Could a single web framework popularize Python?

[Alex Martelli]

Ian Bicking wrote:
   ...
> that use Java).  Personally I think Perl is on the way down, and has
> been for a while -- Perl 6 is a disaster that won't even arrive, and in
> general the thrill of Perl is gone.  PHP is almost more vulnerable,

I think you're wrong, but I admit I'm guessing.  To me, particularly
networking at OSCON, it seemed as if Parrot was generating lots of buzz.

> because it's support is so shallow -- at least Perl programmers have a
> certain loyalty, PHP programmers usually just fall into the language.

Interesting, because to me it seemed very different - i.e. as if the
_sociology_ of PHP was an example well worth studying and perhaps trying
to get inspiration from.  I've often seen PHP'ers moving to Python
bemoan the fact that www.python.org isn't community-run and intense-
community-involvement the way www.php.org is -- that seems as inevitable
a complaint as Perlites-moving-to-Python's hankering for CPAN.

> PHP's success, as I see it, has been from its shallow initial learning
> curve (nevermind the ceilings you'll hit).  Beginning web programmers
> get stuff done quickly, it keeps them working, and if PHP is used in

Sounds credible, but I think it may ignore the "buying into the
community" sociological effect.  But, I _am_ just musing/guessing.

> Since Python isn't proprietary, and doesn't encourage proprietary
> offerings, something like the J2EE standard is unnecessary.  An

It may not be indispensable, but it WOULD help a lot in avoiding
fragmentation of effort -- just like e.g. the DBAPI helps a bit
(and could do more) to help avoid fragmentation there.

> implementation *is* the standard, just like CPython is the standard
> definition of the language.

I don't think that's necessarily a good thing.  One of the hopes of
the pypy project is eventually supplying a better "executable
pseudocode" definition;-).

>> With python, you face a non-trivial research project even deciding
>> which environment you're going to use.

True, and important (for both Web and GUI programming, BTW).

> Twisted's HTTP probably isn't the most interesting part -- it's
> certainly not what attracts most people to it.  More interesting is
> Twisted's support for other protocols, like Jabber or SSH.  Being

That's definitely how I got into Twisted -- via the lower level
only (I still haven't used any of the upper layers yet).  Not sure
it generalizes all that much, though.

> Python it's a bit more nimble, which is something the core of Apache
> can't offer for the aspiring low-level networking programmer.  Also
> Twisted is a bit easier to pull apart.  Then there's the asynchronous
> programming -- you love it or you hate it -- but that's another
> discussion.  (Honestly I've never tried it enough to know, but I'm so
> skeptical of the ability of asynchronous programming to appeal to a
> general audience of programmers that I haven't bothered to give it a
> real try)

This sounds to me like claiming that "a general audience of programmers"
will never be able to do GUI's because just about all of them require
event-driven, i.e. asynchronous, programming.  And yet, what loomed like
a huge issue back in the time of (e.g.) early Macs / Windows / X now
doesn't seem anywhere like so terrible -- teaching GUI programming,
even in C, is quite easy (http://www.aleax.it/TutWin32/index.htm --
in Italian -- for an example; I have successfully used that material
to get beginning C programmers to code simple Windows GUI's on the bare
API's without breaking a sweat...).


Alex