[Stackless] Re: Stackless Python Logic Variables?

Jim Bowery jabowery at technicalpursuit.com
Wed Jun 11 16:16:05 CEST 2003


----- Original Message ----- 
From: "Christian Tismer" <tismer at tismer.com>
> > James Bowery wrote:
> >>If anyone has the proof Christian Tismer has requested, this may be the
big
> >>chance to see an implementation of logic variables wend its way to a
large
> >>user base.
>
> Please, don't over-estimate that one of my statements. As I tried
> to explain, this is no final thought which I think every day.
> I was just captured as a snap-shot in a bad mood.

Well it was a fair request in my opinion.  You at least set up a standard of
communication to meet and the Oz/Mozart culture (of which I am not a member)
has enough academic savvy to throw at the communication problem to meet it.
It was a challenge that they should be ready willing and able to handle.

I'm coming at this from the perspective of a nearly 50-year old, flying at
about 10km altitude so you may not even want to talk to me directly although
it may not be as hopeless as you (or I) think.  I did have to start
somewhere.  (See http://www.geocities.com/jim_bowery/future.html for a brief
history.)

The Mozart guys face along with you the eternal question of what constitutes
"proof" and in fact this is largely what led me to initially (1981)
investigate relational programming -- well beyond anything like Codd's SQL
or Prolog -- as a paradigm.

Although Hoare has done some really good work in this area
(http://www.comlab.ox.ac.uk/oucl/users/tony.hoare/topdownz.ps) your
following comment lets me do what may be a better job than just throwing out
Hoare's relational paradigm as both "proof" of the general correctness of
the relational approach to parallelism and as a standard for "proof" in what
might be called computics as opposed to say physics.

> Sure, a good proof would be more than motivating, since all my
> friends know how much I love math...

Gian Carlo Rota, in his last book, "Indiscrete Thoughts", has a chapter on
"The Phenomenology of Mathematical Proof" that I love, particularly since I
worked with Lippold Haken on the PLATO Music Project back in 1977 when the
four color map theorem was proven by his father, and Lippold was getting in
trouble with PLATO authorities by chewing up the Cyber 6500's background
time on his own similar proof-by-verification of another conjecture (the
name of which escapes me).

Gian Carlo Rota says: "The example of the four-color conjecture leads to an
inescapable conclusion:  not all proofs give satisfying reasons why a
conjecture should be true.  Verification is proof, but verification may not
give the _reason_.  What then do we mean by _the reason_?"  GCR then recalls
the history of Bertram Kostant's "leap of faith" in finding "the _reason_
for the existence of the five exceptional Lie groups in the outer
automorphism of the orthogonal group in dimension 8 by a tour de force that
remains to this day a jewel of mathematical reasoning."  GCR concludes,
"...mathematicians are not satisfied with proving conjectures.  They want
the _reason_."

Now, what I just said would be gibberish to a lot of programmers -- but it
should not be gibberish to anyone with a love for math.  Proof is a matter
of culture as much as form.

> I anyway hereby sware to catch up with the current standards
> of knowledge, ASAIC.

I suffer from the 10km altitude problem but nearing 50 years of age I've
done a lot of walking around the same territory at the ground level and
still haven't found a better approach to the topography.  Relational
semantics, Hoare, too, has finally come to see is a good path, after all
these years of messing with CSP (which in his culture means "Communicating
Sequential Processes" -- not to be confused with the Mozart/Oz/Relational
Programming culture's acronym of "Constraint Satisfaction Programming" which
is actually closer to Hoare's current relational model than Communicating
Sequential Processes).

I hope the Mozart guys, who have serious ground truth in a very important
part of the "correct" path, have enough motivation to communicate their
OO/generator/consumer view of Mozart to you so you can see the value.  I'm
seriously handicapped by not really having their ground truth mapping
between paradigms.

> Anyway, I'd like to take back everything that made people not
> respond in the last weeks. Please accept that I'm very open
> for any kind of discussion, that I'm willing to thoroughly
> reply to any kind of serious request, and that I really will
> not run that buggy BOT any longer, quibbling under my name.

Don't underestimate GVR's willingness to keep Stackless around as a valuable
sandbox within which to explore important ideas for future versions of
Python.  He seems to rely on you for that for good reasons.


_______________________________________________
Stackless mailing list
Stackless at www.tismer.com
http://www.tismer.com/mailman/listinfo/stackless




More information about the Stackless mailing list