I'm trying to use uMotion's constrain feature to create a fingers controller, but found it a bit difficult with current functions. Here is what I encountered:
I selected all five fingers' 15 bones (3 for each) as driven transforms. Then I found the thumb need to be rotated in a different direction from the other fingers (thumb is y and others are z axis), thus I can't build all five fingers in one constrain.
That's fine, I can let the thumb alone, but then I found each joint of a finger needs different bending angle: closer to the finger tip should bend less to form a holding hand shape rather than a solid fist.
And, different fingers should bend differently like the little finger only need to bend a little to form a circle with the same radius as a much more bent mid-finger.
So an overall channel and overall scaling is not enough to make a decent controller in my test, I think adding a weight value for each transform can solve this : the weight can be -1 to 1 and the final results can multiply with it's weight and the overall scaling before applying. Then I can set different weight for different fingers or other transforms.
Above this, I think you really should let us use custom properties as a constrain's channel, if all the driven transforms have a constrain custom property with the same name, the property can be shown in the channel dropdown, and we can select it as a driven channel and let the upper constrain to drive them, thus we can create a controller with different controlling channels.
The last one can also be achieved by giving the ability to choose a specific channel for each transform.
That's all of my thoughts right now, what do you think? :)