Interesting. An abstraction that is asymmetric in its distributivity sounds like a good idea for this. Isn't the problem with Maybe just that it is a left-nearsemiring instead? The laws seem to hold (back of envelope check). Or, if we flipped the behaviour of mplus for Maybe (so it has right bias instead) then the right-nearsemiring laws hold (again, I just scribbled this but it looks right).
no subject
Date: 2014-02-06 09:56 am (UTC)From: