A typical scenario of using flatMap in State Monad? - scala

I read the definition of a State Monad as follows, which includes a definition of flatMap. The literal definition of flatMap is clear to me, but what would be a typical use case of it?
trait State[S,A]{
def run (initial:S):(S,A)
def flatMap[B] (f:A=>State[S,B]):State[S,B] =
State{ s =>
val (s1, a) = run(s)
f(a).run(s1)
}
}
object State{
def apply[S,A](f:S => (S,A)):State[S,A] =
new State[S,A] {
def run(initial:S): (S,A) = f(initial)
}
}

According to cats State monad documentation
The flatMap method on State[S, A] lets you use the result of one State
in a subsequent State
It means we can have state transitions nicely lined up within a for-comprehension like so
val createRobot: State[Seed, Robot] =
for {
id <- nextLong
sentient <- nextBoolean
isCatherine <- nextBoolean
name = if (isCatherine) "Catherine" else "Carlos"
isReplicant <- nextBoolean
model = if (isReplicant) "replicant" else "borg"
} yield Robot(id, sentient, name, model)
In general, purpose of flatMap is chaining of monadic computations, so whatever monad we have we can stick it within a for-comprehension.

This quote is taken from wikibooks about state monad in haskell.
If you have programmed in any other language before, you likely wrote some functions that "kept state". For those new to the concept, a state is one or more variables that are required to perform some computation but are not among the arguments of the relevant function. Object-oriented languages like C++ make extensive use of state variables (in the form of member variables inside classes and objects). Procedural languages like C on the other hand typically use global variables declared outside the current scope to keep track of state.
In Haskell, however, such techniques are not as straightforward to apply. Doing so will require mutable variables which would mean that functions will have hidden dependencies, which is at odds with Haskell's functional purity. Fortunately, often it is possible to keep track of state in a functionally pure way. We do so by passing the state information from one function to the next, thus making the hidden dependencies explicit.
Basically its purpose is the to write purely functional programs that manipulate state, having the API compute the next state rather than actually mutate anything.
The most common examples for the state monad are:
Generate random numbers.
Building games
Parsers
Data Stuctures
Any finite state machine program.
You can also check the cats page for the state monad
Note: There is an additional more complexed state monad which is called IndexedState monad, which basically gives you the option to change the state.

Related

When to use monads from scalaz?

I'd like to create a simple wrapper for computations. The built-in scala monads (TraversableLike) seems sufficient for me. And they have already syntax sugar. From some point of view scala collection traits are accidental monads. And there intended monads provided by the scalaz library.
What uses cases benefit from complex type classed monads of scalaz? What functionality is unfeasible for built-in monads and indicate need for scalaz?
Some clarification.
This question is not a holy war inheritance vs type classes. My question is about infrastructure that provides scalaz. Not any library with type classes approach, but this mentioned library. It slightly complicates things. But also it have bunch of utility classes that have no matches in scala collection library. Because it is a collection library, not a monadic. So the question is about the additional functionality provided by scalaz. In which cases does it matter?
First for a point about terminology: it's sometimes useful shorthand to say things like "Option is a monad", but "Option has a monad instance" or "Option is monadic" is clearer. It's potentially a little confusing to say that Scalaz provides a bunch of monads—what it provides is a Monad type class and instances of that type class for a number of types, including some of its own (e.g. \/, Task, etc.) and some from the standard library (List, Option, etc.).
So I'm going to answer a question which is similar to your question: what's the value of an explicit Monad type class over the monadic syntactic sugar provided by the standard library?
One place where having an explicit Monad representation is useful is when you want to define your own generic combinators or operations. Suppose I want to write a method addM that takes two monadic M[Int] values and adds them in the monad. It's easy to write for Option:
def addM(oa: Option[Int], ob: Option[Int]): Option[Int] = for {
a <- oa
b <- ob
} yield a + b
Or for lists:
def addM(oa: List[Int], ob: List[Int]): List[Int] = for {
a <- oa
b <- ob
} yield a + b
These two implementations obviously have a lot in common, and it'd be nice to be able to write a single generic implementation that would work in both cases—and for any other monadic type as well. This is really hard if we only have the standard library's hand-wavy monadic syntax, and really easy if we have a Monad type class.

What exactly makes Option a monad in Scala?

I know what the monads are and how to use them. What I don't understand is what makes, let's say, Option a monad?
In Haskell a monad Maybe is a monad because it's instantiated from Monad class (which has at least 2 necessary functions return and bind that makes class Monad, indeed, a monad).
But in Scala we've got this:
sealed abstract class Option[+A] extends Product with Serializable { ... }
trait Product extends Any with Equals { ... }
Nothing related to a monad.
If I create my own class in Scala, will it be a monad by default? Why not?
Monad is a concept, an abstract interface if you will, that simply defines a way of composing data.
Option supports composition via flatMap, and that's pretty much everything that is needed to wear the "monad badge".
From a theoretical point of view, it should also:
support a unit operation (return, in Haskell terms) to create a monad out of a bare value, which in case of Option is the Some constructor
respect the monadic laws
but this is not strictly enforced by Scala.
Monads in scala are a much looser concept that in Haskell, and the approach is more practical.
The only thing monads are relevant for, from a language perspective, is the ability of being used in a for-comprehension.
flatMap is a basic requirement, and you can optionally provide map, withFilter and foreach.
However, there's no such thing as strict conformance to a Monad typeclass, like in Haskell.
Here's an example: let's define our own monad.
class MyMonad[A](value: A) {
def map[B](f: A => B) = new MyMonad(f(value))
def flatMap[B](f: A => MyMonad[B]) = f(value)
override def toString = value.toString
}
As you see, we're only implementing map and flatMap (well, and toString as a commodity).
Congratulations, we have a monad! Let's try it out:
scala> for {
a <- new MyMonad(2)
b <- new MyMonad(3)
} yield a + b
// res1: MyMonad[Int] = 5
Nice! We are not doing any filtering, so we don't need to implement withFilter. Also since we're yielding a value, we don't need foreach either. Basically you implement whatever you wish to support, without strict requirements. If you try to filter in a for-comprehension and you haven't implemented withFilter, you'll simply get a compile-time error.
Anything that (partially) implements, through duck-typing, the FilterMonadic trait is considered to be a monad in Scala. This is different than how monads are represented in Haskell, or the Monad typeclass in scalaz. However, in order to benefit of the for comprehension syntactic sugar in Scala, an object has to expose some of the methods defined in the FilterMonadic trait.
Also, in Scala, the equivalent of the Haskell return function is the yield keyword used for producing values out of a for comprehension. The desugaring of yield is a call to the map method of the "monad".
The way I'd put it is that there's an emerging distinction between monads as a design pattern vs. a first-class abstraction. Haskell has the latter, in the form of the Monad type class. But if you have a type that has (or can implement) the monadic operations and obeys the laws, that's a monad as well.
These days you can see monads as a design pattern in Java 8's libraries. The Optional and Stream types in Java 8 come with a static of method that corresponds to Haskell return, and a flatMap method. There is however no Monad type.
Somewhere in between you also have the "duck-typed" approach, as Ionuț G. Stan's answer calls out. C# has this as well—LINQ syntax isn't tied to a specific type, but rather it can be used with any class that implements certain methods.
Scala, per se, does not provide the notion of a monad. You can express a monad as a typeclass but Scala also doesn't provide the notion of a typeclass. But Cats does. So you can create a Monad in Scala with the necessary boiler plate, e.g. traits and implicits cleverly used, or you can use cats which provides a monad trait out of the box. As a comparison, Haskel provides monads as part of the language. Regarding your specific question, an Option can be represented as a monad because it has a flatMap method and a unit method (wrapping a value in a Some or a Future, for example).

Understanding Gen.unit(x)

Working through Functional Programming in Scala, the book shows the Gen Monad definition. Gen, as I understand, is a ScalaCheck trait.
val genMonad = new Monad[Gen] {
def unit[A](a => A): Gen[A] = Gen.unit(a)
def flatMap[A, B](ma: Gen[A])(f: A => Gen[B]) =
ma.flatMap(f)
}
I believe that OptionMonad.unit is defined as Some(a), but I don't understand Gen.unit(a).
How is Gen.unit(a) defined?
A Gen[A] is just an object which can be repeatedly called to give instances of type A. These "generators" are used to drive the ScalaCheck automated testing framework, which lets programmers specify properties of objects of a given type, and then repeatedly generates instances of that type and checks those properties. Gen forms a monad, which is to say it supports the operations of "unit" and "bind", about which approximately a zillion tutorials can be found on the internet. Scala's idioms for monads are a bit inconsistent, as monad types have a standard method of bind, called flatMap, but none for unit. This is because Scala is object oriented, and unit doesn't take an object of it's monad, but instead returns one, so it doesn't make any sense to make unit a method of the underlying class. Instead, most Scala monads leave the unit method implicit, often as a single-element constructor of the monad type.
So with that background out of the way, what's unit of Gen[A]. Well it needs to be something which takes an object of type A as an argument, and then allows repeated generation of objects of type A. Since A could literally be anything, there's really only one thing we can come up with which fits this bill. unit(a) must be a boring generator which repeatedly returns a . Simple once you think it through.

which GOF Design pattern(s) has entirely different implementation (java vs Scala)

Recently I read following SO question :
Is there any use cases for employing the Visitor Pattern in Scala?
Should I use Pattern Matching in Scala every time I would have used
the Visitor Pattern in Java?
The link to the question with title:
Visitor Pattern in Scala. The accepted answer begins with
Yes, you should probably start off with pattern matching instead of
the visitor pattern. See this
http://www.artima.com/scalazine/articles/pattern_matching.html
My question (inspired by above mentioned question) is which GOF Design pattern(s) has entirely different implementation in Scala? Where should I be careful and not follow java based programming model of Design Patterns (Gang of Four), if I am programming in Scala?
Creational patterns
Abstract Factory
Builder
Factory Method
Prototype
Singleton : Directly create an Object (scala)
Structural patterns
Adapter
Bridge
Composite
Decorator
Facade
Flyweight
Proxy
Behavioral patterns
Chain of responsibility
Command
Interpreter
Iterator
Mediator
Memento
Observer
State
Strategy
Template method
Visitor : Patten Matching (scala)
For almost all of these, there are Scala alternatives that cover some but not all of the use cases for these patterns. All of this is IMO, of course, but:
Creational Patterns
Builder
Scala can do this more elegantly with generic types than can Java, but the general idea is the same. In Scala, the pattern is most simply implemented as follows:
trait Status
trait Done extends Status
trait Need extends Status
case class Built(a: Int, b: String) {}
class Builder[A <: Status, B <: Status] private () {
private var built = Built(0,"")
def setA(a0: Int) = { built = built.copy(a = a0); this.asInstanceOf[Builder[Done,B]] }
def setB(b0: String) = { built = built.copy(b = b0); this.asInstanceOf[Builder[A,Done]] }
def result(implicit ev: Builder[A,B] <:< Builder[Done,Done]) = built
}
object Builder {
def apply() = new Builder[Need, Need]
}
(If you try this in the REPL, make sure that the class and object Builder are defined in the same block, i.e. use :paste.) The combination of checking types with <:<, generic type arguments, and the copy method of case classes make a very powerful combination.
Factory Method (and Abstract Factory Method)
Factory methods' main use is to keep your types straight; otherwise you may as well use constructors. With Scala's powerful type system, you don't need help keeping your types straight, so you may as well use the constructor or an apply method in the companion object to your class and create things that way. In the companion-object case in particular, it is no harder to keep that interface consistent than it is to keep the interface in the factory object consistent. Thus, most of the motivation for factory objects is gone.
Similarly, many cases of abstract factory methods can be replaced by having a companion object inherit from an appropriate trait.
Prototype
Of course overridden methods and the like have their place in Scala. However, the examples used for the Prototype pattern on the Design Patterns web site are rather inadvisable in Scala (or Java IMO). However, if you wish to have a superclass select actions based on its subclasses rather than letting them decide for themselves, you should use match rather than the clunky instanceof tests.
Singleton
Scala embraces these with object. They are singletons--use and enjoy!
Structural Patterns
Adapter
Scala's trait provides much more power here--rather than creating a class that implements an interface, for example, you can create a trait which implements only part of the interface, leaving the rest for you to define. For example, java.awt.event.MouseMotionListener requires you to fill in two methods:
def mouseDragged(me: java.awt.event.MouseEvent)
def mouseMoved(me: java.awt.event.MouseEvent)
Maybe you want to ignore dragging. Then you write a trait:
trait MouseMoveListener extends java.awt.event.MouseMotionListener {
def mouseDragged(me: java.awt.event.MouseEvent) {}
}
Now you can implement only mouseMoved when you inherit from this. So: similar pattern, but much more power with Scala.
Bridge
You can write bridges in Scala. It's a huge amount of boilerplate, though not quite as bad as in Java. I wouldn't recommend routinely using this as a method of abstraction; think about your interfaces carefully first. Keep in mind that with the increased power of traits that you can often use those to simplify a more elaborate interface in a place where otherwise you might be tempted to write a bridge.
In some cases, you may wish to write an interface transformer instead of the Java bridge pattern. For example, perhaps you want to treat drags and moves of the mouse using the same interface with only a boolean flag distinguishing them. Then you can
trait MouseMotioner extends java.awt.event.MouseMotionListener {
def mouseMotion(me: java.awt.event.MouseEvent, drag: Boolean): Unit
def mouseMoved(me: java.awt.event.MouseEvent) { mouseMotion(me, false) }
def mouseDragged(me: java.awt.event.MouseEvent) { mouseMotion(me, true) }
}
This lets you skip the majority of the bridge pattern boilerplate while accomplishing a high degree of implementation independence and still letting your classes obey the original interface (so you don't have to keep wrapping and unwrapping them).
Composite
The composite pattern is particularly easy to achieve with case classes, though making updates is rather arduous. It is equally valuable in Scala and Java.
Decorator
Decorators are awkward. You usually don't want to use the same methods on a different class in the case where inheritance isn't exactly what you want; what you really want is a different method on the same class which does what you want instead of the default thing. The enrich-my-library pattern is often a superior substitute.
Facade
Facade works better in Scala than in Java because you can have traits carry partial implementations around so you don't have to do all the work yourself when you combine them.
Flyweight
Although the flyweight idea is as valid in Scala as Java, you have a couple more tools at your disposal to implement it: lazy val, where a variable is not created unless it's actually needed (and thereafter is reused), and by-name parameters, where you only do the work required to create a function argument if the function actually uses that value. That said, in some cases the Java pattern stands unchanged.
Proxy
Works the same way in Scala as Java.
Behavioral Patterns
Chain of responsibility
In those cases where you can list the responsible parties in order, you can
xs.find(_.handleMessage(m))
assuming that everyone has a handleMessage method that returns true if the message was handled. If you want to mutate the message as it goes, use a fold instead.
Since it's easy to drop responsible parties into a Buffer of some sort, the elaborate framework used in Java solutions rarely has a place in Scala.
Command
This pattern is almost entirely superseded by functions. For example, instead of all of
public interface ChangeListener extends EventListener {
void stateChanged(ChangeEvent e)
}
...
void addChangeListener(ChangeListener listener) { ... }
you simply
def onChange(f: ChangeEvent => Unit)
Interpreter
Scala provides parser combinators which are dramatically more powerful than the simple interpreter suggested as a Design Pattern.
Iterator
Scala has Iterator built into its standard library. It is almost trivial to make your own class extend Iterator or Iterable; the latter is usually better since it makes reuse trivial. Definitely a good idea, but so straightforward I'd hardly call it a pattern.
Mediator
This works fine in Scala, but is generally useful for mutable data, and even mediators can fall afoul of race conditions and such if not used carefully. Instead, try when possible to have your related data all stored in one immutable collection, case class, or whatever, and when making an update that requires coordinated changes, change all things at the same time. This won't help you interface with javax.swing, but is otherwise widely applicable:
case class Entry(s: String, d: Double, notes: Option[String]) {}
def parse(s0: String, old: Entry) = {
try { old.copy(s = s0, d = s0.toDouble) }
catch { case e: Exception => old }
}
Save the mediator pattern for when you need to handle multiple different relationships (one mediator for each), or when you have mutable data.
Memento
lazy val is nearly ideal for many of the simplest applications of the memento pattern, e.g.
class OneRandom {
lazy val value = scala.util.Random.nextInt
}
val r = new OneRandom
r.value // Evaluated here
r.value // Same value returned again
You may wish to create a small class specifically for lazy evaluation:
class Lazily[A](a: => A) {
lazy val value = a
}
val r = Lazily(scala.util.Random.nextInt)
// not actually called until/unless we ask for r.value
Observer
This is a fragile pattern at best. Favor, whenever possible, either keeping immutable state (see Mediator), or using actors where one actor sends messages to all others regarding the state change, but where each actor can cope with being out of date.
State
This is equally useful in Scala, and is actually the favored way to create enumerations when applied to methodless traits:
sealed trait DayOfWeek
final trait Sunday extends DayOfWeek
...
final trait Saturday extends DayOfWeek
(often you'd want the weekdays to do something to justify this amount of boilerplate).
Strategy
This is almost entirely replaced by having methods take functions that implement a strategy, and providing functions to choose from.
def printElapsedTime(t: Long, rounding: Double => Long = math.round) {
println(rounding(t*0.001))
}
printElapsedTime(1700, math.floor) // Change strategy
Template Method
Traits offer so many more possibilities here that it's best to just consider them another pattern. You can fill in as much code as you can from as much information as you have at your level of abstraction. I wouldn't really want to call it the same thing.
Visitor
Between structural typing and implicit conversion, Scala has astoundingly more capability than Java's typical visitor pattern. There's no point using the original pattern; you'll just get distracted from the right way to do it. Many of the examples are really just wishing there was a function defined on the thing being visited, which Scala can do for you trivially (i.e. convert an arbitrary method to a function).
Ok, let's have a brief look at these patterns. I'm looking at all these patterns purely from a functional programming point of view, and leaving out many things that Scala can improve from an OO point of view. Rex Kerr answer provides an interesting counter-point to my own answers (I only read his answer after writing my own).
With that in mind, I'd like to say that it is important to study persistent data structures (functionally pure data structures) and monads. If you want to go deep, I think category theory basics are important -- category theory can formally describe all program structures, including imperative ones.
Creational Patterns
A constructor is nothing more than a function. A parameterless constructor for type T is nothing more than a function () => T, for example. In fact, Scala's syntactical sugar for functions is taken advantage on case classes:
case class T(x: Int)
That is equivalent to:
class T(val x: Int) { /* bunch of methods */ }
object T {
def apply(x: Int) = new T(x)
/* other stuff */
}
So that you can instantiate T with T(n) instead of new T(n). You could even write it like this:
object T extends Int => T {
def apply(x: Int) = new T(x)
/* other stuff */
}
Which turns T into a formal function, without changing any code.
This is the important point to keep in mind when thinking of creational patterns. So let's look at them:
Abstract Factory
This one is unlikely to change much. A class can be thought of as a group of closely related functions, so a group of closely related functions is easily implemented through a class, which is what this pattern does for constructors.
Builder
Builder patterns can be replaced by curried functions or partial function applications.
def makeCar: Size => Engine => Luxuries => Car = ???
def makeLargeCars = makeCar(Size.Large) _
def makeCar: (Size, Engine, Luxuries) => Car = ???
def makeLargeCars = makeCar(Size.Large, _: Engine, _: Luxuries)
Factory Method
Becomes obsolete if you discard subclassing.
Prototype
Doesn't change -- in fact, this is a common way of creating data in functional data structures. See case classes copy method, or all non-mutable methods on collections which return collections.
Singleton
Singletons are not particularly useful when your data is immutable, but Scala object implements this pattern is a safe manner.
Structural Patterns
This is mostly related to data structures, and the important point on functional programming is that the data structures are usually immutable. You'd be better off looking at persistent data structures, monads and related concepts than trying to translate these patterns.
Not that some patterns here are not relevant. I'm just saying that, as a general rule, you should look into the things above instead of trying to translate structural patterns into functional equivalents.
Adapter
This pattern is related to classes (nominal typing), so it remains important as long as you have that, and is irrelevant when you don't.
Bridge
Related to OO architecture, so the same as above.
Composite
Lot at Lenses and Zippers.
Decorator
A Decorator is just function composition. If you are decorating a whole class, that may not apply. But if you provide your functionality as functions, then composing a function while maintaining its type is a decorator.
Facade
Same comment as for Bridge.
Flyweight
If you think of constructors as functions, think of flyweight as function memoization. Also, Flyweight is intrinsic related to how persistent data structures are built, and benefits a lot from immutability.
Proxy
Same comment as for Adapter.
Behavioral Patterns
This is all over the place. Some of them are completely useless, while others are as relevant as always in a functional setting.
Chain of Responsibility
Like Decorator, this is function composition.
Command
This is a function. The undo part is not necessary if your data is immutable. Otherwise, just keep a pair of function and its reverse. See also Lenses.
Interpreter
This is a monad.
Iterator
It can be rendered obsolete by just passing a function to the collection. That's what Traversable does with foreach, in fact. Also, see Iteratee.
Mediator
Still relevant.
Memento
Useless with immutable objects. Also, its point is keeping encapsulation, which is not a major concern in FP.
Note that this pattern is not serialization, which is still relevant.
Observer
Relevant, but see Functional Reactive Programming.
State
This is a monad.
Strategy
A strategy is a function.
Template Method
This is an OO design pattern, so it's relevant for OO designs.
Visitor
A visitor is just a method receiving a function. In fact, that's what Traversable's foreach does.
In Scala, it can also be replaced with extractors.
I suppose, Command pattern not needed in functional languages at all. Instead of encapsulation command function inside object and then selecting appropriate object, just use appropriate function itself.
Flyweight is just cache, and has default implementation in most functional languages (memoize in clojure)
Even Template method, Strategy and State can be implemented with just passing appropriate function in method.
So, I recommend to not go deep in Design Patterns when you tries yourself in functional style but reading some books about functional concepts (high-order functions, laziness, currying, and so on)

What are practical uses of applicative style?

I am a Scala programmer, learning Haskell now. It's easy to find practical use cases and real world examples for OO concepts, such as decorators, strategy pattern etc. Books and interwebs are filled with it.
I came to the realization that this somehow is not the case for functional concepts. Case in point: applicatives.
I am struggling to find practical use cases for applicatives. Almost all of the tutorials and books I have come across so far provide the examples of [] and Maybe. I expected applicatives to be more applicable than that, seeing all the attention they get in the FP community.
I think I understand the conceptual basis for applicatives (maybe I am wrong), and I have waited long for my moment of enlightenment. But it doesn't seem to be happening. Never while programming, have I had a moment when I would shout with a joy, "Eureka! I can use applicative here!" (except again, for [] and Maybe).
Can someone please guide me how applicatives can be used in a day-to-day programming? How do I start spotting the pattern? Thanks!
Applicatives are great when you've got a plain old function of several variables, and you have the arguments but they're wrapped up in some kind of context. For instance, you have the plain old concatenate function (++) but you want to apply it to 2 strings which were acquired through I/O. Then the fact that IO is an applicative functor comes to the rescue:
Prelude Control.Applicative> (++) <$> getLine <*> getLine
hi
there
"hithere"
Even though you explicitly asked for non-Maybe examples, it seems like a great use case to me, so I'll give an example. You have a regular function of several variables, but you don't know if you have all the values you need (some of them may have failed to compute, yielding Nothing). So essentially because you have "partial values", you want to turn your function into a partial function, which is undefined if any of its inputs is undefined. Then
Prelude Control.Applicative> (+) <$> Just 3 <*> Just 5
Just 8
but
Prelude Control.Applicative> (+) <$> Just 3 <*> Nothing
Nothing
which is exactly what you want.
The basic idea is that you're "lifting" a regular function into a context where it can be applied to as many arguments as you like. The extra power of Applicative over just a basic Functor is that it can lift functions of arbitrary arity, whereas fmap can only lift a unary function.
Since many applicatives are also monads, I feel there's really two sides to this question.
Why would I want to use the applicative interface instead of the monadic one when both are available?
This is mostly a matter of style. Although monads have the syntactic sugar of do-notation, using applicative style frequently leads to more compact code.
In this example, we have a type Foo and we want to construct random values of this type. Using the monad instance for IO, we might write
data Foo = Foo Int Double
randomFoo = do
x <- randomIO
y <- randomIO
return $ Foo x y
The applicative variant is quite a bit shorter.
randomFoo = Foo <$> randomIO <*> randomIO
Of course, we could use liftM2 to get similar brevity, however the applicative style is neater than having to rely on arity-specific lifting functions.
In practice, I mostly find myself using applicatives much in the same way like I use point-free style: To avoid naming intermediate values when an operation is more clearly expressed as a composition of other operations.
Why would I want to use an applicative that is not a monad?
Since applicatives are more restricted than monads, this means that you can extract more useful static information about them.
An example of this is applicative parsers. Whereas monadic parsers support sequential composition using (>>=) :: Monad m => m a -> (a -> m b) -> m b, applicative parsers only use (<*>) :: Applicative f => f (a -> b) -> f a -> f b. The types make the difference obvious: In monadic parsers the grammar can change depending on the input, whereas in an applicative parser the grammar is fixed.
By limiting the interface in this way, we can for example determine whether a parser will accept the empty string without running it. We can also determine the first and follow sets, which can be used for optimization, or, as I've been playing with recently, constructing parsers that support better error recovery.
I think of Functor, Applicative and Monad as design patterns.
Imagine you want to write a Future[T] class. That is, a class that holds values that are to be calculated.
In a Java mindset, you might create it like
trait Future[T] {
def get: T
}
Where 'get' blocks until the value is available.
You might realize this, and rewrite it to take a callback:
trait Future[T] {
def foreach(f: T => Unit): Unit
}
But then what happens if there are two uses for the future? It means you need to keep a list of callbacks. Also, what happens if a method receives a Future[Int] and needs to return a calculation based on the Int inside? Or what do you do if you have two futures and you need to calculate something based on the values they will provide?
But if you know of FP concepts, you know that instead of working directly on T, you can manipulate the Future instance.
trait Future[T] {
def map[U](f: T => U): Future[U]
}
Now your application changes so that each time you need to work on the contained value, you just return a new Future.
Once you start in this path, you can't stop there. You realize that in order to manipulate two futures, you just need to model as an applicative, in order to create futures, you need a monad definition for future, etc.
UPDATE: As suggested by #Eric, I've written a blog post: http://www.tikalk.com/incubator/blog/functional-programming-scala-rest-us
I finally understood how applicatives can help in day-to-day programming with that presentation:
https://web.archive.org/web/20100818221025/http://applicative-errors-scala.googlecode.com/svn/artifacts/0.6/chunk-html/index.html
The autor shows how applicatives can help for combining validations and handling failures.
The presentation is in Scala, but the author also provides the full code example for Haskell, Java and C#.
Warning: my answer is rather preachy/apologetic. So sue me.
Well, how often in your day-to-day Haskell programming do you create new data types? Sounds like you want to know when to make your own Applicative instance, and in all honesty unless you are rolling your own parser, you probably won't need to do it very much. Using applicative instances, on the other hand, you should learn to do frequently.
Applicative is not a "design pattern" like decorators or strategies. It is an abstraction, which makes it much more pervasive and generally useful, but much less tangible. The reason you have a hard time finding "practical uses" is because the example uses for it are almost too simple. You use decorators to put scrollbars on windows. You use strategies to unify the interface for both aggressive and defensive moves for your chess bot. But what are applicatives for? Well, they're a lot more generalized, so it's hard to say what they are for, and that's OK. Applicatives are handy as parsing combinators; the Yesod web framework uses Applicative to help set up and extract information from forms. If you look, you'll find a million and one uses for Applicative; it's all over the place. But since it's so abstract, you just need to get the feel for it in order to recognize the many places where it can help make your life easier.
I think Applicatives ease the general usage of monadic code. How many times have you had the situation that you wanted to apply a function but the function was not monadic and the value you want to apply it to is monadic? For me: quite a lot of times!
Here is an example that I just wrote yesterday:
ghci> import Data.Time.Clock
ghci> import Data.Time.Calendar
ghci> getCurrentTime >>= return . toGregorian . utctDay
in comparison to this using Applicative:
ghci> import Control.Applicative
ghci> toGregorian . utctDay <$> getCurrentTime
This form looks "more natural" (at least to my eyes :)
Coming at Applicative from "Functor" it generalizes "fmap" to easily express acting on several arguments (liftA2) or a sequence of arguments (using <*>).
Coming at Applicative from "Monad" it does not let the computation depend on the value that is computed. Specifically you cannot pattern match and branch on a returned value, typically all you can do is pass it to another constructor or function.
Thus I see Applicative as sandwiched in between Functor and Monad. Recognizing when you are not branching on the values from a monadic computation is one way to see when to switch to Applicative.
Here is an example taken from the aeson package:
data Coord = Coord { x :: Double, y :: Double }
instance FromJSON Coord where
parseJSON (Object v) =
Coord <$>
v .: "x" <*>
v .: "y"
There are some ADTs like ZipList that can have applicative instances, but not monadic instances. This was a very helpful example for me when understanding the difference between applicatives and monads. Since so many applicatives are also monads, it's easy to not see the difference between the two without a concrete example like ZipList.
I think it might be worthwhile to browse the sources of packages on Hackage, and see first-handedly how applicative functors and the like are used in existing Haskell code.
I described an example of practical use of the applicative functor in a discussion, which I quote below.
Note the code examples are pseudo-code for my hypothetical language which would hide the type classes in a conceptual form of subtyping, so if you see a method call for apply just translate into your type class model, e.g. <*> in Scalaz or Haskell.
If we mark elements of an array or hashmap with null or none to
indicate their index or key is valid yet valueless, the Applicative
enables without any boilerplate skipping the valueless elements while
applying operations to the elements that have a value. And more
importantly it can automatically handle any Wrapped semantics that
are unknown a priori, i.e. operations on T over
Hashmap[Wrapped[T]] (any over any level of composition, e.g. Hashmap[Wrapped[Wrapped2[T]]] because applicative is composable but monad is not).
I can already picture how it will make my code easier to
understand. I can focus on the semantics, not on all the
cruft to get me there and my semantics will be open under extension of
Wrapped whereas all your example code isn’t.
Significantly, I forgot to point out before that your prior examples
do not emulate the return value of the Applicative, which will be a
List, not a Nullable, Option, or Maybe. So even my attempts to
repair your examples were not emulating Applicative.apply.
Remember the functionToApply is the input to the
Applicative.apply, so the container maintains control.
list1.apply( list2.apply( ... listN.apply( List.lift(functionToApply) ) ... ) )
Equivalently.
list1.apply( list2.apply( ... listN.map(functionToApply) ... ) )
And my proposed syntactical sugar which the compiler would translate
to the above.
funcToApply(list1, list2, ... list N)
It is useful to read that interactive discussion, because I can't copy it all here. I expect that url to not break, given who the owner of that blog is. For example, I quote from further down the discussion.
the conflation of out-of-statement control flow with assignment is probably not desired by most programmers
Applicative.apply is for generalizing the partial application of functions to parameterized types (a.k.a. generics) at any level of nesting (composition) of the type parameter. This is all about making more generalized composition possible. The generality can’t be accomplished by pulling it outside the completed evaluation (i.e. return value) of the function, analogous to the onion can’t be peeled from the inside-out.
Thus it isn’t conflation, it is a new degree-of-freedom that is not currently available to you. Per our discussion up thread, this is why you must throw exceptions or stored them in a global variable, because your language doesn’t have this degree-of-freedom. And that is not the only application of these category theory functors (expounded in my comment in moderator queue).
I provided a link to an example abstracting validation in Scala, F#, and C#, which is currently stuck in moderator queue. Compare the obnoxious C# version of the code. And the reason is because the C# is not generalized. I intuitively expect that C# case-specific boilerplate will explode geometrically as the program grows.