[SciPy-dev] Is Python being Dylanized?

[Prabhu Ramachandran]

>>>>> "PM" == Patrick Miller <pnmiller at pacbell.net> writes:

    PM> Now Dylan and FL both used the idea that between some user
    PM> specification of types and a type inference engine one could
    PM> get great benefits without giving up the wonderful dynamicism
    PM> of runtime typing.

Absolutely.  I have heard that Guido thinks of Dylan as a very nice
language.  I wonder if he would be pleased at the direction weave is
thinking of taking. :)

    PM> The direction I was planning for the weave accelerate
    PM> backplane was exactly as Prabhu described...  A Python subset,

So atleast all of us here are almost on the same wavelength. :)

    PM> reasonablely described, some types supported, and extensible.
    PM> I think that I can detect "conforming" classes in the bytecode
    PM> compiler and then make 'em go really fast.  My true goal is to
    PM> let most of Python to be written in Python (accelerated into
    PM> C++).  Guido wants to get the core of the language smaller,
    PM> and this may be a way to help.  The Py-Python could use the

Indeed, you'd need to have a small core in C that completely describes
the API and then everything is a translation to this API either
dynamically interpreted or statically compiled.  Right?

    PM> HPP subset so that Jython could execute it directly (or put a
    PM> Java interface to weave.accelerate) and the normal Python is
    PM> simply a weave accelerated version of Py-Python.  For a lot of

Cool.

    PM> this to work, all one needs is a good interface between C
    PM> structs and the Python struct class.

Really? Then how hard is it to do this?  And why is that sufficient?
Any pointers on more information?  If this is all in your head and
you'd rather code it than explain it, thats fine.

Overall this is getting to be pretty amazing and the possibilities are
truly fantastic.

Cheers,
prabhu