Christian Seberino wrote:
4) Python *still* does not have a useful JIT compiler
Unless you count 'psycho'.
I don't.
"6 August 2005
For the last few months, Psyco has been hosted on
http://codespeak.net in a Subversion repository. However, Psyco has not
been in very active development for quite a while now. I consider that
the project is as complete as it can reasonably be. Developing it
further would be possible and interesting, but require much more efforts
that I want to invest."
A JIT is harder for dynamically typed langs.
Not really, a good JIT is *hard*, period. And Java doesn't get away
from all of that complexity because of its introspection tools as well
as downcasting everything to "Object" for containers (fortunately
Generics help with this).
In addition, the Lisp, Erlang, and Haskell folks seem to handle dynamic
JIT just fine, TYVM.
Please elaborate. I'm curious what Guido's opinions are regarding
threads. I tend to avoid them myself.
Guido does not like threads either.
In addition, he is unwilling to take the required performance hit to
move things out from under the global interpreter lock.
With that lock in place, it even becomes difficult to even experiment
with different ways of implementing threading in the language.
Hopefully, PyPy will come online *soon*. Once that happens, people can
start to adjust the semantics of the language.
-a
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
