[Bf-funboard] Reasoning behind child - parent coordinates

[Campbell Barton]

> I can't see any benefit, reason or logic behind this. This makes things
> confusing and conflicts with the way other 3d apps work. And I can't find a
> simple solution (besides writing my own python script which I did), to get
> normal parenting behavior.
>
>
> best wishes,
> Gatis

This allows for animating an object, Then parenting it.

Without the parent-inverse-matrix, parenting would have to adjust the
animation curves - to keep the object in the same position. (in
practice there are drivers, fcurve-modifiers, even python
expressions... etc).

So this is really needed for parenting _not_ to break existing
animation, constraints... etc.

@Knapp, this is unrelated to the depsgraph.

@Alberto Torres, quite sure all values relating to parent/child
relationship are accessible via Python. we only miss utility functions
to easily access them.