Variables

Demonstrate the language's methods of variable declaration, initialization, assignment, datatypes, scope, referencing, and other variable related facilities.

Local variables are generally called local variables in ALGOL 68. Variables must be declared before use. In traditional ALGOL 68, variables must be declared before any labels: in a compound-clause. The declaration of a variable, without assigning a value takes the form: <typename> <variablename>; <lang algol68>int j;</lang> Some common types are: char, string, short int, int, long int, real, long real, bits and bytes .

Multiple variables may be defined in a single statement as follows: <lang algol68>LONG REAL double1, double2, double3;</lang> It is possible to initialize variables with expressions having known values when they are defined. The syntax follows the form <typename> <variablename> := <initializing expression>; <lang algol68>SHORT INT b1 := 2500; LONG INT elwood = 3*bsize, jake = bsize -2;</lang> The strings in ALGOL 68 are flex arrays of char. To declare initial space for a string of exactly to 20 characters, the following declaration is used. <lang algol68>FLEX[20]CHAR mystring;</lang> All arrays are structure that include both the lower lwb and upper upb of the array. Hence strings in ALGOL 68 may safely contain null characters and can be reassigned with longer or shorter strings.

To declare an initialized string that won't be changed the following declaration may be used: <lang algol68>[]CHAR mytext = "The ALGOL 68 Language";</lang> There are more rules regarding arrays, variables containing pointers, dynamic allocation, and initialization that are too extensive to cover here.

AppleScript

Variables are untyped in AppleScript, but they must be instantiated before use. Example:<lang AppleScript>set x to 1</lang> Scope may be explicitly defined before instantiation using either the global or local declarations.<lang AppleScript>global x set x to 1 local y set y to 2</lang>If undeclared, AppleScript will automatically set the scope based on the following rule: variables declared at the top level of any script will be (implicit) globals, variables declared anywhere else will be (implicit) locals. Scope cannot be changed after being explicitly or implicitly defined. Where a variable has both local and global instances, it is possible to use the my modifier to access the global (top-level) instantiation. Example:<lang AppleScript>on localx() set x to 0 -- implicit local return x end localx

on globalx() set x to 0 -- implicit local return my x end globalx

on run set x to 1 -- top-level implicit global return {localx(), globalx()} end run --> RETURNS: {0, 1}</lang>

Applescript also supports top-level entities known as properties that are global to that script. Example:<lang AppleScript>property x : 1</lang>Properties behave exactly as global variables except that they are persistent. Their most recent values are retained between script executions (or until the script is recompiled).

AutoHotkey

<lang autohotkey>x = hello ; assign verbatim as a string z := 3 + 4 ; assign an expression if !y ; uninitialized variables are assumed to be 0 or "" (blank string) Msgbox %x% ; variable dereferencing is done by surrounding '%' signs fx() { local x ; variable default scope in a function is local anyways global y ; static z=4 ; initialized once, then value is remembered between function calls }</lang>

AWK

In awk, variables are dynamically typecast, and do not need declaration prior to use. <lang awk>a = 1 # Here we declare a numeric variable fruit = "banana" # Here we declare a string datatype </lang>

In awk multiple assignments are possible from within a single statement:

C

Local variables are generally called auto variables in C. Variables must be declared before use. The declaration of a variable, without assigning a value takes the form <typename> <variablename>; <lang c>int j;</lang> Some common types are: char, short, int, long, float, double and unsigned.

Multiple variables may be defined in a single statement as follows: <lang c>double double1, double2, double3;</lang> It is possible to initialize variables with expressions having known values when they are defined. The syntax follows the form <typename> <variablename> = <initializing expression>; <lang c>short b1 = 2500; long elwood = 3*BSIZE, jake = BSIZE -2;</lang> Strings in C are arrays of char terminated by a 0 or NULL character. To declare space for a string of up to 20 characters, the following declaration is used. <lang c>char mystring[21];</lang> The extra length leaves room for the terminating 0.

To declare an initialized string that won't be changed the following declaration may be used: <lang c>const char * mytext = "The C Language";</lang> There are more rules regarding arrays, variables containing pointers, dynamic allocation, and initialization that are to extensive to cover here.

C#

Variables in C# are very dynamic, in the form that they can be declared practically anywhere, with any scope. As in other languages, often used variables are: int, string, double etc.

They are declared with the type first, as in C: <lang csharp>int j;</lang>

Multiple variables may be defined in a single line as follows: <lang csharp>int p, a, d;</lang>

It is also possible to assign variables, either while declaring or in the program logic: <lang csharp> int a = 4; int b; int c = Func(a);

b = 5; </lang>

C++

much like C, C++ variables are declared at the very start of the program after the headers are declared. To declare a as an integer you say: the type of variable; then the variable fallowed by a semicolon ";" <lang c++>int a;</lang>

Delphi

<lang Delphi>var

 i: Integer;
 s: string;
 o: TObject;

begin

 i := 123;
 s := 'abc';
 o := TObject.Create;
 try
   // ...
 finally
   o.Free;
 end;

end;</lang>

D

<lang D> float bite = 36.321; ///_Defines a floating-point number (float), "bite", with a value of 36.321 float[3] bites; ///_Defines a static array of 3 floats float[] more_bites; ///_Defines a dynamic array of floats </lang>

DWScript

See Delphi for "classic" declaration. In DWScript, if variables have to be declared before use, but can be declared inline, and their type can also be inferred.

<lang Delphi> var i := 123; // inferred type of i is Integer var s := 'abc'; // inferred type of s is String var o := TObject.Create; // inferred type of o is TObject var s2 := o.ClassName; // inferred type of s2 is String as that's the type returned by ClassName </lang>

E

E is an impure, lexically scoped language. Variables must be defined before use (they are not created on assignment). Definition of variables is a special case of pattern matching.

An identifier occurring in a pattern is a simple non-assignable variable. The def operator is usually used to define local variables:

<lang e>def x := 1 x + x # returns 2</lang>

Assignment

The pattern var x makes x an assignable variable, and := is the assignment operator.

<lang e>def var x := 1 x := 2 x # returns 2</lang>

(As a shorthand, var x := ... is equivalent to def var x := ....)

There are update versions of the assignment operator, in the traditional C style (+=, -=, |=, etc.), but also permitting any verb (method name) to be used:

<lang e>def var x := 1 x += 1 # equivalent to x := x + 1, or x := x.add(1) x # returns 2

def var list := ["x"] list with= "y" # equivalent to list := list.with("y") list # returns ["x", "y"]</lang>

Patterns

Since variable definition is part of pattern matching, a list's elements may be distributed into a set of variables:

<lang e>def [hair, eyes, var shirt, var pants] := ["black", "brown", "plaid", "jeans"]</lang>

However, assignment to a list as in Perl or Python is not currently supported.

<lang e>[shirt, pants] := ["white", "black"] # This does not do anything useful.</lang>

Scoping

In E, a variable is visible from the point of its definition until the end of the enclosing block. Variables can even be defined inside expressions (actually, E has no statement/expression distinction):

<lang e>def list := [def x := timer.now(), x] # two copies of the current time list[0] == x # x is still visible here; returns true</lang>

Slots

The difference between assignable and non-assignable variables is defined in terms of primitive operations on non-primitive slot objects. Slots can also be employed by programmers for effects such as variables which have an effect when assigned (e.g. backgroundColor := red) or automatically change their values over time, but that is beyond the scope of this task. For example, it is possible to transfer a variable between scopes by referring to its slot:

<lang e>def makeSum() {

 var a := 0
 var b := 0
 return [&a, &b, fn { a + b }]

}

def [&x, &y, sum] := makeSum() x := 3 y := 4 sum() # returns 7</lang>

As suggested by the & syntax, the use of slots is somewhat analogous in effect to C pointers or C++ references, allowing the passing of locations and not their values, and "pass-by-reference" or "out" parameters:

<lang e>def getUniqueId(&counter) {

 counter += 1
 return counter

}

var idc := 0 getUniqueId(&idc) # returns 1 getUniqueId(&idc) # returns 2</lang>

Ela

Strictly speaking Ela doesn't have variables. Instead Ela provides a support for a declaration of names that can be bound to values. Unlike variables names are immutable - it is not possible to change a value bound to a name.

Global declarations:

<lang ela>let x = 42

let sum x y = x + y</lang>

Local declarations:

<lang ela>let sum x y = let x+y in z

let sum x y = z

             where z = x + y</lang>

Factor

The SYMBOL bit defines a new symbol word which is used to identify variables. use-foo shows how one would modify and get the contents of the variable. named-param-example is an example of using :: to define a word with named inputs, similar to the way other languages do things. Last, but not least, local-example shows how to use [let to define a group of lexically scoped variables inside of a word definition. <lang factor>SYMBOL: foo

use-foo ( -- )

   1 foo set
   foo get 2 + foo set ! foo now = 3
   foo get number>string print ;

named-param-example ( a b -- )

   a b + number>string print ;

local-example ( -- str ) [let "a" :> b "c" :> a a " " b 3append ] ;</lang>

Forth

Historically, Forth has preferred open access to the parameter stack over named local variables. The 1994 standard however added a cell-sized local variable facility and syntax. The semantics are similar to VALUEs: locals are initialized from stack contents at declaration, the name retrieves the value, and TO sets the value of the local name parsed at compile time ("value TO name"). <lang forth>: hypot ( a b -- a^2 + b^2 )

 LOCALS| b a |            \ note: reverse order from the conventional stack comment
 b b * a a * + ;</lang>

Works with: GNU Forth

Modern Forth implementations often extend this facility in several ways, both for more convenient declaration syntax and to be more compatible with foreign function interfaces. Curly braces are used to replace the conventional stack comment with a similar looking local variable declaration. <lang forth>: hypot { a b -- a^2 + b^2 } \ text between "--" and "}" remains commentary

 a a * b b * + ;</lang>

Modern systems may also allow different local data types than just integer cells. <lang forth>: length { F: a F: b F: c -- len } \ floating point locals

 a a F* b b F* F+ c c F* F+ FSQRT ;</lang>

Fortran

<lang fortran> program test

implicit none

integer :: i  !scalar integer 
integer,dimension(10) :: ivec !integer vector
real :: r !scalar real
real,dimension(10) :: rvec !real vector
character(len=:),allocatable :: char1, char2  !fortran 2003 allocatable strings

!assignments:

!-- scalars:

i = 1
r = 3.14

!-- vectors:

ivec = 1 !(all elements set to 1)
ivec(1:5) = 2

rvec(1:9) = 0.0
rvec(10) = 1.0

!-- strings:

char1 = 'hello world!'
char2 = char1   !copy from one string to another
char2(1:1) = 'H'  !change first character

end program test </lang>

GAP

<lang gap># At top level, global variables are declared when they are assigned, so one only writes global_var := 1;

In a function, local variables are declared like this

func := function(n)

   local a;
   a := n*n;
   return n + a;

end;

One can test whether a variable is assigned

IsBound(global_var);

true;

And destroy a variable

Unbind(global_var);

This works with list elements too

u := [11, 12, , 14]; IsBound(u[4]);

true

IsBound(u[3]);

false

Unbind(u[4]);</lang>

Go

Simplest and most common

While Go is statically typed, it provides a “short variable declaration” with no type explicitly stated, as in, <lang go>x := 3</lang> This is the equivalent of, <lang go>var x int // declaration x = 3 // assignment</lang> The technique of not stating the type is known as type inference, or duck typing. The right hand side can be any expression. Whatever type it represents is used as the type of the variable. More examples: <lang go>y := x+1 // y is int, assuming declaration above same := x == y // same declared as bool p := &same // type of p is pointer to bool pi := math.Floor(math.Pi) // math.Floor returns float64, so that is the type of pi</lang> Nothing goes uninitialized

Variables declared without initializer expressions are initialized to the zero value for the type. <lang go>var x, y int // two variables, initialized to zero. var p *int // initialized to nil</lang> Opposite C

While putting the variable before the type feels “backwards” to programmers familiar with certain other languages, it succinctly allows multiple variables to be declared with arbitrarily complex type expressions.

List syntax

Variables can be declared in a list with the keyword var used only once. The syntax visually groups variables and sets the declaration off from surrounding code. <lang go>var (

   x, y int
   s string

)</lang> Multiple assignment

Multiple values can be assigned in a single assignment statement, with many uses. <lang go>x, y = y, x // swap x and y sinX, cosX = math.Sincos(x) // Sincos function returns two values // map lookup optionally returns a second value indicating if the key was found. value, ok = mapObject[key]</lang> Other kinds of local variables

Parameters and named return values of functions, methods, and function literals also represent assignable local variables, as in, <lang go>func increase (x int) (more int) {

   x++
   more = x+x
   return

}</lang> Parameter x and return value more both act as local variables within the scope of the function, and are both assignable. When the function returns, both go out of scope, although the value of more is then returned as the value of the function.

While assignment of return values is highly useful, assignment of function parameters is often an error. Novice programmers might think that modifying a parameter inside the function will affect a variable used as an argument to the function call. It does not.

Method receivers also represent assignable local variables, and as with function parameters, assigning them inside the method is often a mistake.

Other common errors

Short declarations can involve multiple assignment, as in <lang go>x, y := 3, 4</lang> But there are complications involving scope and variables already defined that confuse many programmers new to Go. A careful reading of the language specification is definitely in order, and a review of misconceptions as discussed on the mailing list is also highly recommended.

Programmers new to the concept of closures often fail to distinguish between assigning free and bound variables. Function literals in Go are closures, and a common novice error is to start multiple goroutines from function literals, and fail to understand that multiple goroutines are accessing the same free variables.

Haskell

You can define a variable at the top (module) level or in a where, let, or do construct.

<lang haskell>foobar = 15

f x = x + foobar

 where foobar = 15

f x = let foobar = 15

     in  x + foobar

f x = do

   let foobar = 15
   return $ x + foobar</lang>

One particular feature of do notation looks like assignment, but actually, it's just syntactic sugar for the >>= operator and a unary lambda.

<lang haskell>main = do

   s <- getLine
   print (s, s)

-- The above is equivalent to:

main = getLine >>= \s -> print (s, s)</lang>

Pattern matching allows for multiple definitions of the same variable, in which case each call uses the first applicable definition.

<lang haskell>funkshun True x = x + 1 funkshun False x = x - 1

foobar = funkshun True 5 + funkshun False 5 -- 6 + 4</lang>

case expressions let you do pattern-matching on an arbitrary expression, and hence provide yet another way to define a variable.

<lang haskell>funkshun m = case foo m of

   [a, b]           -> a - b
   a : b : c : rest -> a + b - c + sum rest
   a                -> sum a</lang>

Guards are as a kind of syntactic sugar for if-else ladders.

<lang haskell>signum x | x > 0 = 1

        | x < 0     = -1
        | otherwise =  0</lang>

A defintion can be accompanied by a type signature, which can request a less general type than the compiler would've chosen on its own. (Because of the monomorphism restriction, there are also some cases where a type signature can request a more general type than the default.) Type signatures are also useful even when they make no changes, as a kind of documentation.

<lang haskell>dotProduct :: [Int] -> [Int] -> Int dotProduct ns ms = sum $ zipWith (+) ns ms -- Without the type signature, dotProduct would -- have a more general type.

foobar :: Num a => a foobar = 15 -- Without the type signature, the monomorphism -- restriction would cause foobar to have a less -- general type.</lang>

Since Haskell is purely functional, most variables are immutable. It's possible to create mutable variables in an appropriate monad. The exact semantics of such variables largely depend on the monad. For example, STRefs must be explicitly initialized and passed between scopes, whereas the implicit state of a State monad is always accessible via the get function.

HicEst

<lang HicEst>! Strings and arrays must be declared. ! Everything else is 8-byte float, READ/WRITE converts

 CHARACTER str="abcdef", str2*345, str3*1E6/"xyz"/
 REAL, PARAMETER :: named_constant = 3.1415
 REAL :: n=2, cols=4, vec(cols), mtx(n, cols)
 DATA vec/2,3,4,5/, mtx/1,2,3.1415,4,  5,6,7,8/

 named = ALIAS(alpha, beta, gamma) ! gamma == named(3)
 ALIAS(vec,n, subvec,2) ! share subvec and vec(n...n+1)
 ALIAS(str,3, substr,n) ! share substr and str(3:3+n-1)

 a = EXP(b + c)     ! assign/initialze a=1, b=0, c=0
 str = "blahblah"   ! truncate/expand if needed
 beta = "blahblah"  ! illegal

 CALL noArguments_noUSE   ! global scope SUBROUTINE
 CALL Arguments_or_USE(a) ! local scope SUBROUTINE
 t = func()               ! local scope FUNCTION

SUBROUTINE noArguments_noUSE() ! all global

 vec2 = $ ! 1,2,3,...

END

SUBROUTINE Arguments_or_USE(var) ! all local

 USE : vec                      ! use global object
 var = SUM(vec)
 t = TIME()         ! local, static, USEd by func()

END

FUNCTION func() ! all local

 USE Arguments_or_USE : t       ! use local object
 func = t

END</lang>

Icon and Unicon

Icon/Unicon data types are implemented as type safe self-descriptive values and as such do not require conventional type declarations. See Introduction to Unicon and Icon about declarations

Declarations are confined to scope and use and include local, static, global, procedure parameters, and record definitions. Additionally Unicon has class definitions. Undeclared variables are local by default. <lang Icon>global gvar # a global

procedure main(arglist) # arglist is a parameter of main local a,b,i,x # a, b, i, x are locals withing main static y # a static (silly in main)

x := arglist[1] a := 1.0 i := 10 b := [x,a,i,b]

... rest of program

end</lang>

Icon

Unicon

This Icon solution works in Unicon.

J

J has two assignment operators. The =. operator declares, initializes, assigns, etc. a local variable. The =: operator does the same for a "global" variable.

<lang j>fun =: 3 :0

 val1 =: 0
 val1 =. 2
 val2 =. 3
 val1, val2

)

fun

2 3

  val1

0

  val2

|value error</lang>

Note that the language forbids assigning a "global" value in a context where the name has a local definition.

<lang j>fun1 =: 3 :0

 val3=. 0
 val3=: 0

)

  fun1

|domain error</lang>

But the purpose of this rule is to help people catch mistakes. If you have reason to do this, you can easily set up another execution context.

<lang j>fun2 =: 3 :0

 val4=. 0
 3 :'val4=:y' y

)

  fun2 </lang>

Variables are referred to by name, and exist in locales (which may be used as classes, closures or other stateful references).

FIXME (working on good illustrative examples that would make sense to someone used to different languages)

That said, it is possible and not uncommon to write an entire J application without using any variables (J has a functional, "point free" style of coding known as tacit). Names are optional (though often convenient). And, it can be possible to build code using names and then remove them using f. -- this is somewhat analogous to compiling code though the implementation of f. does not have to compile anything.

Java

Variables in Java are declared before their use with explicit types: <lang java>int a; double b; AClassNameHere c;</lang> Several variables of the same type can be declared together: <lang java>int a, b, c;</lang> Variables can be assigned values on declaration or afterward: <lang java>int a = 5; double b; int c = 5, d = 6, e, f; String x = "test"; String y = x; b = 3.14;</lang> Variables can have scope modifiers, which are explained here.

final variables can only be assigned once, but if they are Objects or arrays, they can be modified through methods (for Objects) or element assignment (for arrays): <lang java>final String x = "blah"; final String y; final double[] nums = new double[15]; y = "test"; x = "blahblah"; //not legal nums[5] = 2.5; //legal nums = new double[10]; //not legal final Date now = new java.util.Date(); now.setTime(1234567890); //legal now = new Date(1234567890); //not legal</lang>

JavaScript

Information lifted from Stack Overflow (credit to krosenvold and triptych)

Javascript uses scope chains to establish the scope for a given function. There is typically one global scope, and each function defined has its own nested scope. Any function defined within another function has a local scope which is linked to the outer function. It's always the position in the source that defines the scope.

An element in the scope chain is basically a Map with a pointer to it's parent scope.

When resolving a variable, javascript starts at the innermost scope and searches outwards.

<lang javascript>// a globally-scoped variable

var a=1;

// global scope function one(){

   alert(a);

}

// local scope function two(a){

   alert(a);

}

// local scope again function three(){

 var a = 3;
 alert(a);

}

// Intermediate: no such thing as block scope in javascript function four(){

   if(true){
       var a=4;
   }

   alert(a); // alerts '4', not the global value of '1'

}

// Intermediate: object properties function Five(){

   this.a = 5;

}

// Advanced: closure var six = function(){

   var foo = 6;

   return function(){
       // javascript "closure" means I have access to foo in here, 
       // because it is defined in the function in which I was defined.
       alert(foo);
   }

}()

// Advanced: prototype-based scope resolution function Seven(){

 this.a = 7;

}

// [object].prototype.property loses to [object].property in the scope chain Seven.prototype.a = -1; // won't get reached, because 'a' is set in the constructor above. Seven.prototype.b = 8; // Will get reached, even though 'b' is NOT set in the constructor.

// These will print 1-8 one(); two(2); three(); four(); alert(new Five().a); six(); alert(new Seven().a);

alert(new Seven().b);</lang>

Joy

JOY does not have variables. Variables essentially name locations in memory, where values are stored. JOY also uses memory to store values, but has no facility to name these locations. The memory that JOY uses is commonly referred to as "the stack".

Initializing

The JOY stack can be initialized: <lang joy>[] unstack</lang>

Assignment

Values can be pushed on the stack: <lang joy>42</lang> pushes the value 42 of type integer on top of the stack.

Stack

Calling the stack by name pushes a copy of the stack on the stack. To continue the previous example: <lang joy>stack</lang> pushes the list [42] on top of the stack. The stack now contains: [42] 42.

Liberty BASIC

<lang lb> 'In Liberty BASIC variables are either string or numeric. 'A variable name can start with any letter and it can contain both letters and numerals, as well as dots (for example: user.firstname). 'There is no practical limit to the length of a variable name... up to ~2M characters. 'The variable names are case sensitive.

'assignments: -numeric variables. LB assumes integers unless assigned or calculated otherwise. 'Because of its Smalltalk heritage, LB integers are of arbitrarily long precision. 'They lose this if a calculation yields a non-integer, switching to floating point.

   i = 1
   r = 3.14

'assignments -string variables. Any string-length, from zero to ~2M.

   t$    ="21:12:45"
   flag$ ="TRUE"

'assignments -1D or 2D arrays 'A default array size of 10 is available. Larger arrays need pre-'DIM'ming.

   height( 3)          =1.87
   dim height( 50)
   height( 23)         =123.5
   potential( 3, 5)    =4.5
   name$( 4)           ="John"

'There are no Boolean /bit variables as such.

'Arrays in a main program are global. 'However variables used in the main program code are not visible inside functions and subroutines. 'They can be declared 'global' if such visibility is desired. 'Functions can receive variables by name or by reference. </lang>

Logo

Historically, Logo only had global variables, because they were easier to access when stepping through an algorithm. Modern variants have added dynamic scoped local variables.

Works with: UCB Logo

<lang logo>make "g1 0 name 2 "g2 ; same as make with parameters reversed global "g3 ; no initial value to func :x

 make "g4 4   ; still global
 localmake "L1 6
 local ["L2 "L3]    ; local variables, collection syntax
 func2 :g4
 print :L2      ; 9,  modified by func2
 print :L3      ; L3 has no value, was not modified by func2

end to func2 :y

 make "g3 :y
 make "L2 :L1 + 3     ; dynamic scope: can see variables of callers
 localmake "L3 5       ; locally override L3 from caller
 (print :y :L1 :L2 :L3)      ; 4 6 9 5

end print :g4 ; 4 print :L1 ; L1 has no value print name? "L1 ; false, L1 is not bound in the current scope</lang>

LotusScript

<lang Lotusscript>Sub Click() 'a few declarations as example Dim s as New NotesSession ' declaring a New NotesSession actually returns the current, active NotesSession Dim i as Integer ' i = 0 Dim s as String ' s= "" Dim v as Variant ' v is nothing Dim l as Long ' l = 0 Dim doc as NotesDocument 'doc is EMTPY

'...

End Sub </lang>

Lua

In lua, variables are dynamically typecast, and do not need declaration prior to use.

<lang lua>a = 1 -- Here we declare a numeric variable fruit = "banana" -- Here we declare a string datatype needspeeling = True -- This is a boolean local b = 2 -- This variable declaration is prefixed with a scope modifier </lang>

The lua programming language supports multiple assignments from within a single statement:

<lang lua>A, B, C, D, E = 2, 4, 6, 8, "It's never too late"</lang>

Mathematica

x=value	assign a value to the variable x
x=y=value	assign a value to both x and y
x=. or Clear[x]	remove any value assigned to x

lhs=rhs (immediate assignment)	rhs is evaluated when the assignment is made
lhs:=rhs (delayed assignment)	rhs is evaluated each time the value of lhs is requested

Atomic Objects

All expressions in Mathematica are ultimately made up from a small number of basic or atomic types of objects. 

Symbol / String / Integer / Real / Rational / Complex

These objects have heads which are symbols that can be thought of as "tagging" their types. 
The objects contain "raw data", which can usually be accessed only by functions specific to the particular type of object. 
You can extract the head of the object using Head, but you cannot directly extract any of its other parts.

Symbols are the basic named objects in Mathematica

aaaaa	user-defined symbol
Aaaaa	system-defined symbol
$Aaaa	global or internal system-defined symbol
aaaa$	symbol renamed in a scoping construct
aa$nn	unique local symbol generated in a module

Contexts

aaaa`x is a symbol with short name x, and context aaaa. 
Contexts in Mathematica work somewhat like file directories in many operating systems. 
You can always specify a particular file by giving its complete name, including its directory. 
But at any given point, there is usually a current working directory, analogous to the current Mathematica context. 
Files that are in this directory can then be specified just by giving their short names.

Scoping Constructs

With[] evaluate with specified variables replaced by values 
Module[] localize names of variables (lexical scoping)
Block[] localize values of variables (dynamic scoping)
DynamicModule[] localize names of variables in dynamic interface constructs

Other Forms of Scoping
Begin, End  localize symbol namespace
Throw, Catch  localize exceptions
Quiet, Check localize messages
BlockRandom localize pseudorandom variables

MATLAB / Octave

<lang Matlab> a = 4; % declare variable and initialize double value,

       s = 'abc'; % string 
       i8 = int8(5);	% signed byte 
       u8 = uint8(5);	% unsigned byte
       i16 = int16(5);	% signed 2 byte 
       u16 = uint16(5); % unsigned 2 byte integer
       i32 = int32(5);	% signed 4 byte integer
       u32 = uint32(5);% unsigned 4 byte integers 
       i64 = int64(5);	% signed 8 byte integer
       u64 = uint64(5);% unsigned 8 byte integer

f32 = float32(5); % single precission floating point number f64 = float64(5); % double precission floating point number , float 64 is the default data type.

c = 4+5i; % complex number

       colvec = [1;2;4];   % column vector 
       crowvec = [1,2,4];   % row vector 
       m = [1,2,3;4,5,6];  % matrix with size 2x3</lang>

Variables within functions have local scope, except when they are declared as global

<lang Matlab> global b </lang>

Modula-3

<lang modula3>MODULE Foo EXPORTS Main;

IMPORT IO, Fmt;

VAR foo: INTEGER := 5; (* foo is global (to the module). *)

PROCEDURE Foo() =

 VAR bar: INTEGER := 10; (* bar is local to the procedure Foo. *)
 BEGIN
   IO.Put("foo + bar = " & Fmt.Int(foo + bar) & "\n");
 END Foo;

BEGIN

 Foo();

END Foo.</lang>

For procedures, the formal parameters create local variables unless the actual parameter is prefixed by VAR: <lang modula3>PROCEDURE Foo(n: INTEGER) =</lang> Here, n will be local to the procedure Foo, but if we instead wrote: <lang modula3>PROCEDURE Foo(VAR n: INTEGER) =</lang> Then n is the global variable n (if it exists).

Objeck

Different ways to declare and initialize an integer. <lang objeck> a : Int; b : Int := 13; c := 7; </lang>

OCaml

The default handlers for values in OCaml are not variables strictly speaking, because as OCaml is a functional language these values can't vary (so are not variable). Strictly speaking these are bindings. An identifier is bound to a value in an immutable way.

The standard way to bind an identifier to a value is the let construct: <lang ocaml>let x = 28</lang>

This stated, ocaml programmers most often use the word variable when they refer to bindings, because in the programming world we usually use this word for the default values handlers.

Now to add confusion, real variables also exist in OCaml because it is an impure functional language. They are called references and are defined this way: <lang ocaml>let y = ref 28</lang> References can then be accessed and modified this way: <lang ocaml> !y (* access *)

 y := 34  (* modification *)</lang>

An identifier can not be declared uninitialised, it is always defined with an initial value, and this initial value is used by the OCaml type inference to infer the type of the binding.

Inside an expression, bindings are defined with the let .. in construct, and we can also define multiple bindings with the let .. and .. in construct (here the expression can be the definition of a new identifier or the definition of a function): <lang ocaml>let sum = (* sum is bound to 181 *)

 let a = 31
 and b = 150 in
 (a + b)

let sum () = (* sum is a function which returns 181 *)

 let a = 31
 and b = 150 in
 (a + b)</lang>

Openscad

<lang openscad> mynumber=5+4; // This gives a value of nine </lang>

Oz

Variable names in Oz always start with an uppercase letter.

Oz variables are dataflow variables. A dataflow variable can basically be free (unbound) or determined (has a value). Once a value has been assigned, it can not be changed. If we assign the same value again, nothing happens. If we assign a different value to an already determined variable, an exception is raised: <lang oz>declare Var %% new variable Var, initially free {Show Var} Var = 42 %% now Var has the value 42 {Show Var} Var = 42 %% the same value is assigned again: ok Var = 43 %% a different value is assigned: exception</lang>

In the Emacs-based interactive environment, declare creates a new open scope in which variables can be declared. The variables are visible for the entire rest of the session.

Most operations on free variables block until the variables have been bound (but not Show as used above).

Assignment to dataflow variables is also called unification. It is actually a symmetric operation, e.g. the following binds B to 3: <lang oz>declare

 A = 3
 B

in

 A = B
{Show B}</lang>

However, variables can only be introduced at the left side of the = operator. So this is a syntax error: <lang oz>declare

 A = 3
 A = B  %% Error: variable B not introduced

in

{Show B}</lang>

It is possible to introduce multiple variables in a single statement: <lang oz>declare

  [A B C D] = [1 2 3 4]  %% unification of two lists</lang>

In a module definition, toplevel variables can be introduced between the keywords define and end without the need for declare. The range between these two keywords is also their scope. Toplevel variables can optionally be exported. <lang oz>functor export Function define

  ToplevelVariable = 42

  fun {Function}
    42
  end

end</lang>

Function and class definitions introduce a new variable with the name of the function/class and assign the new function/class to this variable.

Most Oz statement introduce a new scope and it is possible to introduce local variables at the top of this scope with the in keyword. <lang oz>fun {Function Arg}

  LocalVar1

in

  LocalVar1 = if Arg == 42 then

LocalVar2 in LocalVar2 = yes LocalVar2 else LocalVar3 = no %% variables can be initialized when declared in LocalVar3 end

  LocalVar1

end</lang> Here, LocalVar1 is visible in the whole body of Function while LocalVar2 is only visible in the then branch and LocalVar3 is only visible in the else branch.

Additionally, new local variables can be introduced everywhere using the keyword local. <lang oz>if {IsEven 42} then

  {System.showInfo "Here, LocalVar is not visible."}
  local
     LocalVar = "Here, LocalVar IS visible"
  in
     {System.showInfo LocalVar}
  end

end</lang>

New variables are also introduced in pattern matching. <lang oz>case "Rosetta code" of First|_ then {Show First} end %% prints "R"</lang> _ creates a new nameless variable that is initially unbound. It is usually pronounced "don't care".

It is possible to create a read-only view of a variable with the !! operator. This is called a "future". We can wait for such a variable to become bound by another thread and we can read its value, but we can never set it. <lang oz>declare

 A
 B = !!A %% B is a read-only view of A

in

 thread
    B = 43 %% this blocks until A is known; then it fails because 43 \= 42
 end
 A = 42</lang>

Additional operations on variables: <lang oz>declare

 V = 42

in

 {Wait V}  %% explicitly wait for V to become determined

 if {IsDet V} then  %% check whether V is determined; not recommended
    {Show determined}
 elseif {IsFree V} then  %% check whether V is free; not recommended
    {Show free}
 end</lang>

IsFree and IsDet are low-level functions. If you use them, you code is no longer declarative and prone to race conditions when used in a multi-threaded context.

To have mutable references like in imperative languages, use cells: <lang oz>declare

 A = {NewCell 42}
 OldVal

in

 {Show @A}         %% read a cell with @
 A := 43           %% change its value
 OldVal = A := 44  %% read and write at the same time (atomically)</lang>

A is an immutable dataflow variable that is bound to a mutable reference.

PARI/GP

There are two types of local variables, local (mostly deprecated) and my. Variables can be used without declaration or initialization; if not previously used such a variable is a pure variable: technically, a monomial in a variable with name equal to the variable name. This behavior can be forced with the apostrophe operator: regardless of the value (if any) currently stored in x, <lang parigp>'x</lang> displays as (and is treated internally as) x. This is useful when you want to use it as a variable instead of a number (or other type of object). For example, <lang parigp>'x^3+7</lang> is a cubic polynomial, not the number 8, even if x is currently 1.

Pascal

See Delphi

Perl

Variables can be declared with our, my, or local, or they can be used without being declared at all; see scope modifiers for the differences. In any case, variables which haven't been assigned to have the undefined value by default. The undefined value acts just like 0 (if used as a number) or the empty string (if used as a string), except it can be distinguished from either of these with the defined function. Also, if warnings are enabled, perl will print a message like "Use of uninitialized value $foo in addition (+)" whenever you use the undefined value as a number or string.

Initialization and assignment are the same thing in Perl: just use the = operator. Note that the rvalue's context (scalar or list) is determined based on the lvalue.

<lang perl>my $x = @a; # Scalar assignment; $x is set to the

                             # number of elements in @a.

my ($x) = @a; # List assignment; $x is set to the first

                             # element of @a.

my @b = @a; # List assignment; @b becomes the same length

                             # as @a and each element becomes the same.

my ($x, $y, @b) = @a; # List assignment; $x and $y get the first

                             # two elements of @a, and @b the rest.

my ($x, $y, @b, @c, $z) = @a; # Same thing, and also @c becomes empty

                             # and $z undefined.</lang>

The kind of value a variable can hold depends on its sigil, "sigil" being a slang term for "funny character in front of a variable name". $dollarsigns can hold scalars: the undefined value, numbers, strings, or references. @atsigns can hold arrays of scalars, and %percentsigns can hold hashes of scalars (associative arrays mapping strings to scalars); nested data structures are constructed by making arrays or hashes of references to arrays or hashes.

There are two other sigils, but they behave quite unlike the others. A token of the form &foo refers to a subroutine named foo. In older versions of Perl, ampersands were necessary for calling user-defined subroutines, but since they no longer are, they have only a handful of obscure uses, like making references to named subroutines. Note that you can't assign to an ampersand-marked name. But you can assign to a typeglob, a kind of object represented with the notation *var. A typeglob *foo represents the symbol-table entry for all of the otherwise independent variables $foo, @foo, %foo, and &foo. Assigning a string "bar" to *foo makes these variables aliases for $bar, @bar, %bar, and &bar respectively. Alternatively, you can assign a reference to a typeglob, which creates an alias only for the variable of the appropriate type. In particular, you can say *twiddle = sub {...} to change the definition of the subroutine &twiddle without affecting $twiddle and friends.

Perl 6

Much of what is true for Perl 5 is also true for Perl 6. Some exceptions:

There are no typeglobs in Perl 6.

Assigning an array to a scalar variable now makes that scalar variable a reference to the array:

<lang Perl6> my @y = <A B C D>; #Array of strings 'A', 'B', 'C', and 'D'

my $x = @y; # $x is now a reference for the array @y

say $x[1]; # prints 'B' follow by a newline character </lang>

Types and constraints can also be applied to variables in Perl 6:

   # $x can contain only Int objects
   my Int $x;

   # $x can only contain native integers (not integer objects)
   my int $x;

   #A variable may itself be bound to a container type that specifies how the container works, without specifying what kinds of things it contains.
   # $x is implemented by the MyScalar class
   my $x is MyScalar;

   #Constraints and container types can be used together:
   # $x can contain only Int objects,
   # and is implemented by the MyScalar class
   my Int $x is MyScalar;

</lang>

(Includes code modified from http://perlcabal.org/syn/S02.html#Built-In_Data_Types. See this reference for more details.)

(Much more can and should be said here).

PL/I

<lang PL/I> /* The PROCEDURE block and BEGIN block are used to delimit scopes. */

declare i float; /* external, global variable, excluded from the */

                /* local ares (BEGIN block) below.              */

begin;

  declare (i, j) fixed binary; /* local variable */
  get list (i, j);
  put list (i,j);

end;

/* Examples of initialization. */

declare p fixed initial (25); declare q(7) fixed initial (9, 3, 5, 1, 2, 8, 15);

  /* sets all elements of array Q at run time, on block entry. */

declare r(7) fixed initial (9, 3, 5, 1, 2, 8, 15);

  /* sets all elements of array R at compile time. */

p = 44; /* run-time assignment. */ q = 0; /* run-time initialization of all elements of Q to zero. */ q = r; /* run-time assignment of all elements of array R to */

      /* corresponding elemets of S.                           */

</lang>

PicoLisp

You can control the local bindings of symbols with functions like 'use' or 'let': <lang PicoLisp>(use (A B C)

  (setq A 1  B 2  C 3)
  ... )</lang>

This is equivalent to <lang PicoLisp>(let (A 1 B 2 C 3)

  ... )</lang>

The parentheses can be omitted if there is only a single variable <lang PicoLisp>(use A

  (setq A ..)
  ... )

(let A 1

  ...)</lang>

Other functions that handle local bindings are 'let?', 'bind', 'job', 'with' or 'for'.

PureBasic

<lang PureBasic>; Variables are initialized when they appear in sourcecode with default value of 0 and type int Debug a

or value "" for a string, they are not case sensitive

Debug b$

This initializes a double precision float, if type is following the dot

Debug c.d

They can be initialized with define (double precision float, string, integer)

Define d.d = 3.5, e$ = "Test", f.i = a + 2

Define can have a default type (all bytes except j which is long)

Define.b g, h, j.l

Define without following variables sets default type. In this case to single precision float

Define.f

So this will be an single precision float and no integer

Debug k

EnableExplicit forces declaration of used variables with define

EnableExplicit

Will throw an error because L isn't initialized

Debug L DisableExplicit

Global Variables are available in Procedures and Threads too

Global M = 3, N = 2 Procedure Dummy(parameter1, parameter2 = 20)

 ; Parameter contain values which where used when calling the function,
 ; their types have to be specified in the above Procedure header.
 ; The last ones can have default values which get applied if this parameter is not given.

 ; Variables in Procedures are separate from those outside,
 ; so d can be initialized again with another type
 ; which would otherwise lead to an error
 d.i
 ; Protected makes a variable local even if another one with same name is declared as global (see above)
 Protected M = 2
 ; Shares a variable with main program like it was declared by global
 Shared a
 ; prevents a variable to be initialized with default value again when procedure is called a second time,
 ; could be used for example as a counter, which contains the number of times a function was called
 Static a
 ; N here also would have a value of 2, while for example
 ; f would, when named, initialize a new variable, and so have a value of 0

EndProcedure

finally there are constants which are prefixed by an #

Test = 1

Their value cannot be changed while program is running

String_Constant = "blubb"

In constrast to variables, a constant has no types except an (optional) $ sign to mark it as string constant

Float_Constant = 2.3

Maps, LinkedLists , Arrays and Structures are not handled here, because they are no elemental variables</lang>

PowerShell

Variables in PowerShell start with a $ character, they are created on assignment and thus don't need to be declared: <lang powershell>$s = "abc" $i = 123</lang> Uninitialized variables expand to nothing. This may be interpreted for example as an empty string or 0, depending on context: <lang powershell>4 + $foo # yields 4 "abc" + $foo + "def" # yields "abcdef"</lang> Variables all show up in the Variable: drive and can be queried from there with the usual facilities: <lang powershell>Get-ChildItem Variable:</lang> Since Variables are provided via a flat filesystem, they can be manipulated using the common cmdlets for doing so. For example to delete a variable one can use <lang powershell>Remove-Item Variable:foo</lang> as if it were a file or a registry key. There are, however, several cmdlets dealing specifically with variables: <lang powershell>Get-Variable # retrieves the value of a variable New-Variable # creates a new variable Set-Variable # sets the value of a variable Clear-Variable # deletes the value of a variable, but not the variable itself Remove-Variable # deletes a variable completely</lang>

Python

Names in Python are not typed, although all the objects referred to by them, are. Names are lexically scoped by function/method/class definitions, and must be defined before use.

Names in global statements are looked up in the outermost context of the program or module. Names in a nonlocal statement are looked up in the order of closest enclosing scope outwards.

R

Variables are dynamically typed, so they do not need to be declared and instantiated separately. <- and = are both used as the assignment operator, though <- is preferred, for compatibility with S-Plus code. <lang R>foo <- 3.4 bar = "abc"</lang> It is possible to assign multiple variables with the same value, and to assign values from left to right. <lang R>baz <- quux <- 1:10 TRUE -> quuux</lang> There are also global assignment operators, <<- and ->>. From their help page:

The operators '<<-' and '->>' cause a search to made through the
environment for an existing definition of the variable being
assigned.  If such a variable is found (and its binding is not
locked) then its value is redefined, otherwise assignment takes
place in the global environment.

In practice, this usually means that variables are assigned in the user workspace (global environment) rather than a function. <lang R>a <- 3

assignmentdemo <- function() {

  message("assign 'a' locally, i.e. within the scope of the function")
  a <- 5
  message(paste("inside assignmentdemo, a = ", a))
  message(paste("in the global environment, a = ", get("a", envir=globalenv())))
  
  message("assign 'a' globally")
  a <<- 7
  message(paste("inside assignmentdemo, a = ", a))
  message(paste("in the global environment, a = ", get("a", envir=globalenv())))

} assignmentdemo()</lang>

assign 'a' locally, i.e. within the scope of the function
inside assignmentdemo, a =  5
in the global environment, a =  3
assign 'a' globally
inside assignmentdemo, a =  5
in the global environment, a =  7

Finally, there is also the assign function, where you choose the environment to assign the variable. <lang R>assign("b", TRUE) #equivalent to b <- TRUE assign("c", runif(10), envir=globalenv()) #equivalent to c <<- runif(10)</lang>

Rascal

The effect of a variable declaration is to introduce a new variable Name and to assign the value of expression Exp to Name. A variable declaration has the form <lang rascal> Type Name = Exp;</lang> A mention of Name later on in the same scope will be replaced by this value, provided that Name’s value has not been changed by an intermediate assignment. When a variable is declared, it has as scope the nearest enclosing block, or the module when declared at the module level.

There are two rules you have to take into account. Double declarations in the same scope are not allowed. Additionally, the type of Exp should be compatible with Type, i.e., it should be a subtype of Type.

As a convenience, also declarations without an initialization expression are permitted inside functions (but not at the module level) and have the form <lang rascal>Type Name;</lang> and only introduce the variable Name.

Rascal provides local type inference, which allows the implicit declaration of variables that are used locally in functions. There are four rules that apply when doing so. (1) An implicitly declared variable is declared at the level of the current scope, this may the whole function body or a block nested in it. (2) An implicitly declared variable gets as type the type of the first value that is assignment to it. (3) If a variable is implicitly declared in different execution path of a function, all these implicit declarations should result in the same type. (4) All uses of an implicitly declared variable must be compatible with its implicit type.

Examples

Two explicit variable declarations: <lang rascal>rascal>int max = 100; int: 100 rascal>min = 0; int: 0</lang>

An implicit variable declaration <lang rascal>rascal>day = {<"mon", 1>, <"tue", 2>, <"wed",3>, >>>>>>> <"thu", 4>, <"fri", 5>, <"sat",6>, <"sun",7>}; rel[str, int]: {

 <"thu",4>,
 <"mon",1>,
 <"sat",6>,
 <"wed",3>,
 <"tue",2>,
 <"fri",5>,
 <"sun",7>

}</lang>

Variable declaration and assignment leading to type error <lang rascal>rascal>int month = 12; int: 12 rascal>month ="December"; |stdin:///|(7,10,<1,7>,<1,17>): Expected int, but got str</lang>

Pitfalls Local type inference for variables always uses the smallest possibe scope for a variable; this implies that a variable introduced in an inner scope is not available outside that scope. Here is how things can go wrong: <lang rascal>rascal>if( 4 > 3){ x = "abc"; } else { x = "def";} str: "abc" rascal>x; |stdin:///|(0,1,<1,0>,<1,1>): Undeclared variable, function or constructor: x</lang>

REXX

REXX has only one type of variables: string (character).
There is no need to declare anything (indeed, there is no way to declare anything).
To assign some data (value) to a variable, just assign it: <lang rexx> aa=10 /*assigns chars 10 to AA */ bb= /*assigns a null value to BB */ cc=2*10 /*assigns charser 20 to CC */ dd='Adam' /*assigns chars Adam to DD */ ee="Adam" /*same as above to EE */ ff=10. /*assigns chars 10. to FF */ gg='10.' /*same as above to GG */ hh=+10 /*assigns chars +10 to hh */ ii=1e1 /*assigns chars 1e1 to ii */ jj=+.1e+2 /*assigns chars +.1e+2 to jj */ </lang> Variables aa, ff, gg, hh, ii, and jj will all be considered equal in REXX.

Other ways to assign values: <lang rexx> kk='123'x /*assigns hexadecimal 00000123 to KK */ kk='dead beaf'X /*assigns hexadecimal deadbeaf to KK */ ll='0000 0010'b /*assigns a blank to LL (if ASCII) */ mm='0000 0100'B /*assigns a blank to MM (if EBCDIC) */

xxx='11 2. 333 -5' parse var xxx nn oo pp qq rr

                    /*assigns   11   to NN */
                    /*assigns   2.   to OO */
                    /*assigns  333   to PP */
                    /*assigns   -5   to QQ */
                    /*assigns "null" to RR */

</lang> There are methods to catch unassigned variables in REXX. <lang REXX> signal on novalue /*usually, this statement is placed at the start of the program.*/

xxx=aaaaa /*tries to assign variable aaaaa to xxx*/ say xxx 'or somesuch' exit

novalue: /*you can dress up this error presentation better. */ badLine =sigl /*the REXX statement # that failed. */ badSource=sourceline(badLine) /*the REXX statement that failed. */ badVar =condition('D') /*the REXX variable that's not defined.*/ say say '*** error! ***' say 'undefined variable' badvar "at REXX statement number" badLine say say badSource say exit 13 </lang> Output:


*** error! ***
undefined variable AAAAA at REXX statement number 4

xxx=aaaaa  /*tries to assign variable aaaaa to xxx*/

Ruby

Information taken from Variables page at the Ruby User's Guide

Ruby has three kinds of variables, one kind of constant and exactly two pseudo-variables. The variables and the constants have no type. While untyped variables have some drawbacks, they have many more advantages and fit well with ruby's quick and easy philosophy.
Variables must be declared in most languages in order to specify their type, modifiability (i.e., whether they are constants), and scope; since type is not an issue, and the rest is evident from the variable name as you are about to see, we do not need variable declarations in ruby.
The first character of an identifier categorizes it at a glance:

$ global variable

@ instance variable

[a-z] or _ local variable

[A-Z] constant

The only exceptions to the above are ruby's pseudo-variables: self, which always refers to the currently executing object, and nil, which is the meaningless value assigned to uninitialized variables. Both are named as if they are local variables, but self is a global variable maintained by the interpreter, and nil is really a constant. As these are the only two exceptions, they don't confuse things too much.

Referencing an undefined global or instance variable returns nil. Referencing an undefined local variable throws a NameError exception.

<lang ruby>$a_global_var = 5 class Demo

 @@a_class_var = 6
 A_CONSTANT = 8
 def initialize
   @an_instance_var = 7
 end
 def incr(a_local_var)
   @an_instance_var += a_local_var
 end

end</lang>

Seed7

Seed7 variables must be defined with type and initialisation value, before they are used. There are global variables and variables declared local to a function. <lang seed7>$ include "seed7_05.s7i";

var integer: foo is 5; # foo is global

const proc: aFunc is func

 local
   var integer: bar is 10;   # bar is local to aFunc
 begin
   writeln("foo + bar = " <& foo + bar);
 end func;

const proc: main is func

 begin
   aFunc;
 end func;</lang>

SNOBOL4

Local variables in Snobol are declared in a function definition prototype string:

<lang SNOBOL4> define('foo(x,y)a,b,c') :(foo_end) foo a = 1; b = 2; c = 3

       foo = a * ( x * x ) + b * y + c :(return)

foo_end</lang>

This defines a function foo( ) taking two arguments x,y and three localized variables a,b,c. Both the argument parameters and vars are dynamically scoped to the function body, and visible to any called functions within that scope. The function name also behaves as a local variable, and may be assigned to as the return value of the function. Any variable initialization or assignment is done explicitly within the function body. Unassigned variables have a null string value, which behaves as zero in numeric context.

Snobol does not support static or lexical scoping, or module level namespaces. Any variables not defined in a prototype are global to the program.

Tcl

Tcl's variables are local to procedures, lambdas and methods by default, and there is no initialization per se: only assignment when the variable previously did not exist.

Demonstrating: <lang tcl>namespace eval foo {

   # Define a procedure with two formal arguments; they are local variables
   proc bar {callerVarName argumentVar} {
       ### Associate some non-local variables with the procedure
       global globalVar;      # Variable in global namespace
       variable namespaceVar; # Variable in local (::foo) namespace
       # Access to variable in caller's context; may be local or global
       upvar 1 callerVarName callerVar

       ### Reading a variable uses the same syntax in all cases
       puts "caller's var has $callerVar"
       # But global and namespace vars can be accessed by using qualified names
       puts "global var has $globalVar which is $::globalVar"

       ### Writing a variable has no special syntax
       ### but [set] is by far the most common command for writing
       set namespaceVar $globalVar
       incr globalVar

       ### Destroying a variable is done like this
       unset argumentVar
   }

}</lang> The main thing to note about Tcl is that the "$" syntax is a language level operator for reading a variable and not just general syntax for referring to a variable.

TI-83 BASIC

Variables will remain global, even after the program is complete. Global variables persist until deleted (or reset or power loss, unless they are archived).

Variables may be assigned with the → to a value.

1→A

</lang>

TI-89 BASIC

A variable not declared local (to a program or function) is global. Global variables are grouped into folders of which one is current at any given time. Global variables persist until deleted (or reset or power loss, unless they are archived).

<lang ti89b>Local mynum, myfunc</lang>

Variables may be assigned with the → or Define statements, both of which assign a new value to a variable. → is typically used interactively, but only Define can assign programs or multi-statement functions.

 If x < 0 Then
   Return –x
 Else
   Return x
 EndIf

EndFunc</lang>

TXR

Variables have a form of pervasive dynamic scope in TXR. Each statement ("directive") of the query inherits the binding environment of the previous, invoking, or surrounding directive, as the case may be. The initial contents of the binding environment may be initialized on the interpreter's command line. The environment isn't simply a global dictionary. Each directive which modifies the environment creates a new version of the environment. When a subquery fails and TXR backtracks to some earlier directive, the original binding environment of that directive is restored, and the binding environment versions generated by backtracked portions of the query turn to garbage.

Simple example: the cases

<lang txr>@(cases) hey @a how are you @(or) hey @b long time no see @(end)</lang> This directive has two clauses, matching two possible input cases, which have a common first line. The semantics of cases is short-circuiting: the first successful clause causes it to succeed and stop processing subsequent clauses. Suppose that the input matches the second clause. This means that the first clause will also match the first line, thereby establishing a binding for the variable a. However, the first clause fails to match on the second line, which means that it fails. The interpreter then moves to the second clause, which is tried at the original input position, under the original binding environment which is devoid of the a variable. Whichever clause of the cases is successful will pass both its environment modifications and input position increment to the next element of the query.

Under some other constructs, environments may be merged:

<lang txr>@(maybe) @a bar @(or) foo @b @(end)</lang>

The maybe directive matches multiple clauses such that it succeeds no matter what, even if none of the clauses succeed. Clauses which fail have no effect, but the effects of all successful clauses are merged. This means that if the input which faces the above maybe is the line "foo bar", the first clause will match and bind a to foo, and the second clause will also match and bind b to bar. The interpreter integrates these results together and the environment which emerges has both bindings.

UNIX Shell

Works with: Bourne Shell

!/bin/sh
The unix shell uses typeless variables

apples=6

pears=5+4 # Some shells cannot perform addition this way

pears = `expr 5+4` # We use the external expr to perform the calculation myfavourite="raspberries" </lang>

XSLT

Although called variables, XSLT "variable" elements are single-assignment, and so behave more like constants. They are valid in the node scope in which they are declared. <lang xml><xsl:variable name="foo" select="XPath expression" /> <xsl:if test="$foo = 4">... </xsl:if> </lang>