Haskell Day One
Written on August 8, 2014
And so, we come to the final chapter of Seven Languages in Seven Weeks. The first day covered some of Haskell’s syntax, and the features it has in common with the other functional languages in the book. As I so far understand it, the idiomatic way of approaching problems in Haskell is very similar to the way things are done in Prolog. The list comprehension syntax is almost identical to Erlang. Haskell has syntactic sugar like Ruby’s ranges, and lazy evaluation like Clojure.
Filter a list
How many different ways can you find to write
In the book, we’re provided with an interesting recursive solution to the problem of taking a list of integers and returning a list with all the odd numbers removed.
This approach strongly reminds me of the way problems are typically solved in Prolog. Perhaps there’s a direct influence there. While interesting, it’s not the approach I’d intuitively reach for — though my brain seems not to be wired for clever recursive stuff yet.
Haskell does Erlang-style list comprehensions which have filters built in. Here we take each number in our list and plug it into our new list if the remainder is zero when dividing the number by two.
As I dig deeper with my Google-fu, I learn that Haskell provides a delicious function called
filter that takes a condition and a list, returning a list with the elements that satisfy the condition. It’s so conversational, you could almost mistake it for Ruby.
Haskell cultists: Your death-threats will not scare me; I am half-Polish.
Reverse a list
Write a function that takes a list and returns the same list in reverse.
Haskell provides a
reverse function that works predictably, so implementing one of my own seems a pointless exercise. Nevertheless, we can fashion a Prolog-style recursive solution.
Write a function that builds two-tuples with all possible combinations of two of the colours black, white, blue, yellow, and red.
My example here is taken almost verbatim from the book. The function takes a list
l and binds its elements to
b in all possible combinations. We then remove duplicates by filtering where
a is not equal to
A childhood multiplication table
Write a list comprehension to build a childhood multiplication table. The table would be a list of of three-tuples where the first two are integers from 1-12 and the third is the product of the first two.
Yet another problem suspiciously well-suited to Haskell. We bind a couple of ranges to
b and return a three-tuple consisting of
b, and the product of the two.
So far I’m enjoying Haskell and I think my initial anxiety was unjustified. It’s a powerful tool and makes short work of complex problems, though this first set of challenges are probably not comprehensive enough to reflect that. I’m struggling most with Haskell’s type system; writing an appropriate type definition for the
backwards function had me stumped for a while, so I discarded it and left the types to be inferred by the compiler.
One of the problems from this day of study involved implementing a depth-first graph traversal algorithm. I didn’t study computer science at university; I went to music school. For that reason, I’ll save this challenge for another day.
My concern is that Haskell will be awkward when I need to accomplish everyday things like accumulate state and do I/O. Even Simon Peyton Jones — one of the creators of Haskell — has said that the language is “safe, but useless”.