Mon Aug 17 17:03:05 EDT 2015
How are we doing wrt. the goals in the head?
The 4 problems:
- properly relate composition of implementation of state machines to
arrows. their behavior is already arrow like.
* using phantom types and an interface to state
* these are static arrows: using Applicative instead
But still there is a conceptual problem in the fold.
- time domain: represent sub/super-sampling. some ad-hoc
implementation of this exists. find a way to abstract.
- integrate spatial domain operations as presented in Feldspar
- allow for the abstraction of data primitives (e.g. parameterize over
the the instruction set of a DSP).
all these are TODO.