I’ve seen a couple posts in the top in the last 6 hours feed, and it seems like people are really up in arms about this functional programming stuff. Not really sure what it even is.
It looks like it’s people writing bad programming or something? Like a lot of extra stuff that is not necessary?
EDIT: sorry everyone, I’m not a programmer and I don’t know to much other than a little java and python. I guess I should have posted this in Explain Like I’m Five.
As others mentioned, it’s a programming paradigm. It and discussions around it have zero implications outside of programming.
People can write great applications using it or using any other paradigm. Same is true for terrible applications.
Some people love it, some people hate it, most are somewhere in between and think it has their merits and tradeoffs, and that it can be used where it makes sense, but shouldn’t where it doesn’t.
Heated discussions are very common in tech circles over things that have zero practical implications outside our own little world, and this is one of them. 😄
I get what you’re saying, but if your application doesn’t have 14 layers of inheritance, it’s barely an application.
#Java4Life
Nah nah nah, it’s just a different paradigm. It’s like… it’s a different meta, but for programming.
If I was to ELI5, I’d say it’s the difference between making a series of objects that interact with each other (OOP) and and creating a very large console with lots of buttons (functional)
Some problems are much easier object oriented. If you have a video game, building a projectile class and adding different types of projectiles makes things suuuper easy to build. Arrows move slow and deal x damage, beams move super fast and deal y.
Functional can be easier to troubleshoot, also easier to crank out novel things, easier to make secure, and even make simultaneous operations trivial. It has very different problems though. You need to put more effort into eliminating dependencies.
The two do not play well together at all they’re like fire and water. Sometimes you need water to soak something and make it easier to work, sometimes you need heat to melt something and make it easier to work.
Telecommunications like phones or military applications tend to the functional languages, data, video games, and a lot of the popular languages tend to OOP.
The transition between the two is jarring, and infuriating, but a knowledge of both can really improve your design skills.
You earn your name, Sir Duke.
In my opinion kotlin tries to slowly close the gap between functional and OOP programming. I love the implicit null checks, extension functions and the various list functions😊
Let me try getting this to ELI5 levels:
Most programming languages use a way of programming called “procedural programming” (or a variation of that, for example object-oriented programming). A series of instructions gets executed step by step to modify the program’s overall state. Imagine it like a cooking recipe. You have control structures like
if,whileandforthat influence the order in which instructions are executed. There are functions (or methods in object-oriented programming) that let you reuse some instructions over and over but that’s not quite functional programming.I’ll use a syntax similar to Java, C++ or C# as an example:
void squareAllElements(int[] array) { for(int i = 0; i < array.size(); i++) { array[i] = array[i] * array[i]; } }Functional programming is a bit more like what you know from maths class. There are still functions that take a list of inputs and produce an output but there is no state to be modified. Every operation must take everything it needs (apart from maybe constants) as parameters and produces new values/objects as an output. You say you know Java, so imagine that none of your methods returns void and every variable can only be assigned once. There are no loops but for example operations that let you separate the first element of a list from the rest or create a new list by applying an operation to all elements of an input. To make that easier, functions can take other functions as parameters.
So let me show the same example in a functional style. This is no specific language, just an example to help you imagine how it works.
int square(int x) => x * x; int[] squareElements(int[] array) => applyToAllElements(array, square);Functional programming has its own advantages and disadvantages. It can avoid a lot of bugs because functions can’t have unwanted side effects. They just take inputs, return outputs and leave the rest of the world (including their inputs) unmodified. On the other hand, that makes other use cases where you actually want to have some state that changes over time more difficult.
Hope that helps.
Edit: some cleanup and more precise explanations.
Edit: replaced “paradigm” with “way of programming” to make @[email protected] happy.
ELI5
paradigm
You must know some terrifying five-year-olds.
At some point, using actual five year old words just makes communication harder rather than easier. OP mentioned that they’re at least somewhat familiar with some programming languages so I assumed they’d figure it out from context. But fine, I’ll change it.
And another example that shows off pattern matching:
Procedural:
int minimumElement(int[] array) { if(array.size() == 0) { return 0; } int minimum = array[1]; for(int i = 1; i < array.size(); i++) { if(array[i] < minimum) { minimum = array[i]; } } return minimum; }Functional (still, no specific language, just an example of what it might look like):
int min(int a, int b) => a < b ? a : b; int minimumElement(int[] array) => array match { [firstElement, ...rest] => min(firstElement, getMinimumElement(rest)); [singleElement] =>singleElement; [] => 0; }To a layman like me it looks like procedural looks more like logical coding and the functional version looks more like a math problem
If you made it through that lengthy post and are still interested, let me add that many concepts from functional programming languages have made it into other languages over the years. Personally, I’m most familiar with C# which has functions that take other functions as parameters, lambda functions (functions without a name that you can just define directly where you use them as a parameter), simple list operations that take functions (filtering, sorting, transforming…) and to some extent pattern matching. From what I’ve heard, newer versions of Java have much of that as well and some of those features have even made it into C++ (from C++11 onwards) though in that case the syntax is a bit hard to read.
Since my other post got a bit wordy and maybe not quite ELI5, let me simplify as much as I can, with the risk of leaving out details:
Say you have a red car but want a car that’s identical in every except that it’s black.
In a procedural programming language, you may just paint the car black. Or you may build an exact copy of the old car but in black. Or you may take half the parts from the red car and put them into a new black outer shell. Or anything in between. Any change to your old car, your driveway or anything else in the world is called a “side effect”. The programming language doesn’t prevent you from writing something that has side effects that you didn’t actually want. It’s the programmers responsibility to make sure.
In a functional programming language, you’re not allowed to change the red car at all. You have to create a new black car. In fact, you can’t have any side effects. That way, you can’t accidentally damage your old car but makes some wishes harder to express.
It’s a programming paradigm where every action is made up of functions, functions which have no side effects beyond their explicit outputs.
I understand every one if those words, yet together I don’t understand them at all.
think of it as: It’s a paradigm where you only use mathematical functions, rather than programmatic ones. Not completely accurate, but gets you closer. Uh, if you know advanced mathematics, anyway.
Thanks! I love going down new (to me) paths!
In addendum to the side effects, functions can certainly have side effects, as the global state is still accessible, but only those that don’t are considered “pure”.
I tried to write a number of things, but think wikipedia or something might have a better answer. I mostly wilrite in golang which is not functional, but I do occasionally maintain some things in elixir which is and I think it’s pretty neat. I’m sure if you read about it and have specific questions, folks here will try to answer them.
Functional programming means that you structure your code in a manner of functions that take arguments and return values based on the functions logic.
This is in sharp contrast to procedural programming which shares a global state to access and mutate.
deleted by creator
Well it makes testing very different and also the act of accessing and changing data in the global scope can cause much havoc as a programmer always need to have the entire program in mind and not just the function they are working on.
That’s the sharp contrast e.g. it makes organising and working on code bases easier.
That’s because simple examples don’t really highlight the difference.
Functional programming is completely different from procedural programming at many levels. Not just the code, syntax and program flow, but also the way it’s compiled and executed. But I’d go further than that and say it’s like a different way of thinking about what programming actually is.
When people say it’s a different paradigm, that’s exactly what it is.
For example I do a lot of stuff in OpenSCAD, which is a functional programming language to generate 3D models. What only sunk into me as I used it more was how the entire program is a single mathematical formula. We all know the y = 2 + 3x from school and with functional programming, that’s exactly what you are doing. But y is the 3D model and x is the input parameters. And once that penny drops it all falls into place.
Another place I’ve used functional programming a lot if when programming Digital Signal Processors (DSPs). There is a continuous input stream and there needs to be continuous output stream. This makes it ideal for functional programming, where the input is one stream and the output is another. With functional programming it’s known exactly how much processing is required for a given algorithm and helps a lot in knowing what chip is required and what the latency will be. There’s also all sorts of crazy optimizations and analysis you can do, which can’t easily be done with procedural programming.
Now this isn’t to say you can’t do a lot of the same stuff in both paradigms, one isn’t better or more capable than the other in any absolute way. And for most people procedural thinking is much more intuitive, which is why it’s so dominant perhaps. But functional programming has it’s niches where it excels.
hmmm. maybe look at it like a style of writing where you try to keep nouns separate from verbs.
think of verbs here as the something agnostic to context and can be joined to make bigger tasks. e.g. draw + color + cut paper for an art
there’s other nuances on the style but verbs would be the functions and nouns would be the data.
