[sl4] JOIN: Hello, etc.

From: Gwern Branwen (gwern0@gmail.com)
Date: Sun Dec 07 2008 - 10:41:54 MST

Hash: SHA512

On Sun, Dec 7, 2008 at 7:47 AM, Philip Hunt wrote:
Version: GnuPG v1.4.9 (GNU/Linux)


> So have you started hacking the Eurisko codebase? (relatedly: Is it
> easy to get it to run on modern PCs running operating systems such as
> Linux?)

Actually, I'd like to know if Eurisko's codebase is available at all,
in any form. I've never been able to found even the slightest example
snippet of code (outside of the examples from Lenat's papers), and
other people report discouraging results. See
or http://www.intelligence.org/ourresearch/publications/GISAI/meta/glossary.html#gloss_eurisko

For example, "AM: A CASE STUDY IN AI METHODOLOGY" critiques Eurisko in
some detail, but does without apparently any access at all to
Eurisko's source as they seem to base everything on close analysis of
the reports - and that was back in 1981 *. If 27 years ago, fellow
academics couldn't get access...

On the other hand, this may not be unusual for the period. The authors
of that paper remark:

"Lenat has attempted (more than most other workers in AI) to render
his program available to public assessment, both by making it
available for running and by supplying such detailed appendices in his
thesis. The whole discussion in this paper could not have commenced if
Lenat had not provided this unusual level of documentation."

It's unclear to me what is meant by 'making it available for running,
but this would seem to imply the sources where available in some form
- as it was an interpreted program.

Besides, you probably wouldn't want to run it directly. Lenat wrote it
while at Stanford, and the Lisp Stanford researchers used was
Interlisp. There are very few working Interlisp interpreters for
modern PCs and OSs. (I've found exactly one:
http://blake.mcbride.name/software/lispf4/index.html ).

* Email me if you want a copy.


This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:03 MDT