psyco in the standard distribution

[RPM1]

Jp Calderone <exarkun at intarweb.us> wrote:
>On Sun, Mar 23, 2003 at 11:17:18PM +0000, Ian McMeans wrote:
>> Maybe this has been brought up before, but should pysco be added to the
>> standard library once it's mature enough?

>> and I'm sure the speedups would help python's reputation.
>
>  Hmm, I'm not sure about this.  People who say Python is slow are generally
>speaking out of ignorance.  Adding Psyco to the standard library won't
>educate them, it'll just be one more point for them to be unaware of.
>

Challenge:

Try writing a chess program in pure python and you will become aware of
how woefully slow Python can be.  Then port whatever algorithm you came
up with in Python to Java or C# and you will become even MORE aware
of how slow Python really is.

Then just add psyco to your Python program and watch your Python program
go from being 100 times slower, (literally), to being only ten times slower
than
the Java or C# program, (not to mention C or C++).

Of course, the standard Pythonic answer to this will be, "Well you should be
writing an extension module in C."   But the only reason I was using Python 
in the first place was because I really like Python and didn't want to use
C.

Psyco or something like it should have a place in Python's future.  I really
like Python, (nothing else would explain why I'm writing a chess program in
it),
but it IS slow.

Just my 2 cents,
Patrick