BodyKinematics and PointKinematics acceleration calculation issues

[carmichaelong]

See: https://simtk.org/plugins/phpBB/viewtopicPhpbb.php?f=91&t=11925&p=33869&start=0&view=

[carmichaelong]

With the pendulum model, also tried using /|com_acceleration with an OutputReporter which was in agreement with the erroneous BodyKinematics and PointKinematics acceleration calculations

[jenhicks]

There is not a technical issue, but confusing to users ... should clarify in documentation.

[clnsmith]

The naming of these Analyses is especially confusing:

Kinematics is a subfield of physics, developed in classical mechanics, that describes the motion of points, bodies (objects), and systems of bodies (groups of objects) without considering the forces that cause them to move.

But from what I can tell the accelerations are not calculated as the second derivative of the input positions in the .mot file, but rather using the forces in the model (ie realizeAcceleration).

[carmichaelong]

But from what I can tell the accelerations are not calculated as the second derivative of the input positions in the .mot file, but rather using the forces in the model (ie realizeAcceleration).

You're absolutely correct that this is how it's done. If you don't have the full set of constraints, external forces, etc. that drive a motion, you'll get unexpected answers. We've chatted about how to deal with this, and it might be that you could flip a flag for a separate mode to get accelerations from derivatives only. The other competing factor is that we were hoping to slowly move away from Analyses too... @clnsmith any suggestions from a user end is super helpful here.