Hi,

This is a suggestion/question about the guts of Processing.
I'm right now coding a library and I'm running into a little problem.
I have in my library some classes which I would like to contain a function draw(PGraphics g) in order to allow them to draw "themselves" in any given PGraphics, PGraphics2, PGraphics3, PGraphicsGL object we pass to them.  So I have must I implement three different draw() methods with three different signatures each time?

I think it would be better having a PGraphicsInterface and all the PGraphics* implement from.  That would allow people to implement their own PGraphics renderers and other people access those objects transparently without knowing the renderer, like my case.

I just saw, that PGraphics* extend from PGraphics, so it should work with one only signature for draw().

But I still don't understand why not having an interface and implementing it for each different renderer. It seems like the academic example for making use of interface .  But then, I'm quite a beginner and maybe I'm missing something.

Thanks for the aclarations.

You're right about the interface creating unbeneficious overhead.
And this cleared out a lot about the "spirit" of processing.  So it made me change a few things in my library, in order to make it simpler, but keeping the functionality we want to keep.

In a week or two I'll publish an Alpha (changes in the API might happen) version of the library.

thanks