Henrik Brix Andersen wrote:
Well, I think we should discuss the road map (or whatever you prefer to call it) for the release to come after 2.2 (2.4? 3.0?). We need to talk about stuff like PDB rewrite, GEGL integration, next generation XCF, plug-in policies etc.
Is anyone actually working on a requirements list for a new PDB? It would be nice to know what exactly we expect from it, and whether the PDB should be refactored or re-written with gegl nodes in mind. Also, is there any way currently to have standalone out-of-process or threaded GeglNodes?
As to the next generation XCF, pippin has a pretty good file format which he is currently using with libgggl and bauxite, currently there is no competing format around.
For gegl migration, there has been quite a bit of talk already. Last I heard there were two ideas, which can generally be called data first or structure first - either we keep the same image data objects, and start using GeglGraphs for compositing first, then start changing data structures (I guess that going this way means turning GeglNodes into more or less an interface which delegate operations to a GIMP data based implementation, or to a gegl data based implementation, then eventually killing off the GIMP based ones)., or we change the data structures first (which implies that the data structures and the methods for accessing them are finished) and once we have the data in gegl, start migrating the internal compositing engine and plug-ins to GeglNodes.
There was no concrete decision out of the last conversation we had about this. Dan and Calvin's proposal to Mark Shuttleworth on bounties will help lay down tracks here, but it's currently looking unlikely that gegl will be ready to integrate into the GIMP this Sumemr at its current rate of change. Dan has said he'll have less time to spend on it from now on, which means that gegl right now needs some people to work on the data structures end.
If this doesn't happen soon, Calvin's proposition of migrating the rendering framework first, and following up with the data structures when they're done, will look a lot more attractive.
I recently posted a link to a requirements document that Lourens Veen did for plug-in distribution, but I haven't seen any comments on it. Again, this is one of those things that needs someone actually working on it, rather than just saying to ourselves "that needs to be done". I know yosh was thinking about this - perhaps he has comments?
We also need to talk about the GIMP Foundation and how it fits in with GIMP development.
I think that this will be resolved before GUADEC. I hope it will be resolved by the end of this week...
-- Dave Neary [EMAIL PROTECTED]
_______________________________________________ Gimp-developer mailing list [EMAIL PROTECTED] http://lists.xcf.berkeley.edu/mailman/listinfo/gimp-developer