[REBOL] Re: facts we will have to face ...
From: maksa:sezampro:yu at: 24-Oct-2001 9:27
Petr Krenzelok wrote:
>> btw: Could anyone enlighten me and describe,
>> how Python solves such issue of interpreter speed?
From my experience - it doesn't. ;)
A little story: I have a co-worker who is a Perl bigot when it comes to
scripting, and we sometimes have language-pissing matches. But they're
constructive, in a "talk-is-cheap,
so-prove-whatever-you're-saying-by-writing-code"
sense. We both learn in the process, so we don't feel so bad after all about
doing
what is realy a stupid thing - arguing about languages. Now, one of the
matches
was a task to write a fast and extensible cross-referencing program for web
sites
that the web part of the company is producing. The task was to find how many
ADO VBScript constants are actually being used from the 12 Kb file included
on
almost every ASP page. He wrote his in Perl, I wrote mine in Python. His was
the
cleanest and most readable Perl code I've ever seen (but I should say that I
haven't
seen much Perl code, it's not easy on the eyes ;), and it was pretty damn
fast.
My Python code looked a lot better, but it was a *real* catastrophy when it
came to
performance. So I rewrote mine in Ruby while fully retaining the Python
programs
structure. Speed gain was incredible. Results:
Ruby: 0.82 s
Perl 3.50 s
Python: 17.2 s
Results are obtained after running each program several times to equalize
their
file caching opportunities, etc.
This, of course, doesn't prove much, except that Python is way slower when
processing files and doing string matching, and that Ruby had some clever
speed tricks when it came to that.
I'm sorry to say that my Rebol knowledge was (and probably still is, but I'm
working
on that) inadequate at the time to write the thing in Rebol. I think I'll go
and do it now.
Sorry if the whole post was an off-topic drag, but I just couldn't keep my
mouht shut
when I saw "Python" and "interpreter speed" in the same sentence. :)
Regards,
Maksa