[Stackless] rough PyCon'07 draft; feedback wanted

Andrew Dalke dalke at dalkescientific.com
Tue Feb 13 20:01:09 CET 2007


On Feb 13, 2007, at 10:53 AM, Andrew Francis wrote:
> I think simple example like your factorial example
> would do the trick. The idea is to illustrate
> Stackless Python's "lightweightness." In the process
> show the differences in speed and memory usage.

Hmm..  I really do need to sit down and work out a
slide count for each of these parts.  And remember to
not speak really, really fast in order to squeeze
everything in.


>> I'm not a game person.
>
> I am not either. I am using Stackless Python for
> service oriented architectures (read business, Web
> 2.0, etc).

I've seen some of your comments about doing pipelines
and that's something which interests me.  After
PyCon is done I'll probably talk with you some about
it in private email.

>> Really? There's no way I know to do that using
>> generators. It's going through the C stack because
>> of Expat so the callbacks can't yield control.
>
> Granted. I'm assuming generators could change the push
> based SAX parser into a pull based system. Let me look
> into this. I would be surprised if someone like Uche
> Ogbuji or David Mertz hasn't done this already.

All the solutions I know were done by switching to a
push-based parser and not by retrofitting a callback-based
one.  The only way to do the latter is though threads/
threadlets.  But feel free to investigate.  :)


>> But I don't know Twisted,

>> Can you help?
>
> Yes I can!

Ahh, if only some company would fund me for this  :)

Looking through the back emails I see the work you've
put into making them fit together.  I'm worried about
what would need to be done within the next week.

> I guess my attitude is that Twisted is free and
> provides a really powerful networking framework and
> its model fits in relatively well with Stackless. I
> feel if you want to sell a solution, show the average
> programmer how to solve a complex problem in a simple
> way.

I've had my experiences with Twisted, which is why I
don't look forward do doing it on my own.  What I
like about Stackless, among others, is that I can take
advantage of Twisted and async I/O without having to
twist my whole program around.

I think for my talk I'll stay with asyncore.  It's
simple enough that people can easily understand it.

In my vision of what could occur, a Stackless support
library should be able to interoperate with the 3
or so async I/O packages easily.  Just need time and
money. :)

>> What's an example of using it?
>
> One of my practical uses for pickling is ...

I've decided that I have enough to talk about and I
don't have a good example of pickling tasklets
that I can easily demonstrate so I'm not going to
talk about it, except perhaps in passing.

There's EuroPython this summer and PyCon next year
as possible venues to talk more about other aspect
of Stackless.

> Unfortunately I am about a month away from having a
> fairly complete prototype that can demonstrate
> everything I have done with Stackless Python.
>
> To recap, I feel I can commit to Twisted examples.
> Give me a hard date for stuff you would like.

EuroPython in Vilnius, Lithuania this summer?
"Stackless and Twisted" sounds like a great talk.

				Andrew
				dalke at dalkescientific.com



_______________________________________________
Stackless mailing list
Stackless at stackless.com
http://www.stackless.com/mailman/listinfo/stackless



More information about the Stackless mailing list