Choosing a programming language as a competitive tool

Tim Peters tim.one at home.com
Sun May 6 21:31:01 CEST 2001


[Steven D. Majewski]
> ...
>  However, lots of other threads on the desire for a better python
> compiler, and comparisons with Lisp performance has me again
> considering the idea of writing another Python implementation
> in Lisp (i.e. parallel to what Jython is to Java).
>
>  It would be interesting to see if the benefits of an optimizing
> Lisp compiler would carry though to better LPython performance.

Note that OCaml also has a highly optimizing compiler (competitive with C
speed), and John Skaller already reimplemented Python (+ a number of
interesting extensions) in and for OCaml:

    http://vyper.sourceforge.net/

Unfortunately, it appears the Vyper project was last active a year ago.  John
*started* Vyper because his Interscript literate programming system:

    http://interscript.sourceforge.net/

was written in Python, and ran too slowly for his tastes.  The point of Vyper
was to speed up Interscript, but that appeared to get sidetracked in favor of
adding new features (I suspect that may not be unique in the history of
software projects <wink>).

Anyway,

    How fast is Vyper?

    Vyper was designed to be compiled, however the compiler is
    currently not available.  The interpreter Pystone benchmark
    is currently about 10 times slower than CPython 1.5.2.
    Considerable improvement is expected in the future.

If you want to see what potential there is in this kind of approach, pick up
where John left off.

eats-factors-of-ten-for-breakfast<wink>-ly y'rs  - tim





More information about the Python-list mailing list