Providence Salumu 24 Days of Hackage: base

24 Days of Hackage: base

As 24 Days of Hackage comes to a close, I see no better library to review than base itself. base is a library that all Haskell users will be familiar with, as it defines the Prelude, along with a variety of other useful modules. base could be considered to be a bare-bones standard library for Haskell, and while the Prelude is fairly extensive, there’s actually a lot of other useful functionality that is provided.

Some of my favourite constructs lie in the Control.Concurrent namespace. Concurrency is a pretty hard problem, but provided with good tools it doesn’t have to be a painful experience. The Control.Concurrent.Chan module provides “unbounded” channels, which give you a tiny little message queue that runs in memory.

echo :: Chan String -> IO ()
echo c = forever $ readChan c >>= putStrLn

main :: IO ()
main = do
  c <- newChan
  forkIO (echo c)
  forever $ getLine >>= writeChan c

In main we create a new channel, then fork a separate thread to consume messages from this channel. Finally, in the original thread we write messages into the channel from user input. This gives us a concurrent application, with two threads communicating almost transparently.

Concurrency is not all base has to offer, and recently I’ve been increasingly intrested in Data.Monoid, after reading Brent Yorgey’s excellent Monoids: Themes and Variations paper. Monoids aren’t particularly complicated beasts, they just have an “empty” element, and an operation to combine two values into one larger value. This can be really nice when combined with tuples, as you can have some really expressive transformations. For example, we can use a variety of monoids to find the sum and product of a list of integers:

stats :: [Int] -> (Sum Int, Product Int)
stats = mconcat . map (\x -> (Sum x, Product x))

After a while, this has started to feel like a really natural way to aggregate data. Combined with semigroups, you can be extremely expressive in very little typing.

If you’ve been following these blog posts, then it’ll be no surprise that I love Control.Applicative, but Control.Monad and Data.Foldable are also great modules. They provide a few more abstractions, and a whole bunch of combinators which can make working with these common data structures an absolute breeze. Combinators like mapM, forever, when and unless frequently appear in my code, and the combinatros in Data.Traversable are extremely handy. I commonly have Maybe values that I want to apply an IO action to:

lookupEnv "HOME" >>= traverse putStrLn

Of course, traverse is a lot more powerful than this - you only need to look at recent papers, or the lens library to see what I mean.

All in all, base is fairly spartan, but still provides a lot of power. It’s not quite the Python standard library, but that was never it’s aim - we have the Haskell platform for that!


You can contact me via email at ollie@ocharles.org.uk or tweet to me @acid2. I share almost all of my work at GitHub. This post is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.

I accept Bitcoin donations: 14SsYeM3dmcUxj3cLz7JBQnhNdhg7dUiJn. Alternatively, please consider leaving a tip on


Providence Salumu