[Edu-sig] Shuttleworth Summit

Ian Bicking ianb at colorstudy.com
Fri Apr 21 18:40:29 CEST 2006

I don't actually know a ton about Pygame.  I was going to copy someone 
listed on the website, but the pygame website is currently broken.  So 
off the top of my head I'm copying Richard Jones, and maybe he'll be 
able to better describe pygame's abilities than I (the discussion kind 
of starts from here: 

Guido van Rossum wrote:
> On 4/21/06, Ian Bicking <ianb at colorstudy.com> wrote:
>>This all sounds like pygame.  Is there a reason people aren't happy with
>>pygame as infrastructure?  It's quite portable, there are already
>>widgets built on top of it (like Squeak, these widgets have no relation
>>to the system UI -- for better and worse).  It already has a development
> Good idea. I never managed to successfully install pygame, so I kind
> of question the "quite portable" assertion. But maybe it's better now.
> (Does it run on OSX at all?)

It does, but the installers aren't always up to date with the OS.  It's 
portable if someone ports it.  Starting a new project because no one is 
willing to put the effort into keeping an existing port up to date is 
also a bad strategy.

The underlying layer -- SDL -- is the best graphics portability layer I 
know of.  Not that I know a ton about these things.  The only flaw I see 
in it is that it doesn't interact with native widgets or windows, which 
doesn't seem like what people want to do.  So either you reimplement SDL 
(which is what Squeak does -- though it predates SDL), or reimplement 

>>There's still a lot of hard work to creating a real Squeak-like
>>environment there, but going any lower level than that -- especially for
>>reasons like portability -- just doesn't make any sense.  If there was
>>some other environment with other compelling reasons to use it (like
>>maybe better vector support, or better libraries to create systems with
>>text flow and whatnot), but unless someone wants something that *already
>>exists* elsewhere, I can't see a reason to consider radical
>>infrastructure work.
>>There are some basic limitations to pygame, but those limitations don't
>>seem to be the reason people are speculating about other (mostly
>>non-existant) systems.
> How is pygame's 3D support? While Squeak's UI has a very flat 2D look,
> some Squeak apps are effectively 3D virtual worlds.

I don't actually have much pygame experience.  SDL has OpenGL support, 
and I see some evidence that it is exposed to PyGame as well.

> All the pygame apps I've seen demoed (e.g. Armin Rigo's PyPy slides
> and demos) were entirely 2D. But maybe there's 3D rendering with
> camera positions, shading, lighting etc. too? (If not, adding it might
> be a SoC project?)

If it's just a matter of exposing or improving the underlying OpenGL 
bindings to Python, then that seems like a feasible project.  If that 
work is already done, then a good project might be working on filling 
out a widget system (several are reviewed here: 
http://pitchersduel.iuplog.com/default.asp?item=109658) -- if you can't 
do decent text editing then that's a problem for a programming environment.

Finding a mentor already familiar with Pygame would be essential.

Ian Bicking  /  ianb at colorstudy.com  /  http://blog.ianbicking.org

More information about the Edu-sig mailing list