Язык Kotlin, зачем он нужен, и его концептуальные отличия от Scala. Презентация доступна на странице проекта: http://scaladev.ru

1. The Kotlin
Programming Language
Svetlana Isakova

2. What? Who? Why?
 A modern language
 Statically typed
 Object-oriented
 General-purpose
 Compiler
 JVM byte code
 JavaScript (not GWT)
2

3. What? Who? Why?
 A modern language
 Statically typed
 Object-oriented
 General-purpose
 Developed by JetBrains
 Open source
 Intended for industrial use
3

4. Dynamic vs static typing
 Dynamically typed languages
 Easier to learn
 Easier to write small programs
 Statically typed languages
 Programs run much faster
 More powerful IDE
 Easier to support huge projects
4

5. Scala as a candidate
 Goal
 statically typed
 JVM-targeted
 concise language
 Scala
 too complicated
 too implicit
 too difficult to create a good IDE
5

6. Era of new industrial languages
 Ceylon Red Hat / Jboss (April 2011)
 Scala Typesafe (May 2011)
 Kotlin JetBrains (July 2011)
 xTend Eclipse / itemis (November 2011)
6

7. Kotlin “Hello, world”
fun main(args : Array<String>) {
Greeter().greet()
}
class Greeter(name : String = "world") {
val greeting = "Hello, $name!"
fun greet() {
println(greeting)
}
} 7

8. Problems and Solutions
 Extending Existing APIs
 Handling Absent Values
 Case Analysis
 Multiple Implementation Inheritance
 Collection transformation
 Variance of Generic Types
 Builders
8

9. Extending Existing APIs
9

10. Regular expression example
 Java
Pattern p = Pattern.compile("d*");
Matcher m = p.matcher("123");
if (m.find())
String result = m.group();
 Scala
val r = "d*".r
val result = r.findFirstIn("123")
10

11. Scala: implicit conversions
val r = "d*".r
trait StringOps {
...
def r: Regex = new Regex(toString)
}
implicit def augmentString(x: String)
= new StringOps(x)
11

12. Scala: implicit conversions
val r = augmentString("d*").r
trait StringOps {
...
def r: Regex = new Regex(toString)
}
implicit def augmentString(x: String)
= new StringOps(x)
12

13. Scala: implicit conversions
val r = "d*".r
 Conversion to 'StringOps' class:
 implicit
 requires a wrapper for each invocation
13

14. Functionality extension
 Java
Collections.sort(list, comparator)
 Scala / Kotlin
list.sort(comparator)
 'sort' is not declared in java.util.List
 'sort' can be autocompleted
14

15. Kotlin: extension functions
 list.sort(comparator)
 fun <T> List<T>.sort(c: Comparator<T>) {
...
}
15

16. Iterator for String
for (c in “abc”) {
println(c)
}
16

17. Iterator for String
for (c in “abc”) {
println(c)
}
class StringIterator(val s: String) :
Iterator<Char> {
private var i = 0
override fun next() = s[i++]
override fun hasNext() = i < s.size
}
17

18. Iterator for String
for (c in “abc”) {
println(c)
}
fun String.iterator() =
StringIterator(this)
18

19. Extension functions
 fun <T> List<T>.sort(c:
Comparator<T>) {...}
 fun String.iterator() =
StringIterator(this)
 non-virtual
 compile to a static function
19

20. Extending APIs
 Scala
 implicit
 overhead on wrappers
 Kotlin
 functionality cannot be inherited
20

21. Handling Absent Values
21

22. Java NPE problem
 if (order.getCustomer(). //NPE
getInfo().getAge() >
limits.getMinLegalAge()) {
...
}
22

23. Java: @Nullable, @NotNull
 @NotNull Customer getCustomer() {
...
return null; //warning, assertion
}
 @Nullable Customer getCustomer() {...}
getCustomer().getInfo() //warning, NPE
23

24. Scala: Option type
 def getCustomer() : Option[Customer]
{…}
 order.getCustomer match {
case Some(customer) =>
Some(customer.getInfo)
case None => None
}
24

25. Scala: Option type
 def getCustomer() : Option[Customer]
{…}
 order.getCustomer match {
case Some(customer) =>
Some(customer.getInfo)
case None => None
}
 order.getCustomer.map(_.getInfo) 25

26. Scala: Option type
 def getCustomer() : Option[Customer] {
...
return null //no error
}
26

27. Scala: Option type
 def getCustomer() : Option[Customer]
 order.getCustomer.map(_.getInfo).
map(_.getAge)
 extra objects, classes are created:
 Some(customer), Some(info), Some(age)
 classes for function literals
 objects for function literals
 inconvenient for debugging 27

28. Kotlin: nullable types
 fun getCustomer() : Customer? {
...
return null
}
 fun getCustomer() : Customer {
...
return null //error
} 28

29. Kotlin: nullable types
 fun getCustomer() : Customer?
 getCustomer().getInfo() error
 getCustomer()?.getInfo() ok
29

30. Kotlin: nullable types
 fun getCustomer() : Customer?
 getCustomer().getInfo() error
 getCustomer()?.getInfo() // Info?

30

31. Kotlin: nullable types
 fun getCustomer() : Customer?
 getCustomer()?.getInfo() // Info?

 val customer = getCustomer()
if (customer != null)
customer.getInfo() // Info
31

32. Call chains
 Java
order.getCustomer().getInfo().getAge()
 Scala
order.getCustomer.map(_.getInfo).
map(_.getAge)
 Kotlin
order.getCustomer()?.getInfo()?.getAge()
32

33. “If not null” check
 Scala
opt foreach { value =>
operate(value)
}
 Kotlin
if (value != null) {
operate(value)
} 33

34. Java Interop
 Scala
val files = new File("test").listFiles()
println(files.length) //NPE
 Kotlin
val files = File("test").listFiles()
println(files.length) //doesn't compile
println(files?.size ?: "no files")
println(files!!.size) //NPE
34

35. Kotlin & Java interop
 @NotNull, @Nullable annotations
 “alternative headers”
package java.util
public trait List<E> :
java.util.Collection<E> {
...
fun listIterator() :
java.util.ListIterator<E>
} 35

36. Handling absent values
 Scala
 overhead
 Kotlin
 inconvenient Java interop without alternative
headers or annotations
36

37. Case Analysis
37

38. Scala: pattern matching
sealed trait Expr
case class Num(value: Int) extends Expr
case class Sum(left: Expr, right: Expr)
extends Expr
def eval(e: Expr) : Int =
e match {
●
case Num(v) => v
case Sum(l, r) => eval(l) + eval(r)
}
38

39. Kotlin: smart casts
trait Expr
class Number(val value: Int) : Expr
class Sum(val left: Expr, val right: Expr):
Expr
fun eval(e: Expr) : Int {
if (e is Number)
return e.value //smart cast
if (e is Sum)
return eval(e.left) + eval(e.right)
throw IllegalArgumentException(
“Unknown expression $e”)
} 39

40. Kotlin: smart casts
trait Expr
class Num(val value: Int) : Expr
class Sum(val left: Expr, val right: Expr):
Expr
fun eval(e: Expr) : Int =
when (e) {
is Num ­> e.value
is Sum ­> eval(e.left) + eval(e.right)
else ­> throw IllegalArgumentException(
“Unknown expression $e”)
}
40

41. Multiple
Implementation
Inheritance
41

42. Scala traits
trait A {
val a = print("A")
}
trait B {
val b = print("B")
}
class C extends B with A {}
new C()
42

43. Scala traits
trait A {
val a = print("A")
}
trait B {
val b = print("B")
}
class C extends B with A {}
new C() // prints BA
43

44. Scala traits
trait A {
val a = print("A")
}
trait B extends A {
val b = print("B")
}
class C extends B with A {}
new C()
44

45. Scala traits
trait A {
val a = print("A")
}
trait B extends A {
val b = print("B")
}
class C extends B with A {}
new C() // prints AB
45

46. Scala traits
trait A {
def foo()
}
trait B extends A {
override def foo() = "B"
}
class C extends B {}
new C().foo() //"B"
46

47. Scala traits
trait A {
def foo()
}
trait B extends A {
override def foo() = "B"
}
trait D extends A {
override def foo() = "D"
}
class C extends B with D {}
new C().foo() //"D" 47

48. Kotlin traits
 are more like Java interfaces
 methods can have bodies
 no state
 no linearization
48

49. Kotlin traits
trait A {
fun foo() = println("A")
}
trait B {
fun foo() = println("B")
}
class C : B, A {} //doesn't compile
49

50. Kotlin traits
trait A {
fun foo() = println("A")
}
trait B {
fun foo() = println("B")
}
class C : B, A {
override fun foo() {
super<B>.foo()
super<A>.foo()
}
} 50

51. Collection
transformation
51

52. Scala collections
 Set(1, 2, 3).map(_ * 0) → Set(0)
 "abc".map(_.toUpper) → "ABC"
52

53. Scala collections
 Set(1, 2, 3).map(_ * 0) → Set(0)
 "abc".map(_.toUpper) → "ABC"
 trait TraversableLike[+A, +Repr] {
def map[B, That](f: A => B)(implicit bf:
CanBuildFrom[Repr, B, That]): That
}
53

54. Kotlin approach
 'map', 'filter' always return immutable list
54

55. Kotlin approach
 'map', 'filter' always return immutable list
 hashSet(1, 2, 3).map
{(i: Int) ­> i * 0}
→ List(0, 0, 0)
55

56. Kotlin approach
 "abc".map {(c: Char) ­>
c.toUpper} → "ABC"
 fun String.map(
f : (Char) ­> Char) : String
56

57. Lazy computations
 Iterators are lazy
 set.iterator().map { … }.filter { … }
57

58. Lazy computations
 Iterators are lazy
 set.iterator().map { … }.filter { … }
 Iterables are not lazy
 val lazySet = set.lazyMap { … }
lazySet().filter { … }
lazySet().filter { … }
58

59. Variance
of Generic Types
59

60. Scala. Declaration site variance
 class List[+T] {
//only produces T
def head() : T
def tail() : List[T]
}
 class Comparable[­T] {
//only consumes T
def compare(t1 : T, t2: T)
} 60

61. Kotlin. Declaration site variance
 class List<out T> {
//only produces T
fun head() : T
fun tail() : List[T]
}
 class Comparable<in T> {
//only consumes T
fun compare(t1 : T, t2: T)
} 61

62. Use-site variance
fun <T> copy(from: Set<T>,
to: Set<T>) {
...
}
copy(numbers, objects)
62

63. Kotlin. Use-site variance
fun <T> copy(from: List<out T>,
to: List<T>) {
...
}
copy(numbers, objects)
63

64. Kotlin. Use-site variance
fun <T> copy(from: List<out T>,
to: List<T>) {
from.add(t) //error
...
}
copy(numbers, objects)
64

65. Scala. Use-site variance
def copy[T](from: Set[_ <: T],
to: Set[T]) {
...
}
copy(numbers, objects)
65

66. Scala. Existential types
def copy[T](from:
Set[X] forSome { type X <: T },
to: Set[T]) {
...
}
copy(numbers, objects)
66

67. Builders
67

68. Builders in Groovy
html {
head {
title "XML encoding with Groovy"
}
body {
h1 "XML encoding with Groovy"
p "this format can be used as an
alternative markup to XML"
/* an element with attributes and text content */
ahref:'http://groovy.codehaus.org'["Groovy"]
}
}
68

69. Builders in Kotlin
html {
head {
title { + "XML encoding with Kotlin" }
}
body {
h1 { + "XML encoding with Kotlin" }
p { + "this format this format is now type­safe" }
/* an element with attributes and text content */
a(href = "http://kotlin.jetbrains.org")["Kotlin"]
}
}
69

70. Builders: Implementation
 Function definition
fun html(init : HTML.() ­> Unit) : HTML {
val html = HTML()
html.init()
return html
}
 Usage
html {
this.head { ... }
} 70

71. Builders: Implementation
 Function definition
fun html(init : HTML.() ­> Unit) : HTML {
val html = HTML()
html.init()
return html
}
 Usage
html {
head { ... }
} 71

72. Builders: Implementation
class Tag(val name : String) : Element {
val children = ArrayList<Element>()
val attributes = HashMap<String, String>()
}
class TagWithText(name : String) : Tag(name) {
fun String.plus() {
children.add(TextElement(this))
}
}
class HTML() : Tag("html") {
fun head(init : Head.() ­> Unit) { }
fun body(init : Body.() ­> Unit) { }
72
}

73. Builders in Kotlin
html {
head {
title { + "XML encoding with Kotlin" }
}
body {
h1 { + "XML encoding with Kotlin" }
p { + "this format this format is now type­safe" }
/* an element with attributes and text content */
a(href = "http://kotlin.jetbrains.org")["Kotlin"]
}
}
73

74. Resources
 Home page
http://kotlin.jetbrains.org
 Web Demo
http://kotlin­demo.jetbrains.com
 Blog
http://blog.jetbrains.com/kotlin
 Forum
Kotlin at http://devnet.jetbrains.com 74

75. Interesting links
 Marius Eriksen “Effective Scala”
http://twitter.github.com/effectivescala/
 Yang Zhang “True Scala complexity”
http://yz.mit.edu/wp/true­scala­complexity/
 Joshua Suereth D. “Scala in Depth”
http://www.amazon.com/Scala­Depth­Joshua­Suereth­
D/dp/1935182706
75

76. Thank you!