[Kde-games-devel] Kapman sprite system (also, connecting-tiles systems)

[Mauricio Piacentini]

Matthew Woehlke wrote:
> Mauricio Piacentini wrote:
>> Yes, each game currently uses its own animation system, but we really 
>> have no "framework" for this. The few games that feature animation 
>> usually just append the frame numbers to the id of the element, i.e. 
>> enemy_1, enemy_2, enemy_3. It is doing the job so far, but if a given 
>> game requires a refinement then something like what Matthew suggests 
>> could be a good starting point, with hints for fps and the like.
> 
> I was definitely thinking about a unified animation library. Is there 
> anything code-wise I should look at stealing, or do y'all think it would 
> be better to start from scratch, attempting to keep reusability in mind 
> from the start?

There are some initiatives like the KGlEngine that are tackling these 
issues for our games, but they are also works in progress. As you seem 
more interested in creating frameworks and reusable systems, I think 
there are several courses of action possible. Of course, these are only 
my suggestions:

a) Had you looked at the KGlEngine specs? Maybe collaborate with these guys?

b) Find a game (preferrably unmaintained, for ultimate freedom) and 
attempt to rewrite/plug your reusable systems in it. Then the second 
step would be to abstract these to libkdegames, once they prove useful 
to more than one game at least. Maybe KSame, as it could benefit from 
both animation and the tiling scheme?

c) Write the frameworks AND a new game to use it at the same time, in 
playground. This could help you experiment with both the art/maze code 
and the animation one at your own pace, and hopefully with reusability 
in mind. Then if people find it useful it would go into libkdegames. I 
think KTank was started this way, as a proving ground for what is now 
shaping up as KGLEngine, right?

What I do not think wise is to attempt to force the framework design to 
be adopted by other games before it is ready. We end up in the land of 
speculation and personal preferences, and no productive work being 
churned out. KGoldRunner for example could potentially benefit from your 
ideas, but it has undergone several rewrites and I think Ian/Luciano and 
me are not up to yet another API change, at least for now ... :)

Regards,
Mauricio Piacentini