Hacker News Reader

Comment by hiker | Hacker News Reader

> That's just not true. Functional programming does not eliminate state.

And yet it says so in the first sentence in the Wikipedia page for functional programming https://en.wikipedia.org/wiki/Functional_programming

>a style of building the structure and elements of computer programs—that treats computation as the evaluation of mathematical functions and avoids changing-state and mutable data.

But I'll take it that you don't have much functional programming experience.

Of course one can still go with a big global array and keep updating it in-place. A good programmer can write Fortran (or C in that case) in any language.

View on HN

At some point, you need to modify some state, otherwise your program/language is useless. And that's not me saying that, I am just quoting, or at least paraphrasing, Simon Peyton Jones:

https://www.youtube.com/watch?v=iSmkqocn0oQ

And of course a lot of so-called "functional" programs just outsource their state management to some sort of relational database. And the people talking about their creation will praise the state-less-ness of their creation. Unironically.

What can you do? ¯\_(ツ)_/¯

Anyway, more practically, the vast majority of workloads do not have computation as their primary function. They store, load and move data around. Computers, generally, don't compute. Much. For those workloads, a paradigm that tries to eliminate the very thing that the problem domain is about, and can only get it back by jumping through hoops, is arguably not ideal.

https://carlstrom.com/stanford/cs315a/papers/barroso00piranh...

> avoids changing-state and mutable data.

This doesn't mean that functional programming eliminates state. Avoiding changing-state and mutable data is different and the Wikipedia article is referring to how functional programming doesn't mutate existing data, so you avoid the stale reference problems that can occur in OO languages.

Instead, the state is the current execution state of the program. Function calls are side affect free (except when interacting with the external world, which is a special case I'm not covering here). Because of this, the only way data can be communicated from one function to another is by passing it in when calling the function, or by returning it. This means the state is just the data local to the currently executing function, and any calling functions (though the data in that part of the state isn't available to the current function it's still in memory until the calling function returns).

Contrast this with procedural programming languages, where state can also be maintained in global variables, or object oriented languages, where objects maintain their own state with the system state being spread around the whole system.

Again, you can't do computation without state. The only question is how honest you want to be about it. And whether you put things in global tables or not is completely orthogonal to whether you mutate data or make new data.

And please, no beaten up buzz words and selling pitches needed.

Reply on news.ycombinator.com