Newsgroups: comp.ai
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!gatech!news.mathworks.com!news.kei.com!uhog.mit.edu!news!news
From: Pushpinder Singh <push@mit.edu>
Subject: Re: Does anyone use JAVA for AI programming?
Content-Type: text/plain; charset=us-ascii
Message-ID: <31F15646.5B6F@mit.edu>
Sender: news@media.mit.edu (USENET News System)
Reply-To: push@mit.edu
Content-Transfer-Encoding: 7bit
Cc: push@mit.edu
Organization: MIT
References: <4sk4u7$hkc@geraldo.cc.utexas.edu>
Mime-Version: 1.0
Date: Sat, 20 Jul 1996 21:57:26 GMT
X-Mailer: Mozilla 3.0b5a (Win95; I)
Lines: 18

Jason wrote:
> 
> What is the opinion of those-in-the-know here on the future or
> appropriateness of doing AI work in Java?

Perhaps the best reason for using java in AI work is its portability.
Imagine being able to include implementations and test cases _within_
your papers (assuming they are published on the web.)  It would
certainly make it easier to test other researchers' claims, and if code
or libraries are provided, to build on their systems.  I find it
annoying to read papers that don't include enough information to
replicate the results they describe.

Personally I like Java but I wish java environments included lisp-like
listener windows in which you could interactively test pieces of your
code.

-push
