scala/scala

A relaxed chat room about all things Scala (2 & 3 both). Beginner questions welcome. https://scala-lang.org/conduct/ applies

David

@chugua99_twitter

I have no idea how to start, I am beginner in this language

Martijn Hoekstra

@martijnhoekstra

here is a start for basics in the language: https://docs.scala-lang.org/tour/basics.html

Renan Reis

@renanreismartins

Hello, I'm going through the Functional Programming in Scala and on the chapter about State we have:
case class State[S, +A](run: S => (A, S))

then we have a method
def simulateMachine(inputs: List[Input]): State[Machine, (Int, Int)]

later on:

val m: Machine = Machine(true, 5, 10)
Candy.simulateMachine(inputs).run(m)

What is making me confused is that "m" is of type machine and "run" expects a "S => (A, S)" function... so how could I pass Machine to "run"?

Gavin Bisesi

@Daenyth

In this case simulateMachine's S is Machine

State[Machine, *] so run takes S which is Machine

run provides the "initial" state, and then things go from there

1 reply

Luis Miguel Mejía Suárez

@BalmungSan

@renanreismartins run does not expects a function, run IS A function, you are calling it with the initial state.

(btw, please do not use threads, they are hard to read)

Renan Reis

@renanreismartins

Sorry, I did not know about the thread etiquete.

Rob Norris

@tpolecat

There is no way you would have known.

Renan Reis

@renanreismartins

@BalmungSan when you say run is a function, where is it defined?

Fabio Labella

@SystemFw

State has a field run of type S => (A, S) (a function from S to a tuple of A and S)

when you say .run, you access that field

when you add (), you call Function.apply (under the hood) therefore passing the S to the function

the S type parameter in this case is equal to Machine, which is why you can pass a Machine to the function

Renan Reis

@renanreismartins

@SystemFw thanks for the explanation. Kinda makes sense to me, but where is the definition of run where it takes the S and makes it on a (A, S) ?

Luis Miguel Mejía Suárez

@BalmungSan

@renanreismartins yeah no worries about the thread, sorry if I sounded rude.
Fabio already explained it pretty well but maybe this helps, the code is expanded like this:

Candy.simulateMachine(inputs).run.apply(m) // we are calling the apply method on the run value which is an instance of the function class

Martijn Hoekstra

@martijnhoekstra

if you have a val x: Int => Int it's a value that returns a function

Luis Miguel Mejía Suárez

@BalmungSan

Here: final case class State[S, +A](run: S => (A, S)) when you create an State you have to pass a Function.

Basically the State data type is just a convenient wrapper over a S => (A, S) Function.

Fabio Labella

@SystemFw

@renanreismartins actually, that's the key conceptual point about State

when you create an State you have to pass a Function.

I think this view is confusing

but where is the definition of run where it takes the S and

so you have various State computations, right?

which you can combine together with map and flatMap and pure etc

Renan Reis

@renanreismartins

yes

Fabio Labella

@SystemFw

so, those computations are building the function

so what's the starting point?

the initial mini functions that get combined together

those are the primitive State computation, i.e get and put

I think this view is confusing

so actually I take this back

you can build a State computation by passing the function

and that's how get and put are built

Renan Reis

@renanreismartins

@SystemFw yes, the example from the book is quite complex, at least for me, the way he creates that function, so Im trying to decompose

Fabio Labella

@SystemFw

or, you can build State computations based on other State computations

ok, so what State represents is a sequence of transformations over an S

Renan Reis

@renanreismartins

Do you guys have a simple example?

The way the book creates the computation:

def simulateMachine(inputs: List[Input]): State[Machine, (Int, Int)] = for {
    _ <- sequence(inputs map (modify[Machine] _ compose update))
    s <- get
  } yield (s.coins, s.candies)

Fabio Labella

@SystemFw

which will result in an A

that's what State[S, A] means, "a sequence of transformation over an S, which will result in an A"

does this bit make sense?

I will get to simple examples soon, once the main concepts are clarified :)

Renan Reis

@renanreismartins

that's what State[S, A] means, "a sequence of transformation over an S, which will result in an A"

yes

Fabio Labella

@SystemFw

ok, so once you have built this sequence of transformations, you pass the initial state, and you get back the result A (and the final new state, which is also of type S)

does this also make sense?

this is what you do when you write mySequenceOfTransformationsWhichIsStateSA.run(initialState)

Renan Reis

@renanreismartins

Also does, looks like very much with Random number Generator, previously on the chapter. Till here i can follow you

this is what you do when you write mySequenceOfTransformationsWhichIsStateSA.run(initialState)

perfect :)

Luis Miguel Mejía Suárez

@BalmungSan

@renanreismartins btw, even if I know Fabio will be able to make it clear (because he always does). May I ask, what was your experience with Scala before that book?
It is quite complex and some people recommend learning the language + fp with it which IMHO it is really an overkill.

Fabio Labella

@SystemFw

right, so we can start by building the simplest State program, the program whose final result is the state itself

the name of this program is get, and its type is State[S, S], i.e. the sequence of transformations on S which will return S itself (which is what "getting the state" is)

fair enough?

Where communities thrive

People

Repo info

Activity