Chris Dragan's Corner

For some time now I’ve been planning to learn Haskell. Haskell is a modern functional programming language. The fact is that many other modern languages are constantly acquiring new features from Haskell, such as lambdas, coroutines, list comprehensions, etc. If you look at JS 1.8, Python 3 or C++11, they have more and more features which come from Haskell.

Learning new programming languages is a good way for a programmer to improve his skills. Haskell in particular requires a specific approach to programming, different from the languages most programmers use every day.

One thing, which is encouraging to me, is that you can compile Haskell programs into real executables and run them on bare metal, without any VMs, interpreters and other junk.

There is an obstacle, however. Haskell is scary, complex and difficult to learn. Even though its syntax is quite terse and there is very little boilerplate code in programs written in Haskell, it still has a steep learning curve. If you don’t know Haskell, a program written in this language may seem incomprehensible. This reminds me a lot of PERL, which I have a personal aversion to.

Python is trivial, you can learn it overnight by just reading the online tutorial and you can master it within a week. But Haskell? Forget about it. After one evening of reading my head swelled and at the end of an interactive tutorial I forgot half of the things I’ve learnt!

I’ve decided to take a different approach: to learn the language by combining reading a tutorial and simultaneously set some goals and achieve them by writing more and more complex programs which will do concrete things. I generally hate raw material and hands on experience is the best way for me to learn.

The Haskell Wikibook, which a friend recommended to me, is probably a good place to start. I also have upgraded the installation of Haskell Platform (I’ve had some older version installed for some time).

So for a start, after some basic reading to refresh what I’ve previously learnt and after looking here and there, I’ve managed to write a simple “Hello, World!” program:

main = print "Hello, World!"

I saved this in a file called test.hs. From there there are two options, either compile it like this: ghc -o test test.hs or to run it directly like this: runhaskell test.hs. However the output is not satisfactory:

$ runhaskell test.hs
"Hello, World!"

In particular I don’t want the double quotes. A quick Google search brought me down to this:

main = putStrLn "Hello, World!"

This achieves exactly what I wanted. Unfortunately putStrLn is more difficult to remember than print. After some practice I will probably remember it though.

I’ve also managed to print two strings, although it wasn’t trivial and involved using a do expression:

main = do
    putStrLn "Hello, World!"
    putStrLn "Hello, programmer!"

As you would expect, this is what I get:

$ runhaskell test.hs
Hello, World!
Hello, programmer!

So here is some recap:

1. The program has one function, main. This is the function which is executed to run the program. This is similar to C, Python and many other languages.
2. It’s trivial to write the above function. It has no arguments and the return value is implicit (no return value in fact). The function declaration couldn’t be more trivial: functionName = functionBody
3. putStrLn (or print) is the function I use to print a string, i.e. to write it to standard output.
4. I don’t have to use parentheses when invoking putStrLn. In fact everything until the end of line is treated as arguments for that function. If I wanted to do something more complex, I would have to use parentheses to tell the compiler where the arguments end. Another way is to use the $ operator, which I haven’t fully figured out yet.
5. The do expression lets me do several things, otherwise I would be limited to a single expression.
6. Similarly to Python, there are no semicolons needed. Haskell uses indentation to figure out what belongs where, just like Python. I could use semicolons if I wanted to put both putStrLn functions on one line though, like in Python.

One thing I’ve found out is that WordPress doesn’t color Haskell syntax, so my Haskell programs will be uncolored, sorry!

If you have ideas what I should do next, please let me know. I’m planning to do some string and/or list manipulation, together with maybe input and output formatting.