Python Parallel Paradigm

Paul Boddie paul at
Sun Nov 12 15:45:01 CET 2006

Sandy wrote:
>     A higher-level system of concurrency, not based on monitors and
> locks and great programmer discipline, will ultimately require making
> "Python 3000" a reality.

It would surprise me if Python 3000 introduced anything substantially
more than what Python 2.x provides in the area of concurrency.

>     In the meantime, is there anywhere, or any thing, that discusses the
> various concurrency options related to Python?  There's Stackless Python
> (which I can't make head or tail of; I have been unable to find any
> lucid overview, or genuine explanation of the purpose of the design.)

What about this introduction...?

>  I know that there's a package for an Erlang system for Python, somewhere
> ("Parnassus" probably).  There's probably a Py-CSP somewhere too.  Lots
> of trees, but where's the Wood?

Here are some fairly similar projects in Python:
 - a way of making general concurrency easy to work with, and fun
 - a Python implementation of Erlang concurrency primitives
 - process forking and channel-based communication (using pickles)

>     Where are concurrency/distributed models compared and discussed?

In the following article and comments there are some opinions expressed
on Python concurrency (along with the usual dose of Ruby vapourware

Meanwhile, the topic has been discussed on python-dev:

That discussion led to a description of something which isn't so far
from parallel/pprocess, at least:

Other searching yielded this interesting paper about PyPy and "hardware
transactional memory":


More information about the Python-list mailing list