Some of you may wonder why i have not posted anything significant to computer graphics recently. I wanted to appologize for this. With work, my INS dealings, the holidays coming and my self-publishing venture my free time has been rare recently to work on any significant projects.

One thing i have been tinkering with on and off recently is Inverse Kinematics. I am still far from having something functional yet, but have been making some progress. But so far still have a few unanswered questions:

1) CCD or Jacobians? I undrestand that CCDs can be less stable but seem much more intuitice to implement and cost efficient than Jacobians. Any one has any experience with both who can say that it is truely worth doing Jacobians even for dealing with simple chains?

2) Working in quaternion space. Most text i have read perfer treating IK rotation as a set of Euler (x,y,z) rotations. However, to me it would seem that working in quaternion space would be more cost-efficient and intuitive. I have read that you can do this by simply determining the best axis of rotation for every frame and then simply using this. But how can you efficiently deal with other aspects of IK such as joint constraints in quaternion space without having to convert everything back to Euler space?