[pygame] Re: GSoC A* for AI

[Orcun Avsar <orc.avs@xxxxxxxxx>]

yes, but these should be general methods. AI is a very large case but
these must be AI modules like path finding that many users should
need. otherwise it serves only for special purposes. what do you
advice for general AI demands?

On 2 Nisan, 04:16, René Dudfield <ren...@xxxxxxxxx> wrote:
> hi,
>
> I don't know if that is enough for a 3 month project.  Perhaps a few
> more things could be added.
>
> cheers,
>
> On Thu, Apr 2, 2009 at 12:02 PM, Orcun Avsar <orc....@xxxxxxxxx> wrote:
> > its only little time left for submiting deadline but i want to share
> > quickly my ideas with you about AI module and i can work for a
> > proposal. I red some discussions about this on the mailing list. Here
> > is my opinion. A* algorithm is not a very hard thing to write. Porting
> > it from C may make it faster but i think main question here how to
> > make it more usable. It should be used with minimal effort. Finding
> > shortest way is not as good as finding shortest way in shortest time
> > but algorithm or code can be healed, changed in future but usage must
> > be same and good. I plan to write a astar pathfinding module that is
> > practical to use.
> > something like first registering sprites or rects to module(rect
> > objects have coordinates and size ). Rect objects are general with all
> > games and this is better than creating a matrix or scanning a surface
> > for colorkeys.
> > pathfinder.register( user.rect )
> > pathfinder.register( enemy.rect )
> > pathfinder.register( house.rect )
>
> > and any time using some function like this  pathfinder.findway
> > (start=user.rect,dest=house.rect)
> > and it can return a result like node points of shortest path like
> > (  (12,32), (12,300), (400,300)  ) that each node consist of x and y