> The way I see it ATM is that we'll probably end up with a type
> function or fundep that says the "shrinking fixup" patch type
> is determined by the model.
That was roughly what I tried first but it didn't quite work out,
and I think the reason is that you also want to judiciously choose the
shrinking fixup type to be in some sense compatible with the real
patches, so you can efficiently check whether it cancels fragments
of them.
|