Re: Rich text descriptions

From: srt@pbm.com

Date: Mon, 18 Jan 1999 00:00:00 +0000

(From a discussion on the PBM Design forum on how to output good text
descriptions.)

ARENA (http://www.pbm.com/arena) does a pretty good job with this, and one of
the techniques it uses is essentially a thesaurus.  The idea is to represent
each specific concept within your program as a single identifier, construct
output using those identifiers, and then run it through a process that
substitutes synonyms for those identifiers.

For example, suppose I want to describe someone who has a lot of native
intelligence.  I pick an identifier for that attribute -- "wily".  Now
whenever I want to describe someone that way, I make sure to use "wily".
Then I build a thesaurus:

{"wily","clever","cunning","crafty","sly","tricky"}

I can now run my output, match the "wily" in the output against the "wily" in
my thesaurus, and substitute any of the other synonyms.

You can go a long way with this technique.  I leave it to the reader to expand
and figure out some of the more clever usages.

Check out http://www.cogsci.princeton.edu/~wn/w3wn.html for an on-line
thesaurus database organized this way.

-- Scott T.

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own