Advanced Programming Topics Course Material
First of all, let me stress that I won’t pretend functional programming is the ultimate panacea, the One True Way, the way real programmers do it. Like everything in life, it has both advantages and disadvantages. Good programming consists of being able to apply the right ideas in the right places, so as to both maximize the benefits and minimize the downsides of each approach.
During this text, functional programming might come across as very awkward or impractical, or you might think that there are clearly situations where it comes across as comically forced. And you might very well be right: in those cases one might need to rely on imperative concepts. In this text, however, we will go “full functional”, but in practice, you should probably never go full functional.
If we ask Wikipedia what functional programming is, it answers us that it is a programming paradigm that treats computation as the evaluation of mathematical functions. While technically true, it may be a bit too theoretical for our purposes. Before delving into actual pragmatic explanations of what functional programming actually entails, we’ll try to give you a general feeling about what the above definition actually means.
Consider the recipe below:
Take 250g of dark chocolate
Heat it up au bain marie.
Meanwhile, in a separate bowl, place 75g of flour.
Add 125g of sugar.
Add 5 eggs (white and yolk), one by one so as to prevent clumps from formimg.
Wait until chocolate is melted.
Off the fire, add 175g butter to chocolate.
Mix until butter is dissolved.
Add chocolate/butter mixture to flour/sugar/egg mixture. Mix.
Pour into ramekin.
Bake in over at 200 degrees Celsius for around 9 minutes.
This recipe is written in an imperative style: it is a series of instructions which have to be performed sequentially. This recipe is the procedure for baking some dessert.
If we rewrite this recipe in functional style, we’d get
Bake the mixture of 250g of dark chocolate melted au bain marie
and 175g butter with the mixture of 75g flour,
125g sugar and 5 eggs for 9 minutes at 200 degrees Celsius.
Functional programming does not specify a clear sequential order, but focuses on how to combine “values” to achieve a certain goal. There are some constraints on the order: you can’t bake first and mix the ingredients after. But contrary to the imperative version, it does not demand that the chocolate be melted first.
At its very core, functional programming consists of getting rid of state. If something is stateful, it means it is able to changed. Consider the code below:
# Python
x = 5
x = 6
A variable named x is initialized to 5. On the next line, it is being
overwritten with 6. x has changed from 5 to 6.
Because of a variable’s ability to change, it is stateful.
Functional programming means that changing things is prohibited: variables cannot be reassigned values, lists or any other data structure cannot be modified in any way, etc.
A synonym for stateful is mutable, i.e., changeable. Stateless is then also known as immutable. We’ll use these terms interchangeably.
You might wonder what good it does to voluntarily limit ourselves to this functional way of programming. Let us give answers to questions you might have.
Is functional programming still as “powerful” as imperative (stateful) programming? Can we still write every program in it?
Yes. Every imperative program can be translated into an equivalent functional program and vice versa.
Is functional programming as efficient?
As with many design decision, it depends on the situation. Sometimes it’s more efficient, sometimes less.
It looks harder. Is it worth the extra effort?
Whether it’s truly harder is a matter of debate. It certainly is less familiar, but that’s only temporary, by definition. Functional programming has multiple advantages, chiefly that functional code is a lot easier to reason about. You’ll encounter plenty of examples in later parts of this text.