[Bf-committers] Armatures Bones and Constraints Oh My!
Thu, 20 May 2004 18:25:37 -0400
What about Parent item ?
And size and shape?
Also to add to the "size and shape" Could it somehow have different
shapes like solid boxes for better visual feedback ?
On May 20, 2004, at 11:31 AM, email@example.com wrote:
> In whatever system that eventually replaces the current Armature and
> Constraint setup in blender I see several key things to keep in mind
> Bones need to store several key pieces of data.
> 1. Rest Position (currently in head tail and roll)
> 2. Pose Position (currently stored in quat and duplicated in pose-
> 3. IK flag
> 4. Parent Bone
> 5. Child Bone List
> 6. RootBone (could be calculated by following the parents up)
> 7. Constraint list
> anything else is really unneccessary information. Pose position
> should be stored in the same format as the rest position. This is
> where we get our data for drawing the bone and deforming the mesh.
> The rest position is only used for going to the bones rest position.
> it should never be used for drawing the bone or deforming the mesh.
> When we return to rest position we just copy the rest position values
> to the pose position. This is the basis of a working efficient Bone
> system. Some of you may see the need for additional pieces of data.
> Please respond if you do. I'd like some feedback.
> Constraints need to be evaluated per every transform operation. I
> think they would be better served as a test function not performing
> actual operations on the data. returning a boolean value for pass or
> fail perhaps. the only hitch there is many transformations operate on
> more than one axis. an indication of which axis operation failed
> might be in order. Or perhaps transformations could be calculated on
> a per axis basis. Such a method might tie in with working transform
> handles also. Constraint's would be easier to use also in the code if
> they were linked to the object's struct. Like in the proposed bone
> structure above each object should have a list of constraints on that
> object. No more looking up the constraint in the global list.
> Jeremy Wall
> PS. Sorry about the title there I have something of a sense of humor
> as you'll find out.
> Bf-committers mailing list