Tony Morris

More than a couple of people have asked me to talk about what it's like to program in Haskell. As a first attempt, here's a little window into the experience via something incredibly banal, but IMO hugely motivating.

I have some code that uses `Maybe` values heavily (for, ahem, good reasons). At the end of a particular function, I need to tie all of them off into just the few that are relevant given the arguments. Effectively, the equivalent of how `||` is used in JS, e.g. `a || b || c`, etc

Now that I have tests settled for this particular module, I thought I'd go see what function in the stdlib might exist that is equivalent to my `orElse`. So, I punched its type signature into hoogle: Maybe a -> Maybe a -> Maybe a

Hoogle, an index of the ~standard libraries (and which you can run locally using an index of the libs used in *your* project ), popped back a set of results, the second of which is exactly equivalent to my `orElse`: <|>, defined in Control.Applicative: http://hackage.haskell.org/package/base-4.12.0.0/docs/Control-Applicative.html#v:-60--124--62- …

This kind of thing cannot be done without things that are super-important in the haskell world, e.g.: * sufficient static analysis discipline, obviously aided greatly by surfacing the type system * adequate community enthusiasm for tools to augment/replace flawed human faculties

Fun addendum that illustrates now noob I still am: I've actually used <|> a bunch because of parsec, but until now erroneously thought it was parsec-specific (thanks to the similarity to BNF's pipe). it's super-obvious now that <|> is big-F Fundamental