[mod_python] supporting modular mod_python extensions vs."folding" mod_psp

Rimon Barr barr at cs.cornell.edu
Mon Jun 9 13:05:18 EST 2003


Dear Sterling,

On 9 Jun 2003, Sterling Hughes wrote:

>On Mon, 2003-06-09 at 07:05, Geert Jansen wrote:
>> > PSP is a python handler, distributed with mod_python.  There
>> > is a clear boundary between the two pieces of code.  I can
>> > see your political reasoning for not including PSP, but what
>> > is your technical reason?  The other systems will not be
>> > technically disadvantaged in anyway, and psp
>> > certainly doesn't muddy things up from a technical perspective.
>>
>> The fact that mod_psp hasn't proved anything yet while other frameworks
>> have seems a technical reason to me.
>
>That's a reason against bundling?  I can see that's a reason against
>bundling PSP, but Rimon is talking about the chastity of mod_python
>here.

I agree with Geert here, and also with Ian's prior email. PSP has not
proven itself, while other projects have. There are also prior projects
with that name and concept, that are many years older than mod_psp. Does
mod_psp even have any users yet?

>For one thing, Rimon didn't have a problem with the "purity" of
>mod_python when he proposed that spyce be integrated, it was only after
>that fell through in private mail that he refocused the issue, and
>brought it to the mailing lists.

First, please do not get personal. It's not necessary. I'm only trying
to help the OSS community, just like you. And, second, please don't
misquote me. I suggested to you, in that private email, that what you
were doing had similar goals to Spyce. And what I then suggested, to see
whether you were interested, is that you help with the Spyce parser. I
only wanted to combine efforts with you, so that we could build
something bigger and better. I saw that you were working on a C-based
parser for psp, and I asked you whether you were interested in working
on such a parser for the Spyce language. The rationale is that although
the parser does not matter much for performance, because the compiled
files are cached, it would be nice as an option for developers, since
the current Spyce parser is implemented in Python and is relatively
slow. I also, thinking that you work with the Apache Foundation, I
offered to donate Spyce to the AF. Donating to the AF does not mean, in
my mind, merging the Spyce and mod_python projects.

You seem to be attacking me as an individual rather than discussing the
idea. I'm not interested in this. I am only interested in making Python
a more web-friendly language, and I think mod_python is an important
project in this regard. I also don't have a Spyce agenda, as you seem to
imply. It's just a project that I work on, because I like it, and
because I wanted to implement some features that I didn't find in other
systems.

>PSP is a very small, very fast alternative, it doesn't preclude you from
>using any of the listed solutions.  In fact you can use PSP in
>conjunction with any of the solutions listed, or you can even ignore its
>existence altogether and use something else, or code directly with
>req.write().  There is *zero* performance loss if you choose to use
>another option, like spyce.

Ok, but will PSP be any faster than any of the other solutions?
Possibly. Perhaps you are a better designer, and that's great. But, if
it's because of its integration with mod_python, then this means that
there is something missing in the mod_python API, and we should expose
that functionality. I'm concerned that this will not be the focus of
mod_python any longer, just as it was not the focus of PHP. Have you
looked at the performance of PHP -- it took years to get it to where it
is currently, and it's still not great.

>As for the shift of development focus.  I'm touched Rimon cares so much
>about how developers spend their time.

Again with the personal attacks and sarcasm! Please...

>If you want to re-invent the wheel, go for it!  Including Python with a
>Linux distribution doesn't stop you from using PHP.  PSP is a common
>feature request, and it lives as a standard module for the 99% of users
>who just want a solution that works, and don't care how efficient '[['
>is to type.  That doesn't mean alternatives aren't allowed and
>encouraged, they just aren't bundled.

It seems to me, based on everything that is currently out there, that it
is you who are re-inventing the wheel.

BTW, Spyce supports the more common <% syntax as well.

You know, come to think of it... You're making the same bundling
arguments that I've heard used before to extend monopoly positions: it's
good for the user. How would you know what's good for the user without
any mod_psp users? Just a thought..

But, as I said earlier. I'm trying to help the OSS community, as are
you. So good luck, and thank you for your efforts.

All the best,
Rimon.



More information about the Mod_python mailing list