[Stackless] Towards a Stackless Book Re: Stackless Python included in the Softpedia

Andrew Francis andrewfr_ice at yahoo.com
Tue Dec 29 19:08:14 CET 2009


Hi Jason:

>Message: 1
>Date: Mon, 28 Dec 2009 07:45:33 -0800
>From: "Connor, Jason" <jason.connor at adidas-group.com>
>To: <stackless at stackless.com>
>Subject: [Stackless] Stackless Python included in the Softpedia
>Message-ID:
    <967C7036ED6A5944BE5AB39EB524DC0C01D821EB at usporamex01.am.adsint.biz>
>Content-Type: text/plain;    charset="us-ascii"

>adidas will have a small group in Atlanta for Pycon as well. We've been
>lurking for about a year now and have a couple of small internal apps
>running with stackless/Twisted. Nothing major or groundbreaking or
>interesting though.

Sweet! I gave a talk two years ago on "Stackless Python Twisted Integration." I think the two products inter-operate fairly well and over the past two years I have a couple of ideas about making things more efficient. In the process, I thought about stuff like design patterns (reactor, active objects, and half-sync/half-async) and theoretical 
models like I/O automata and simple operation management principles. I
touched on some the ideas in a recent talk to the Montreal Python users
group. 

Having a MIS background,  I am really interested about hearing about how Stackless and Twisted are used in a business environment. Outside of my own stuff, at the National Film Board of Canada, I wrote robots that simulated Flash Players using Stackless, Twisted, and PyAMF. 

> If anything it would be nice if there was at least a lightening talk
> or something.

If I show up I would love to give one. I wanted to do a 45 minute talk at Pycon 2010 talk but I decided against submitting. Also I wanted to try out ideas for potential sections in Bruce Eckel's "Python 3K Design Patterns." I received some good feedback from some of the bigger Pythonistas about improving the talk. 

>An O'Reilly book along the lines of what was done for Twisted would
>definitely be nice. Just something simple poke the attention of people.

Stackless would definitely benefit from a book. The Abe Fettig Twisted
book isn't the greatest. However the Fettig book has simple concrete 
examples one could execute and modify. That is a big bonus. 

I don't see why the Stackless community could not pitch in and write
a book: there is a lot of material ...

If I wrote a book, I guess it would be in the style of Tanenbaum's "Modern Operating Systems" or Douglas Comer's "Operating Systems Design": discuss ideas in the context of building a full-blown application. My app would be a toy orchestration/workflow application. I don't think there is not much difference between an orchestration engine and an online game with NPCs. I think the message exchange patterns are more complicated in the former and there is a greater onus on performance in the later. 

I think the only real point of contention would be how networking
is done. In my Montreal talk, I discussed Twisted and Stacklesssocket.

In a book, I would like to touch on PyPy. 

To speed up things I would probably need help in understanding what is happening under the hood. Working on select() is improving my knowledge on this front. Soon I will dive into PyPy.

Of course, I would love to hear what Christian and the Stackless community has to say about this.

Happy Holidays,
Andrew









      




More information about the Stackless mailing list