[Somewhat Off Topic] AI Contest

Alex Martelli aleaxit at yahoo.com
Mon Aug 6 10:57:06 EDT 2001


"François Pinard" <pinard at iro.umontreal.ca> wrote in message
news:mailman.997103177.15032.python-list at python.org...
> > > With a broad enough perspective, FORTRAN is a good AI language! :-)
>
> [Alex Martelli]
>
> > We mostly used PL/I (and Rex for prototyping, way before it acquired
> > that second 'x':-),
>
> Did you ever use one of those lenient, forgiving, auto-correcting PL/I
> compilers?  Whenever you made some error, the compiler wrote "Assuming
> that ..." and moving on with compilation.  It amused us to think that you

Yes, and it was terrible... at least Rex's attempts to guess what I meant
were reasonably predictable and sensible, as I recall...!-)

> could throw it random garbage and have the compiler some PL/I program out
> of it! :-) That was surely an attempt at artificial intelligence! :-) :-)

An attempt to give AI a bad name (not that it needed that...:-).

> > Fortran was better for some of our work (vectorializing & parallelizing
> > some core parts of the computation [...]
>
> Those vectorizing compilers were surprisingly good, indeed.

As opposed to the attempts to automatize parallelism, which never
worked out just right in my experience (time frame: late '80s, on
IBM machines -- multi-CPU 3090's chock full of VFs -- but at the
time I heard similar user experiences from competitors' customers).

The 'automatic' parallelizers needed so much hand-holding and
source tweaking to work halfway-right that it was much less work
to structure the parallel-computation aspects explicitly (well, on
a FEW big CPU's, at least; I had no real experience with arrays of a
LOT of middling CPU's).


> P.S. - The best approach to AI programming, in my opinion, is to never
loose
> ground.  Some people, when confronted to fuzzy logic, inference engines,
> expert systems, continuations, provers, etc., become so excited that they
> merely give up on their natural intelligence.  This does not yield AI.

I'm not sure what continuations (such sweet little things they are!-) are
doing on this list, actually:-).


Alex






More information about the Python-list mailing list