Lambda the Ultimate

inactiveTopic A Retrospective on Paradigms of AI Programming
started 4/30/2002; 7:15:08 AM - last post 5/2/2002; 2:39:14 AM
Chris Rathman - A Retrospective on Paradigms of AI Programming  blueArrow
4/30/2002; 7:15:08 AM (reads: 947, responses: 2)
A Retrospective on Paradigms of AI Programming
Peter Norvig has some updated thoughts on his book "Paradigms of AI Programming" after 10 years.

Paints a rather bleak picture of programming language diversity in the the commercial marketplace - with Lisp adoption suffering. The "52 most important lessons in PAIP" makes it worth the read (but then I'm one that likes lists).
Posted to general by Chris Rathman on 4/30/02; 7:16:25 AM

Patrick Logan - Re: A Retrospective on Paradigms of AI Programming  blueArrow
5/1/2002; 3:05:07 PM (reads: 920, responses: 0)
Two things I've noted about this.

(1) Lisp's influence is in every language being used instead of Lisp for AI or any other kind of programming.

(2) Although some of these techniques are still kicking and screaming their way into more common usage, it is still very difficult and/or expensive to use them. It is a shame some of these AI techniques that are 10, 20, 30 years old are not more a part of the average programmer's toolkit.

I would summarize these two points by saying...

We still like bits more than lists.

Ehud Lamm - Re: A Retrospective on Paradigms of AI Programming  blueArrow
5/2/2002; 2:39:14 AM (reads: 904, responses: 0)
I always objected to claims equating AI with Lisp and/or Prolog. I understand the (historical) source of these claims, but they do harm both to the filed of AI (e.g., when it comes to statistical methods Prolog is not really such a great choice) and to the programming languages themselves (e.g., "Prolog? It is only good for doing AI. Nobody does AI, he he he").

The connection of Lisp and Prolog (via DCGs and meta-programming) to programming languages research is much more substantial IMHO.