[ann] Minimal Python project

[Edward K. Ream]

"holger krekel" <pyth at devel.trillke.net> wrote in message
news:mailman.1042312221.11913.python-list at python.org...
> Edward K. Ream wrote:
> > > I'm inclined to withhold judgment on such claims, admitting the
> > > possibility that I don't know everything.
> >
> > A better rule of thumb, IMO, is that extraordinary claims require
> > extraordinary justification.
>
> We probably should have said 'hope to' instead of
> 'are confident' regarding speed prospects.
>
> Edward, i appreciate your scepticism as it reminds me
> that we are mostly doing this because we want to have fun
> while trying it.  Raising too many expectations can be
> dangerous.  Now, rumor has it that Armin Rigo will
> also do it as part of his new job and he is (still :-)
> confident so hopefully, 'eventually' this will work out.

Thanks very much for this clarification.  Fun is one of the most powerful,
creative and useful emotions there is.  I think it is one of the very best
reasons for doing any project :-)

Perhaps I wasn't clear enough myself in my original posting.  My attitude at
present is a mixture of a) extreme interest and b) moderate skepticism.
Given the potential payoffs, this is a project that ought to be pursued.

Let us suppose for the moment that it is indeed possible to improve the
speed of the C interpreter using some kind of GIT applied to Python code.
Obviously, this would be extremely important to all Python programmers.
Moreover, such a strategy might be important in the larger world.  For
instance, it might suggest new kinds of hardware design: maybe novel caching
techniques or new machine architectures oriented towards higher-level
languages.  Such ideas were tossed around decades ago; perhaps their time
has come again.  Certainly compiler vendors would be very interested as
well.

These implications might be interpreted more pessimistically.  Designers of
hardware, gits and compilers have put a lot of time, thought and money into
their work; perhaps the fact that run-time optimizations have not yet been
spectacularly successful is some clue that they can't be made so.  OTOH,
perhaps Armin is on to something that will revolutionize many parts of the
computing world.

But these are all speculations.  The intent of my original posting was to
get as much information as possible about _how_ and _why_ this project can
speed up Python's interpreter.  I focused on my doubts because they point to
the kinds of questions I would like to see answered.  To repeat: I eagerly
await a more detailed presentation of the technical details.  Some things
I'd like to see in the presentation:

- a detailed overview of the design of the GIT
- a detailed overview of the actual code of the GIT
- detailed examples of how to speed up the interpreter, with real code
examples
- as many real-world statistics as possible
- a discussion of the limitations of the technique, if any
- anything else that would help us understand how and why the optimizations
are effective.

And as long as I am asking, I would like to see a real technical paper, not
just PowerPoint slides.  Thanks again, Holger, for your good humored answer.

Edward
--------------------------------------------------------------------
Edward K. Ream   email:  edream at tds.net
Leo: Literate Editor with Outlines
Leo: http://personalpages.tds.net/~edream/front.html
--------------------------------------------------------------------