags -- back into the original text material. Source code for software
to implement things can also go into that textual material.
I am using a subset of DocBook for this purpose, with extensions necessary
to allow for the inclusion of other representations. My tag (algorithm) is
then followed by other tags like (graph), (code), and so forth. It is then
a matter of development of interpreters for each of the tags. This
follows, as close as I understand Literate Programming, to Donald Knuth's
original idea.
Sincerely,
One further interesting observation I just made: you call something Welch
linking. I am profoundly concerned that you might be staking a claim for
linking of text as a form of intellectual property belonging to you. I do
hope that is not the case.
I believe that a richly linked dialog, as you have on the web, is a
beginning; a very good one at that. I also believe that anchoring the
semantics of such discussion in a concensus ontology is desirable. Now, I
wish to add that a topic map...
This Literate Programming is the approach I am taking with TSC, and it is
one that I tend to advocate for the OHS user interface. No doubt, there
will be contributed ideas that make it more powerful than it now appears to
be.
Jack Park
jackpark@thinkalong.com