🚀 go-pugleaf

I'm pretty sure that polynomial functors are not what's
really needed to make Malcolm-style datatype constructions
in BMF work. There's a paper from the group at Calgary
(Cockett et al.) from which I gather that strength is
the real property of importance. This can be arranged
in several ways, one of which is to use a polynomial
functor in a CCC. Some people (Bob Walters for one)
would argue that this is the wrong setting, and prefer
distributive categories.

Anyway it seems possible to get the kind of benefits
that the CDT construction gives without needing
the polynomial restriction.

But we're getting a bit far from the centre of this
newsgroup.

The Calgary work can be obtained by ftp from
cpsc.ucalgary.ca in directory pub/charity/PAPERS.

                       -david skillicorn