var & val

When an identifier holds data, you must decide whether it can be reassigned.

You create identifiers to refer to elements in your program. The most basic decision for a data identifier is whether it can change its contents during program execution, or if it can only be assigned once. This is controlled by two keywords:

• var, short for variable, which means you can reassign its contents.
• val, short for value, which means you can only initialize it; you cannot reassign it.

You define a var like this:

var identifier = initialization

The var keyword is followed by the identifier, an equals sign and then the initialization value. The identifier begins with a letter or an underscore, followed by letters, numbers and underscores. Upper and lower case are distinguished (so thisvalue and thisValue are different).

Here are some var definitions:

// VarAndVal/Vars.kt

fun main() {
  var whole = 11              // [1]
  var fractional = 1.4        // [2]
  var words = "Twas Brillig"  // [3]
  println(whole)
  println(fractional)
  println(words)
}
/* Output:
11
1.4
Twas Brillig
*/

In this book we mark lines with commented numbers in square brackets so we can refer to them in the text like this:

• [1] Create a var named whole and store 11 in it.
• [2] Store the “fractional number” 1.4 in the var fractional.
• [3] Store some text (a String) in the var words.

Note that println() can take any single value as an argument.

As the name variable implies, a var can vary. That is, you can change the data stored in a var. We say that a var is mutable:

// VarAndVal/AVarIsMutable.kt

fun main() {
  var sum = 1
  sum = sum + 2
  sum += 3
  println(sum)
}
/* Output:
6
*/

The assignment sum = sum + 2 takes the current value of sum, adds two, and assigns the result back into sum.

The assignment sum += 3 means the same as sum = sum + 3. The += operator takes the previous value stored in sum and increases it by 3, then assigns that new result back to sum.

Changing the value stored in a var is a useful way to express changes. However, when the complexity of a program increases, your code is clearer, safer and easier to understand if the values represented by your identifiers cannot change—that is, they cannot be reassigned. We specify an unchanging identifier using the val keyword instead of var. A val can only be assigned once, when it is created:

val identifier = initialization

The val keyword comes from value, indicating something that cannot change—it is immutable. Choose val instead of var whenever possible. The Vars.kt example at the beginning of this atom can be rewritten using vals:

// VarAndVal/Vals.kt

fun main() {
  val whole = 11
  // whole = 15 // Error   // [1]
  val fractional = 1.4
  val words = "Twas Brillig"
  println(whole)
  println(fractional)
  println(words)
}
/* Output:
11
1.4
Twas Brillig
*/

• [1] Once you initialize a val, you can’t reassign it. If we try to reassign whole to a different number, Kotlin complains, saying “Val cannot be reassigned.”

Choosing descriptive names for your identifiers makes your code easier to understand and often reduces the need for comments. In Vals.kt, you have no idea what whole represents. If your program is storing the number 11 to represent the time of day when you get coffee, it’s more obvious to others if you name it coffeetime and easier to read if it’s coffeeTime (following Kotlin style, we make the first letter lowercase).

• -

vars are useful when data must change as the program is running. This sounds like a common requirement, but turns out to be avoidable in practice. In general, your programs are easier to extend and maintain if you use vals. However, on rare occasions it’s too complex to solve a problem using only vals. For that reason, Kotlin gives you the flexibility of vars. However, as you spend more time with vals you’ll discover that you almost never need vars and that your programs are safer and more reliable without them.

Exercises and solutions can be found at www.AtomicKotlin.com.