I would most definitely echo the concern about the massive size of such a project as well as future maintainability. I don't know anyone who would be able to do such a thing on a volunteer basis, and I think it would be highly unlikely that anyone would step up to pay someone to do it. 

I imagine we all agree that we would love to have such a complete reference, but I have a hard time seeing how it's even remotely feasible.

As the effort supporting Tcl/Tk core development is fairly small, and most of the people involved are far more interested in the Tcl side of things rather than Tk, the odds of huge changes on the Tk front are small, beyond keeping it working, fixing bugs, and the odd little isolated burst of new stuff once in a while.

Based on previous history with that community, I think you can safely assume any licensing issues won't be an impediment to whatever needs to be done on this end, though I'm sure there are details to sort out.