[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [pygame] GSoC - Graph module?

Of course its worth submission..and I agree to your point that it has got wide range of applications..

On Fri, Mar 20, 2009 at 4:51 AM, Ira Fich <igfich@xxxxxxxxxxxxxxxxxx> wrote:
Hi all,

I'm a fourth-year Compsci major at the University of Toronto, and I've been thinking about proposing a project for pygame in this year's GSoC.  I like the idea of an AI module, but I think that even more than that, pygame would benefit from a powerful, generic, dedicated graph module, containing a number of functions especially useful in a game context.
This would be similar to some of the popular graphing modules like NetworkX and igraph, but optimized for games.  In addition to the standard graph analysis and manipulation, it would include plenty of AI functions, including A* and minimax search and Bayesian networks.  On top of that, it would support symbolic graphs (useful for anything with a regular structure) and be able to recognize and generate planar graphs (vital for maps).  And naturally, it would allow for complete customization of nodes (associating nodes with sprites, for instance) and the easy integration of user-defined functions.

I envision something equally well suited to modelling a world map, a finite state machine, or a neural network.  It's likely that this would interact closely with the Improved Sprite and Scene System project, as well as the math module, and possibly the networking one as well; I expect I'd be in constant communication with the people from those projects, to ensure that everything fit together as smoothly as possible.

So, what do you think?  Is this a good idea?  Can it be made better?  Is it worth submitting a proposal?

Ira Fich