Notes on Haskell, and all the wonderful things it doesn’t let you do


My notes on Stephen Blackheath's talk at Open Programming Language Miniconf 2010: Haskell, and all the wonderful things it doesn’t let you do. Posted by Thomas Sutton on January 18, 2010

The first talk at the Open Programming Language Miniconf 2010 was by Stephen Blackheath’s Haskell, and all the wonderful things it doesn’t let you do. This was cool – given that Haskell is my favourite language – but I didn’t get much out of it beyond a link to Stephen’s Arrows Tutorial.

The talk, aimed at newcomers, covered a lot of ground. It started with a characterisation of Haskell as a pure, functional, non-strict, strongly, statically typed, programming language and gave all the usual reasons that those are nice properties for a programming language to have.

After a few examples of the syntax, Stephen talked about recursion and mentioned that idiomatic Haskell tends to use a higher-order function implementing a particular recursions scheme, but didn’t really point any out (other than using map in example).

Next was a brief discussion of the type system: it’s strong and static, but it also features type inference so you don’t need to supply signatures for everything. It’s probably important to note though, that it’s still a good idea for top level definitions for documentation purposes, if for no other reason. Next was a mention of QuickCheck and automatic testing.

Finally, there was a bit about Haskell’s non-strict semantics and the usual points about concurrency (being hard in imperative languages and rather easier in Haskell). Laziness is useful as it makes it possible to represent [potentially] infinite data structures and looping data structures without destructive updates, alas it can make it hard (or, at least, harder) to reason about and predict the time and space behaviour of a program.

It was a pretty good talk, but I though Stephen missed a few opportunities to make some important points that the audience might have appreciated. In particular, I’d have mentioned the various higher-order recursion operators and opportunity they provide for automatic application-specific optimisations through fusion. This is the sort of thing that is difficult or impossible in a lot of languages but are seriously cool!

This post was published on January 18, 2010 and last modified on January 26, 2024. It is tagged with: event, lca, haskell, lca2010, oplm2010, notes.