Addition to ToDo page

[Philip Prindeville]

I was wondering what would be involved in
coming up with a skeletonized API and
structure for encoder devices in the drivers.

Since a lot of encoder chips like the VT1625
and VT1631 are agnostic about what their
source is, making them modular would give
a certain amount flexibility in "mix and match"
(i.e. using an Intel GPU with a Via encoder,
etc).

Seem reasonable?

-Philip