Workshop: Functional programming in OO languages

As I mentioned on Wednesday, I recently ran a session on functional programming in object-oriented languages. By that, I really mean “languages that weren’t designed with functional code in mind”. This ranges from Java, which doesn’t let you treat functions as first-class objects at all, to Ruby and Python, which have all sorts of cool functional features but puts them in the corner so you don’t have to play with them if you don’t want to. The idea of the workshop was to get people thinking in terms of functions without having to learn an entirely new toolset at the same time. Lots of people have told me they want to try this at home, so here it is in text form for your coding pleasure.

Key to the idea of functional programming is immutability—never mutating existing information, only creating more. This embodies the idea of a function: a black box that takes in data, processes it in some way and then outputs something else. At no point does this function communicate with the outside world except to deliver the result. It’s side-effect free, which brings us a whole host of benefits, such as parallelisation, memoisation, laziness and reasonability. That last one is probably the most important: it simply means that because there are no side effects, the function is easier to understand and reason about. Understandable code leads to happy coders.

So here’s the challenge. I have a poker hand—seven cards that look like this:

"4d 7h 8c 8d Td Js Kh"

Not the prettiest cards on the planet, but they’re a lot easier to parse than real ones. You’ll notice they’re sorted by rank. This will make life easier for you.

Here’s your task: tell me if I have a pair in my hand. For the hand above, the pair consists of the eight of clubs and eight of diamonds. I don’t need to know which cards are in the pair though, just whether I have one or not.

When you’re done with determining whether the hand contains a pair, work your way up through each of the poker categories. Wikipedia has an explanation of each one. For example, the hand “2s 5s 7c 7s 8h Js As” contains a flush, which trumps a pair, so the category is flush.

Right. Now to make it interesting. Immutability is key. To this end, I’ve come up with a few rules you should follow.

1. No mutable types. That means no arrays, or even mutable lists such as the lists bundled with every mainstream programming language. Java’s ArrayList, Ruby’s Array, Python’s list… these are all out of bounds.
2. Don’t write your own mutable types. Really important. The state of your objects should not change during the lifetime of the object. If you’re developing in Java, this means that all fields should be final. readonly in C#. Everyone else, just don’t do it.
3. Methods must end by returning. Functions. One thing in, one thing out. Emphasis on the out.
4. Every single statement (apart from the aforementioned return) must be assignment to new variables only. Again, mark them as final if you can.
5. No conditions or loops. Essentially, no if, for or while blocks. These encourage mutation. The one exception is the ternary condition: condition ? true_case : false_case in Java, C#, Ruby and JavaScript (and I’m sure many more). Python’s equivalent is true_case if condition else false_case, which I always thought was backwards, but that shouldn’t stop you from using it.

That should cover it. There are two more rules which are just as important, but nothing to do with functional programming. The first is to write object-oriented code. Use your language. It has all these cool features. Don’t forget about them. If you know them, try your best to apply the rules of object calisthenics. You’ll find yourself breaking them a lot, because they don’t always mesh well with functional programming, but you should be aware you’re doing so. (If you don’t know about object calisthenics, don’t worry about it.)

The second: write tests first. Practice TDD as well as you can. How else will you know you’re done?

I’m not going to leave you completely in the lurch. You’ll find implementations of the functional linked list for Java, C#, Python, Ruby and JavaScript on GitHub. I’ve also written an explanation of said list, and how to implement them in Java. Each one has map implemented on it already, and you’re encouraged to add methods to it as you need them. You’ll also find they follow all the rules set out above. That should get you started. If there isn’t one in your chosen language, it shouldn’t be hard to reimplement.

I asked people in the workshop to pair, as it leads to discussion and helps people get through the tricky bits. A couple of colleagues of mine and I were also floating around to help. Unfortunately I can’t do that here, but I can definitely respond to the comments at the bottom of this post, so if you have any questions, please ask.

When you’re done, throw up your solution on the web somewhere and post a link in the comments here. Even if you only do a little bit, I’d love to see the results. I’ll be posting up my own in a couple of weeks, after people have had a chance to write their own.