QuoteOriginally posted by: katastrofaQuoteOriginally posted by: chocolatemoneyInterestingly, I run a little survey (40fin math practitioners) and discovered that functional programming is primarily associated with currying, functions as first class citizens and the use of lambdas - anonymous functions. I was surprised as I expected referential transparency and the use of immutable data structures to top all the above. Basically, you can claim your code is functional without the need of being pure functional.Immutable data structures kind of suck when you want to do efficient numerical computations.This is true; this is obvious, you don't want to copy the whole data container at each iteration/chainge. I never suggested to prefer immutables in every circumstance.Furthermore, my post is really not about naming conventions - I thought my little survey was quite fun in illustrating some of the current trends in functional programming, some ongoing debates in the FP area and the stance taken by some players. For example, if I reckon correctly, MSFT suggested in their F# guidelines to prefer global state via OOD to more "Haskellian" state monads, despite providing syntactic sugar for monads via the computational expressions. Furthermore, if currently the key fp features fp practitioners primarily demand are first class citizens and lambdas and purity is only secondary : what is the near term future, in terms of adoption rates, of a pure and lazy fp language as Haskell? You may disagree with this, but I think the Haskell ecosystem is probably less rich than jvm and .net, hence the haskell community must grow in order to catch-up.I am sorry if I may have gone off-topic. Cheers
Last edited by chocolatemoney
on December 14th, 2013, 11:00 pm, edited 1 time in total.