AI and Python

[Peter Norvig]

What I meant was the crime is that people in AI spent years 
duplicating work that had already been done in OR (and perhaps the 
reverse as well; I don't know the history of OR well enough).  An 
example is Marcel Schoopers work on universal plans -- important 
work, but largely duplicative of OR work on policies.  It would have 
saved the field a lot of work if this had been known earlier.

-Peter

At 5:47 PM -0400 5/17/00, François Pinard wrote:
>Peter Norvig <peter at norvig.com> écrit:
>
>>  You make a good observation that AI and operations research have a
>>  large overlap in the problems they address: basically, what is the
>>  right program or policy to decide what to do in a complex, uncertain
>>  environment?  It is a crime that the fields developed independently with
>>  little interaction, but in the last five or ten years this has changed,
>>  and AI now encompasses OR techniques, and is much better suited for
>>  dealing with real-world programs.
>
>Surely no crime here.  That various fields synergise is undoubtedly a
>good thing.  What I did not like much, retrospectively, is all the fuss
>AI people were making all along.
>
>Can Python do AI?  I'm not sure what AI means.  OK, yes, Python can! :-)
>
>--
>François Pinard   http://www.iro.umontreal.ca/~pinard

_____________________________________________________
Peter Norvig  peter at norvig.com  http://www.norvig.com
home:650-852-1220  work:650-604-6207 cell:650-279-0474