search

paf31 / purescript-smash

An "extensible coeffect" system built out of comonads and Day convolution.
MIT License
36 stars 7 forks source link

Is it actually coeffects or comonad-derived monadic effects? #2

Open neko-kai opened 6 years ago

neko-kai commented 6 years ago

Hi Phil,

Correct me if I'm wrong, but the term 'coeffect' seems to have a different meaning than comonad effects:

paf31 commented 6 years ago

It's possible that I'm using the term incorrectly. I haven't found a concise definition of coeffect yet, so I'm equating them with computational effects which can be modeled in the co-Kleisli category of some comonad. It's not clear to me if that's complete or even correct, but it models enough things that I'm interested in that I thought it would be understandable to reuse the term. I'm happy to replace it if it is in fact incorrect.

I don't think "comonad-derived monadic effects" is the right thing though, since the monadic bit is somewhat separate. It's one way of navigating a space of application states modelled by a comonad, but it's not the only way.