[Bf-taskforce25] RNA-IPO

[Roland Hess]

Joshua -- I really like this structure you have worked out.

A couple of things that immediately come to mind:

- It would have to be transparent for the casual and existing user. The 
implementation would have to allow them to keep working like they have 
been from a work flow perspective, and add power when they want to dig 
deeper. For people already doing advanced animation work, no one's going 
to complain if the paradigm shifts enough to really smooth things out 
and grant some power.

- As for the f-curve modifiers -- that's a great opportunity. They don't 
just need to be procedural. Picture being able to add another curve with 
only, say five points, as a modifier to a set of mo-capped curves. Pull 
around those five bezier points and have it pull around/affect the "hi 
res" mo-cap curves. That's some serious power. In reality, a curve is 
just a user-defined function of x->y mappings, so it is a still a 
"procedure." The point is that something like this shouldn't be 
constrained -- make it open and artists will take it places you've not 
even considered.

- I really like the idea of each ID-block having it's own link to the 
bAnimData block. This sort of thing could be used for speedups where 
pose caching is missing. The armature portion of a linked group does not 
cache it's pose (not a proxied one -- I'm talking about targeting the 
Action to the proper subobject in NLA), and neither does the onion 
skinning with armatures. With a system like this, you make sure that 
along with that animation link for each ID block you also have the 
proper structure (probably in RNA) to cache things like posing.

Roland

Also -- with the Action DNA wrapped, what you like me to look at next? 
Timeline RNA? Start porting Action Editor tools to operators?