On 10/07/2013, at 8:42 PM, Andreas Abel wrote:
Hear, hear! In OCaml, I can (and often do) write
let (x,s) = foo 1 [] in let (y,s) = bar x s in let (z,s) = baz x y s in ...
I really wish you wouldn't do that. After reading Dijkstra's paper on the fact that we have small heads many years ago -- long enough to forget the actual title, sorry -- I realised that I too was a Bear of Very Little Brain and Get Confused Very Easily. I find that that when the same name gets reused like that I get very confused indeed about which one I am looking at right now. If the variable is hidden (as by the DCG transformation in Prolog, or a state monad, I don't get confused about the variable because it isn't visible. If each instance of the variable is labelled with a sequence number, I don't get confused because each variable has a different name and I can *see* which one this is. Yes, sequence numbering variable states is a chore for the person writing the code, but it's a boon for the person reading the code. Me, I'd be perfectly happy with setup (x,s) = state (\_ -> (x,s)) (setup $ foo 1 []) >>= \x -> bar x >>= \y -> baz x y >>= \z -> ... One reason for this is that it makes refactorings like extracting bar ... >>= ... baz ... thinkable. A long sequence of updates is probably crying out for such a refactoring.