Call a function: Difference between revisions

Content added Content deleted

Inline

Revision as of 16:26, 21 October 2015

The task is to demonstrate the different syntax and semantics provided for calling a function. This may include:

Calling a function that requires no arguments
Calling a function with a fixed number of arguments
Calling a function with optional arguments
Calling a function with a variable number of arguments
Calling a function with named arguments
Using a function in statement context
Using a function in first-class context within an expression
Obtaining the return value of a function
Distinguishing built-in functions and user-defined functions
Distinguishing subroutines and functions
Stating whether arguments are passed by value or by reference
Is partial application possible and how

This task is not about defining functions.

ActionScript

<lang actionscript> myfunction(); /* function with no arguments in statement context */

 myfunction(6,b);    // function with two arguments in statement context
 stringit("apples");    //function with a string argument</lang>

Ada

Ada provides two kinds of subroutines: procedures, without return values, and functions, with return values. The return values of procedures must be used by the callers. If you don't want do deal with the return value, call a procedure instead.

As a rule of thumb, an Ada compiler is free to pass arguments either by value or by reference. Parameters have a mode, however: either 'in' or 'out' or 'in out'. It is prohibited to write somthing to an 'in' parameter. The next language Standard, Ada 2012, will support functions with 'out' and 'in out' mode parameters, so far, only procedures could have parameters with non-'in' modes. So any of the following statements for Ada functions holds for Ada procedures as well.

There are no differences between between calling built-in vs. user defined functions.

Functions without parameters can be called by omitting the parameter list (no empty brackets!):<lang Ada>S: String := Ada.Text_IO.Get_Line;</lang>

Ada supports functions with optional parameters:<lang Ada>function F(X: Integer; Y: Integer := 0) return Integer; -- Y is optional

... A : Integer := F(12); B : Integer := F(12, 0); -- the same as A C : Integer := F(12, 1); -- something different</lang>

If the number of parameters of F where fixed to two (by omitting the ":= 0" in the specification), then B and C would be OK but A wouldn't.

Ada does not support functions with a variable number of arguments. But a function argument can be an unconstrained array with as many values as you want:<lang Ada>type Integer_Array is array (Positive range <>) of Integer;

function Sum(A: Integer_Array) return Integer is

  S: Integer := 0;

begin

  for I in A'Range loop
     S := S + A(I);
  end loop;
  return S;

end Sum; ... A := Sum((1,2,3)); -- A = 6 B := Sum((1,2,3,4)); -- B = 10</lang>

One can realize first-class functions by defining an access to a function as a parameter:<lang Ada>function H (Int: Integer;

           Fun: not null access function (X: Integer; Y: Integer)
             return Integer);
          return Integer;

...

X := H(A, F'Access) -- assuming X and A are Integers, and F is a function

                    -- taking two Integers and returning an Integer.</lang>

The caller is free to use either a positional parameters or named parameters, or a mixture of both (with positional parameters first) <lang Ada>Positional := H(A, F'Access);

Named := H(Int => A, Fun => F'Access); Mixed := H(A, Fun=>F'Access); </lang>

ALGOL 68

<lang algol68># A function without arguments: # f();

or #

f;

A function with a fixed number of arguments: #

f(1, x);

ALGOL 68 does not support optional arguments, variable numbers of arguments, or named

arguments. However, some functions may take an argument "0" as an instruction to use the default value (sort of a pseudo-optional argument). #

In "Talk:Call a function" a statement context is explained as

"The function is used as an instruction (with a void context), rather than used within an expression." Based on that, both ALGOL examples above are already in a statement context. For full ALGOL compatibility, though, they should be in the form "VOID (f ());" #

A function's return value being used: #

x := f(y);

There is no distinction between built-in functions and user-defined functions. #

A subroutine is simply a function that returns VOID. #

If the function is declared with argument(s) of mode REF MODE,

then those arguments are being passed by reference. #</lang>

See First-Class Functions for an example of first-class functions in ALGOL 68.

See Partial Function Application for an example of partial function application in ALGOL 68.

AutoHotkey

<lang AHK>; Call a function without arguments: f()

Call a function with a fixed number of arguments

f("string", var, 15.5)

Call a function with optional arguments

f("string", var, 15.5)

Call a function with a variable number of arguments

f("string", var, 15.5)

Call a function with named arguments

   ; AutoHotkey does not have named arguments. However, in v1.1+,
   ; we can pass an object to the function:

f({named: "string", otherName: var, thirdName: 15.5})

Use a function in statement context

f(1), f(2) ; What is statement context?

No first-class functions in AHK

Obtaining the return value of a function

varThatGetsReturnValue := f(1, "a")

Cannot distinguish built-in functions

Subroutines are called with GoSub; functions are called as above.
Subroutines cannot be passed variables

Stating whether arguments are passed by value or by reference
[v1.1.01+]: The IsByRef() function can be used to determine
whether the caller supplied a variable for a given ByRef parameter.
A variable cannot be passed by value to a byRef parameter. Instead, do this

f(tmp := varIdoNotWantChanged)

the function f will receive the value of varIdoNotWantChanged, but any
modifications will be made to the variable tmp.

Partial application is impossible.

</lang>

AWK

The awk interpreter reads the entire script prior to processing, so functions can be called from sections of code appearing before the definition.

<lang awk>BEGIN {

 sayhello()       # Call a function with no parameters in statement context
 b=squareit(3)    # Obtain the return value from a function with a single parameter in first class context

}</lang>

In awk, scalar values are passed by value, but arrays are passed by reference. Note that if a function has no arguments, then empty parentheses are required.

The awk extraction and reporting language does not support the use of named parameters.

Axe

In Axe, up to six arguments are passed as the variables r₁ through r₆. As with all variables in Axe, these exist in the global scope, which makes nested function calls and recursion quite difficult. <lang axe>NOARG() ARGS(1,5,42)</lang>

Since arguments are simply global variables, they are always optional and can be omitted from right to left. <lang axe>OPARG(1,2,3,4,5,6) OPARG(1,2,3) OPARG()</lang>

Somewhat similar to TI-83 BASIC, the last evaluated expression becomes the return value of the function. However, this is distinct from the Ans variable. Return values can be captured just like any other expression. <lang axe>MATHS(2,4)→A Disp GETSTR()</lang>

User-defined functions can be distinguished from language-defined functions by the fact that language-defined function names are composed of atomic tokens (usually with built-in parentheses) whereas user-defined function names are composed of individual characters. Also, because only uppercase letters are available by default in the OS, most user-defined names are all uppercase while language-defined names are mixed case. <lang axe>USER() axeFunc()</lang>

Batch File

Batch files do not have a traditional "function" system like OOP languages, however this is the closest thing to it. The only difference between a block of code and a function is the way method you choose to invoke it. It's also worth noting that all batch files can be called from any other batch file, performing a function. A function should be put somewhere in the code where it will not be parsed unless the script is redirected there.

<lang dos> @echo off

call a function with no arguments

call :myFunction

call a function with arguments

call :myFunction arg1 "arg 2"

initiate a "function".

myFunction

echo arg1 - %1 echo arg2 - %~2 goto :eof </lang>

Bracmat

Calling a function that requires no arguments:

Strictly speaking, all Bracmat functions receive at least one argument. But empty strings are valid expressions, so you can do

<lang bracmat>aFunctionWithoutArguments$</lang> or <lang bracmat>aFunctionWithoutArguments'</lang>

Both function calls pass the right and side of the $ or ' operator. This is in fact still something: an empty string.

The $ operator always evaluates its right hand side before passing it to the function, while the ' does not. Therefore it is slightly faster to use the functionName' variant.

Calling a function with a fixed number of arguments:

Bracmat passes exactly one argument to a function, called arg. The argument can be any Bracmat expression. In patterns, if a function call expression is a pattern component, a second argument sjt is added, the part of the subject that the pattern component is going trying to match.

Calling a function with optional arguments

There is no special syntax for that. It is up to the programmer to define a datastructure with a variable part, e.g., a list.

Calling a function with a variable number of arguments

Same answer.

Calling a function with named arguments

There is no special syntax for that. You could pass a list of (name.value) pairs.

Using a function in statement context

A f...

You can do <lang bracmat>func$!myargument;</lang> The ; marks the end of a Bracmat statement.

Using a function in first-class context within an expression

(Copied from JavaScript:) Bracmat functions are first-class citizens; they can be stored in variables and passed as arguments. Assigning to a variable yourfunc can be done in a few ways. The most common one is

<lang bracmat>(yourfunc=local vars.function body)</lang> If there is already a function myfunc that you want to assign to yourfunc as well, do <lang bracmat>('$myfunc:(=?yourfunc))</lang>

Obtaining the return value of a function

<lang bracmat>myfunc$!myarg:?myresult</lang>

Notice that the returned value can be any evaluated expression.

Distinguishing built-in functions and user-defined functions

You cannot list built-in functions that are implemented directly in C. Nor can such functions be passed as arguments or assigned to variables. There are also a number of built-in functions that are written in Bracmat. They are nothing special and can be deleted or redefined. You can see a list of all currently defined functions that are written in Bracmat with the function call cat'.

Distinguishing subroutines and functions

You can ignore the return value of a function myfunc as follows:

<lang bracmat>myfunc$!myarg&yourfunc$!yourarg</lang>

But notice that if myfunc fails, the above expression returns the value produced by myfunc! To also ignore the success/failure of a function, do

<lang bracmat>`(myfunc$!myarg)&yourfunc$!yourarg</lang>

Stating whether arguments are passed by value or by reference

Values are passed by reference, or by value if the reference counter, which is a very small integer, overflows. Most values are immutable, so for those there is no practical difference between passing by reference or value. The single exception of a mutable value is always passed by reference, and has an enormous reference counter. (The binary operator = introduces a mutable value and can be used for an object oriented style of programming.)

Is partial application possible and how

There is no special syntax for that, but you can write a function that e.g., can take a list with one or with two elements and that returns a function in the first case.

<lang bracmat>( ( plus

 =   a b
   .     !arg:%?a ?b
       & !b:
       & '(.!arg+$a)
     | !a+!b
 )

& out$("1+2, not partial:" plus$(1 2)) & out$("1+2, partial:" (plus$1)$2) );</lang>

Output:

1+2, not partial: 3
1+2,     partial: 3

C

<lang c>/* function with no argument */ f();

/* fix number of arguments */ g(1, 2, 3);

/* Optional arguments: err...

  Feel free to make sense of the following.  I can't. */

int op_arg(); int main() { op_arg(1); op_arg(1, 2); op_arg(1, 2, 3); return 0; } int op_arg(int a, int b) { printf("%d %d %d\n", a, b, (&b)[1]); return a; } /* end of sensible code */

/* Variadic function: how the args list is handled solely depends on the function */ void h(int a, ...) { va_list ap; va_start(ap); ... } /* call it as: (if you feed it something it doesn't expect, don't count on it working) */ h(1, 2, 3, 4, "abcd", (void*)0);

/* named arguments: no such thing */ /* statement context: is that a real phrase? */

/* as a first-class object (i.e. function pointer) */ printf("%p", f); /* that's the f() above */

/* return value */ double a = asin(1);

/* built-in functions: no such thing. Compiler may interally give special treatment

  to bread-and-butter functions such as memcpy(), but that's not a C built-in per se */

/* subroutines: no such thing. You can goto places, but I doubt that counts. */

/* Scalar values are passed by value by default. However, arrays are passed by reference. */ /* Pointers *sort of* work like references, though. */</lang>

C++

/* function with no arguments */ foo(); </lang>

<lang C++> /* passing arguments by value*/ /* function with one argument */ bar(arg1); /* function with multiple arguments */ baz(arg1, arg2); </lang>

<lang C++> /* get return value of a function */ variable = function(args); </lang>

include <iostream>

using namespace std; /* passing arguments by reference */ void f(int &y) /* variable is now passed by reference */ { y++; } int main() { int x = 0; cout<<"x = "<<x<<endl; /* should produce result "x = 0" */ f(x); /* call function f */ cout<<"x = "<<x<<endl; /* should produce result "x = 1" */ } </lang>

COBOL

<lang cobol>CALL "No-Arguments"

> Fixed number of arguments.

CALL "2-Arguments" USING Foo Bar

CALL "Optional-Arguments" USING Foo CALL "Optional-Arguments" USING Foo Bar

> If an optional argument is omitted and replaced with OMITTED, any following
> arguments can still be specified.

CALL "Optional-Arguments" USING Foo OMITTED Bar

> Interestingly, even arguments not marked as optional can be omitted without
> a compiler warning. It is highly unlikely the function will still work,
> however.

CALL "2-Arguments" USING Foo

> COBOL does not support a variable number of arguments, or named arguments.

> Values to return can be put in either one of the arguments or, in OpenCOBOL,
> the RETURN-CODE register.
> A standard function call cannot be done in another statement.

CALL "Some-Func" USING Foo MOVE Return-Code TO Bar

> Intrinsic functions can be used in any place a literal value may go (i.e. in
> statements) and are optionally preceded by FUNCTION.
> Intrinsic functions that do not take arguments may optionally have a pair of
> empty parentheses.
> Intrinsic functions cannot be defined by the user.

MOVE FUNCTION PI TO Bar MOVE FUNCTION MEDIAN(4, 5, 6) TO Bar

> Built-in functions/subroutines typically have prefixes indicating which
> compiler originally incorporated it:
> - C$ - ACUCOBOL-GT
> - CBL_ - Micro Focus
> - CBL_OC_ - OpenCOBOL
> Note: The user could name their functions similarly if they wanted to.

CALL "C$MAKEDIR" USING Foo CALL "CBL_CREATE_DIR" USING Foo CALL "CBL_OC_NANOSLEEP" USING Bar

> Although some built-in functions identified by numbers.

CALL X"F4" USING Foo Bar

> Parameters can be passed in 3 different ways:
> - BY REFERENCE - this is the default way in OpenCOBOL and this clause may
> be omitted. The address of the argument is passed to the function.
> The function is allowed to modify the variable.
> - BY CONTENT - a copy is made and the function is passed the address
> of the copy, which it can then modify. This is recomended when
> passing a literal to a function.
> - BY VALUE - the function is passed the address of the argument (like a
> pointer). This is mostly used to provide compatibility with other
> languages, such as C.

CALL "Modify-Arg" USING BY REFERENCE Foo *> Foo is modified. CALL "Modify-Arg" USING BY CONTENT Foo *> Foo is unchanged. CALL "C-Func" USING BY VALUE Bar

> Partial application is impossible as COBOL does not support first-class
> functions.
> However, as functions are called using a string of their PROGRAM-ID,
> you could pass a 'function' as an argument to another function, or store
> it in a variable, or get it at runtime.

ACCEPT Foo *> Get a PROGRAM-ID from the user. CALL "Use-Func" USING Foo CALL Foo USING Bar</lang>

CoffeeScript

Calling a function that requires no arguments

foo()

Calling a function with a fixed number of arguments

foo 1

Calling a function with optional arguments
(Optional arguments are done using an object with named keys)

foo 1, optionalBar: 1, optionalBaz: 'bax'

Calling a function with a variable number of arguments
for a function `foo` defined as `foo = ( args... ) ->`

foo 1, 2, 3, 4

Calling a function with named arguments
(Named arguments are done using an object with named keys)

foo bar: 1, bax: 'baz'

Using a function in statement context

x = foo 1

Using a function in first-class context within an expression
(For `foo` defined as `foo = ( x ) -> x + 1`

x = [ 1, 2, 3 ].map foo

Obtaining the return value of a function

x = foo 1

Arguments are passed by value, even objects. Objects
are passed as the _value_ of the reference to an object.
Example:

bar = ( person ) ->

   # Since `person` is a reference
   # to the person passed in, we can assign
   # a new value to its `name` key.
   person.name = 'Bob'

   # Since `person` is just the value of
   # the original reference, assigning to it
   # does not modify the original reference.
   person = new Person 'Frank'

Partial application is only possible manually through closures

curry = ( f, fixedArgs... ) ->

   ( args... ) -> f fixedArgs..., args...

Example usage

add = ( x, y ) -> x + y

add2 = curry add, 2

add2 1 #=> 3 </lang>

Common Lisp

Calling a function that requires no arguments

(defun a () "This is the 'A' function") (a)

Calling a function with a fixed number of arguments

(defun b (x y) (list x y)) (b 1 2)

Calling a function with optional arguments

(defun c (&optional x y) (list x y)) (c 1)

Calling a function with a variable number of arguments

(defun d (&rest args) args) (d 1 2 3 4 5 6 7 8)

Calling a function with named arguments

(defun e (&key (x 1) (y 2)) (list x y)) (e :x 10 :y 20)

Using a function in first-class context within an expression

(defun f (func) (funcall func)) (f #'a)

Obtaining the return value of a function

(defvar return-of-a (a))

Is partial application possible and how

(defun curry (function &rest args-1)

 (lambda (&rest args-2)
   (apply function (append args-1 args-2))))

(funcall (curry #'+ 1) 2) </lang>

D

<lang d>import std.traits;

enum isSubroutine(alias F) = is(ReturnType!F == void);

void main() {

   void foo1() {}

   // Calling a function that requires no arguments:
   foo1();
   foo1; // Alternative syntax.

   void foo2(int x, int y) {}

   immutable lambda = function int(int x) => x ^^ 2;

   // Calling a function with a fixed number of arguments:
   foo2(1, 2);
   foo2(1, 2);
   cast(void)lambda(1);

   void foo3(int x, int y=2) {}

   // Calling a function with optional arguments:
   foo3(1);
   foo3(1, 3);

   int sum(int[] arr...) {
       int tot = 0;
       foreach (immutable x; arr)
           tot += x;
       return tot;
   }

   real sum2(Args...)(Args arr) {
       typeof(return) tot = 0;
       foreach (immutable x; arr)
           tot += x;
       return tot;
   }

   // Calling a function with a variable number of arguments:
   assert(sum(1, 2, 3) == 6);
   assert(sum(1, 2, 3, 4) == 10);
   assert(sum2(1, 2.5, 3.5) == 7);

   // Calling a function with named arguments:
   // Various struct or tuple-based tricks can be used for this,
   // but currently D doesn't have named arguments.

   // Using a function in statement context (?):
   if (1)
       foo1;

   // Using a function in first-class context within an expression:
   assert(sum(1) == 1);

   auto foo4() { return 1; }

   // Obtaining the return value of a function:
   immutable x = foo4;

   // Distinguishing built-in functions and user-defined functions:
   // There are no built-in functions, beside the operators, and
   // pseudo-functions like assert().

   int myFynction(int x) { return x; }
   void mySubroutine(int x) {}

   // Distinguishing subroutines and functions:
   // (A subroutine is merely a function that has no explicit
   // return statement and will return void).
   pragma(msg, isSubroutine!mySubroutine); // Prints: true
   pragma(msg, isSubroutine!myFynction);   // Prints: false

   void foo5(int a, in int b, ref int c, out int d, lazy int e, scope int f) {}

   // Stating whether arguments are passed by value, by reference, etc:
   alias STC = ParameterStorageClass;
   alias psct = ParameterStorageClassTuple!foo5;
   static assert(psct.length == 6); // Six parameters.
   static assert(psct[0] == STC.none);
   static assert(psct[1] == STC.none);
   static assert(psct[2] == STC.ref_);
   static assert(psct[3] == STC.out_);
   static assert(psct[4] == STC.lazy_);
   static assert(psct[5] == STC.scope_);
   // There are also inout and auto ref.

   int foo6(int a, int b) { return a + b; }

   // Is partial application possible and how:
   import std.functional;
   alias foo6b = partial!(foo6, 5);
   assert(foo6b(6) == 11);

}</lang>

Output:

true
false

Déjà Vu

<lang dejavu># all functions used are from the standard library

calling a function with no arguments:

random-int

calling a function with a fixed number of arguments:

+ 1 2

calling a function with optional arguments:
optional arguments are not really possible as such
generally differently named functions are used:

sort [ 3 2 1 ] sort-by @len [ "Hello" "World" "Bob" ]

calling a function with a variable number of arguments:
generally with a special terminator value, which one depends
on the function called

concat( 1 2 3 ) [ 1 2 3 ] set{ :foo :bar :spam }

calling a function with named arguments: not possible
using a function in first-class context within an expression

$ @-- @len # $ is "compose", so the function returned is "one less than the length"

obtaining the return value of a function
return values are always pushed on the stack, so you don't need anything special

random-int

discarding the return value of a function

drop random-int

method call:

local :do { :nuthin @pass } do!nuthin !import!fooModule # same as eva!import :fooModule

arguments are passed by object-identity, like in Python and Lua
partial application is not possible, due to the fact that
a function's arity is a property of its behavior and not
of its definition</lang>

Erlang

<lang erlang> no_argument() one_argument( Arg ) optional_arguments( Arg, [{opt1, Opt1}, {another_opt, Another}] ) variable_arguments( [Arg1, Arg2 | Rest] ) names_arguments([{name1, Arg1}, {another_name, Another}] ) % Statement context? % First class context? Result = obtain_result( Arg1 ) % No way to distinguish builtin/user functions % Subroutines? % Arguments are passed by reference, but you can not change them. % Partial application is possible (a function returns a function that has one argument bound) </lang>

F#

<lang fsharp>// No arguments noArgs()

// Fixed number of arguments oneArg x

// Optional arguments // In a normal function: optionalArgs <| Some(5) <| None // In a function taking a tuple: optionalArgsInTuple(Some(5), None) // In a function in a type: foo.optionalArgs 5;; // However, if you want to pass more than one paramter, the arguments must be // passed in a tuple: foo.optionalArgs(5, 6)

// Function with a variable number of arguments variableArgs 5 6 7 // etc...

// Named arguments can only be used in type methods taking a tuple. The // arguments can appear in any order. foo.namedArgs(x = 5, y = 6)

// Using a function in a statement for i = 0 to someFunc() do

   printfn "Something"

// Using a function in a first-class context funcArgs someFunc

// Obtaining a return value let x = someFunc()

// Built-in functions: do functions like (+) or (-) count?

// Parameters are normally passed by value (as shown in the previous examples), // but they can be passed by reference. // Passing by reference: refArgs &mutableVal

// Partial application example let add2 = (+) 2</lang>

Forth

<lang forth>a-function \ requiring no arguments a-function \ with a fixed number of arguents a-function \ having optional arguments a-function \ having a variable number of arguments a-function \ having such named arguments as we have in Forth ' a-function var ! \ using a function in a first-class context (here: storing it in a variable) a-function \ in which we obtain a function's return value

                  \ forth lacks 'statement contenxt'
                  \ forth doesn't distinguish between built-in and user-defined functions
                  \ forth doesn't distinguish between functions and subroutines
                  \ forth doesn't care about by-value or by-reference

\ partial application is achieved by creating functions and manipulating stacks

curried 0 a-function ;

only-want-third-argument 1 2 rot a-function ;

\ Realistic example:

move ( delta-x delta-y -- )

 y +!  x +! ;

down ( n -- ) 0 swap move ;

up ( n -- ) negate down ;

right ( n -- ) 0 move ;

left ( n -- ) negate right ;</lang>

Fortran

<lang Fortran> program main implicit none integer :: a integer :: f, g logical :: lresult interface

 integer function h(a,b,c)
   integer :: a, b
   integer, optional :: c
 end function

end interface write(*,*) 'no arguments: ', f() write(*,*) '-----------------' write(*,*) 'fixed arguments: ', g(5,8,lresult) write(*,*) '-----------------' write(*,*) 'optional arguments: ', h(5,8), h(5,8,4) write(*,*) '-----------------' write(*,*) 'function with variable arguments: Does not apply!' write(*,*) 'An option is to pass arrays of variable lengths.' write(*,*) '-----------------' write(*,*) 'named arguments: ', h(c=4,b=8,a=5) write(*,*) '-----------------' write(*,*) 'function in statement context: Does not apply!' write(*,*) '-----------------' write(*,*) 'Fortran passes memorty location of variables as arguments.' write(*,*) 'So an argument can hold the return value.' write(*,*) 'function result: ', g(5,8,lresult) , ' function successful? ', lresult write(*,*) '-----------------' write(*,*) 'Distinguish between built-in and user-defined functions: Does not apply!' write(*,*) '-----------------' write(*,*) 'Calling a subroutine: ' a = 30 call sub(a) write(*,*) 'Function call: ', f() write(*,*) '-----------------' write(*,*) 'All variables are passed as pointers.' write(*,*) 'Problems can arise if instead of sub(a), one uses sub(10).' write(*,*) '-----------------' end program

!no argument integer function f() f = 10 end function

!fixed number of arguments integer function g(a, b, lresult) integer :: a, b logical :: lresult g = a+b lresult = .TRUE. end function

!optional arguments integer function h(a, b, c) integer :: a, b integer, optional :: c

h = a+b if(present(c)) then

 h = h+10*c

end if end function

!subroutine subroutine sub(a) integer :: a a = a*100 write(*,*) 'Output of subroutine: ', a end subroutine </lang>

 no arguments:           10
 -----------------
 fixed arguments:           13
 -----------------
 optional arguments:           13          53
 -----------------
 function with variable arguments: Does not apply!
 An option is to pass arrays of variable lengths.
 -----------------
 named arguments:           53
 -----------------
 function in statement context: Does not apply!
 -----------------
 Fortran passes memorty location of variables as arguments.
 So an argument can hold the return value.
 function result:           13  function successful?  T
 -----------------
 Distinguish between built-in and user-defined functions: Does not apply!
 -----------------
 Calling a subroutine: 
 Output of subroutine:         3000
 Function call:           10
 -----------------
 All variables are passed as pointers.
 Problems can arise if instead of sub(a), one uses sub(10).
 -----------------

Go

The following examples use functions from the standard packages plus a few dummy local functions:

<lang go>import (

"image" "image/gif" "io/ioutil" "strings" "unicode" )

func f() (int, float64) { return 0, 0 } func g(int, float64) int { return 0 } func h(string, ...int) {}</lang>

Calling with no arguments and calling with a fixed number of arguments:

<lang go> f()

g(1, 2.0) // If f() is defined to return exactly the number and type of // arguments that g() accepts than they can be used in place: g(f()) // But only without other arguments, this won't compile: //h("fail", f()) // But this will: g(g(1, 2.0), 3.0)</lang>

Calling with a variable number of arguments:

This is only possible with functions defined with a trailing optional/variable length argument of a single type (as h above). <lang go> h("ex1")

h("ex2", 1, 2) h("ex3", 1, 2, 3, 4) // such functions can also be called by expanding a slice: list := []int{1,2,3,4} h("ex4", list...) // but again, not mixed with other arguments, this won't compile: //h("fail", 2, list...)</lang>

Optional arguments and named arguments are not supported.

However, it is reasonably common to see a structure used for this. In this example gif.Options is a structure with multiple members which can initialized/assigned by name or omitted (or the whole third argument can just be nil). <lang go> gif.Encode(ioutil.Discard, image.Black, &gif.Options{NumColors: 16})</lang>

Within a statement context.

Assignment statements are shown later. Only functions returning a single value can be used in a single value context: <lang go> if 2*g(1, 3.0)+4 > 0 {}</lang>

In a first-class context:

<lang go> fn := func(r rune) rune {

if unicode.IsSpace(r) { return -1 } return r } strings.Map(fn, "Spaces removed") strings.Map(unicode.ToLower, "Test") strings.Map(func(r rune) rune { return r + 1 }, "shift")</lang>

Obtaining the value:

<lang> a, b := f() // multivalue return

_, c := f() // only some of a multivalue return d := g(a, c) // single return value e, i := g(d, b), g(d, 2) // multiple assignment</lang>

Built-in functions and user defined functions can not be distinguished.

Functions from the standard packages look like any other. The few truly built-in functions are only different in that they have no package specifier like local functions (and they sometimes have extra capabilities). <lang go> list = append(list, a, d, e, i)

i = len(list)</lang>

Go has no subroutines, just functions and methods.
Go arguments are passed by value.

As with C, a pointer can be used to achieve the effect of reference passing. (Like pointers, slice arguments have their contents passed by reference, it's the slice header that is passed by value).

Partial application is not directly supported.

However something similar can be done, see Partial function application#Go

Haskell

<lang haskell> -- Calling a function with a fixed number of arguments multiply x y = x * y multiply 10 20 -- returns 200

-- Calling a function that requires no arguments -- Normally, you use constant instead of function without arguments: twopi = 6.28 -- But you can also pass special value as the first argument indicating function call: twopi () = 6.28 -- definition twopi :: Num a => () -> a -- its type twopi () -- returns 6.28

-- Partial application and auto-currying is built-in. multiply_by_10 = (10 * ) map multiply_by_10 [1, 2, 3] -- [10, 20, 30] multiply_all_by_10 = map multiply_by_10 multiply_all_by_10 [1, 2, 3] -- [10, 20, 30]

-- TODO: -- Calling a function with optional arguments -- Calling a function with a variable number of arguments -- Calling a function with named arguments -- Using a function in statement context -- Using a function in first-class context within an expression -- Obtaining the return value of a function -- Distinguishing built-in functions and user-defined functions -- Distinguishing subroutines and functions -- Stating whether arguments are passed by value or by reference </lang>

Icon and Unicon

Icon and Unicon have generalized procedures and syntax that are used to implement functions, subroutines and generators.

Procedures can return values or not and callers may use the returned values or not.
Procedures in Icon and Unicon are first class values and can be assigned to variables which can then be used to call procedures. This also facilitates some additional calling syntax.
Additionally, co-expressions are supported which allow for co-routine like transfers of control between two or more procedures. There are some differences in syntax for co-expression calls.
There are no differences between calling built-in vs. user defined functions
Named arguments is not natively supported; however, they can be supported using a user defined procedure as shown in Named parameters
Method calling is similar with some extended syntax
Arguments are basically passed by value or reference based on their type. Immutable values like strings, and numbers are passed by value. Mutable data types like structures are essentially references and although these are passed by value the effective behavior is like a call by reference.

For more information see Icon and Unicon Introduction on Rosetta

<lang Icon>procedure main() # demonstrate and describe function calling syntax and semantics

  # normal procedure/function calling

  f()                      # no arguments, also command context
  f(x)                     # fixed number of arguments
  f(x,h,w)                 # variable number of arguments (varargs)
  y := f(x)                # Obtaining the returned value of a function
  
  # procedures as first class values and string invocation

  f!L                      # Alternate calling syntax using a list as args   
  (if \x then f else g)()  # call (f or g)()
  f := write               # assign a procedure
  f("Write is now called") # ... and call it
  "f"()                    # string invocation, procedure
  "-"(1)                   # string invocation, operator

  # Co-expressions
  
  f{e1,e2}                 # parallel evaluation co-expression call  
                           # equivalent to f([create e1, create e2]) 
  expr @ coexp             # transmission of a single value to a coexpression
  [e1,e2]@coexp            # ... of multiple values (list) to a coexpression 
  coexp(e1,e2)             # ... same as above but only in Unicon

  # Other
  
  f("x:=",1,"y:=",2)       # named parameters (user defined)

end</lang>

J

A function in J is typically represented by a verb. Under the right circumstances other syntactic entities (nouns, adverbs, conjunctions) can represent functions, but let's focus on the typical case.

A verb, in J, typically supports two syntactic variants:

<lang j> verb noun

  noun verb noun</lang>

And a noun, in J, is an array.

An argument list can be represented by an array. Thus, when dealing with multiple arguments, a typical form is:

<lang j> function argumentList</lang>

Here, function is a verb and argumentList is a noun.

For example:

Here sum is a verb and (1,2,3) is a noun.

Thus:

A function that requires no arguments can be simulated by calling a function with empty argument list: <lang j>f</lang> Note that an empty list of characters is not the only constant in the language which is an empty list. That said, most operations in the language do not care what type of data is not present, in an array which contains nothing.

A function with a fixed number of arguments gets special treatment in J when the fixed number is 1 or 2. <lang j>f 'one argument'</lang>and <lang j>'this example has two arguments' f 'the other argument'</lang> Alternatively, the function can be written such that an argument list is an error when it's the wrong length.

A function with a variable number of arguments (varargs): See above.

If argument types conflict they will need to be put in boxes and the function will have to take its arguments out of the boxes. Here's an unboxed example with five arguments: <lang j> f 1,2,3,4,5</lang> and here's a boxed example with five arguments: <lang j>f (<1),(<2),(<3),(<4),(<5) </lang> Note that the last set of parenthesis is unnecessary <lang j>f (<1),(<2),(<3),(<4),<5</lang> Note also that J offers some syntactic sugar for this kind of list <lang j>f 1; 2; 3; 4; <5</lang>. Note also that if the last argument in a semicolon list is not boxed there is no need to explicitly box it, since that is unambiguous (it must be boxed so that it conforms with the other members of the list). <lang j>f 1; 2; 3; 4; 5</lang>

A function with named arguments can be accomplished by calling a function with the names of the arguments. <lang j>f 'george';'tom';'howard'</lang> Other interpretations of this concept are also possible. For example, the right argument for a verb might be a list of argument names and the left argument might be a corresponding list of argument values: <lang j>1 2 3 f 'george';'tom';'howard'</lang> Or, for example a function which requires an object could be thought of as a function with named arguments since an object's members have names:<lang j> obj=: conew'blank'

  george__obj=: 1
  tom__obj=: 2
  howard__obj=: 3
  f obj
  coerase obj</lang>  Name/value pairs can also be used for this purpose and can be implemented in various ways, including passing names followed by values <lang j>f 'george';1;'tom';2;'howard';3</lang> and passing a structure of pairs <lang j>f ('george';1),('tom';2),:(howard';3)</lang>  Or, for example, the pairs could be individually boxed:  <lang j>f ('george';1);('tom';2);<howard';3</lang>

Using a function in command context is no different from using a function in any other context, in J. Using a function in first class context within an expression is no different from using a function in any other context, in J.

Obtaining the return value of a function is no different from using a function in j. For example, here we add 1 to the result of a function: <lang j>1 + f 2</lang>

The only differences that apply to calling builtin functions rather than user defined functions is spelling of the function names.

There are no differences between calling subroutines and functions because J defines neither subroutines nor functions. Instead, J defines verbs, adverbs, and conjunctions which for the purpose of this task are treated as functions. (All of the above examples used verbs. J's adverbs and conjunctions have stronger valence than its verbs.)

Java

Java does not have functions, but Java classes have "methods" which are equivalent.

Calling a function that requires no arguments

<lang java>myMethod()</lang> We didn't specify an object (or a class) as the location of the method, so this.myMethod() is assumed. This applies to all the following examples.

Calling a function with a fixed number of arguments

<lang java>myMethod(97, 3.14)</lang>

Calling a function with optional arguments

This is possible if the method name is overloaded with different argument lists. For example: <lang java>int myMethod(int a, double b){

   // return result of doing sums with a and b

}

int myMethod(int a){

   return f(a, 1.414);

}</lang>

The compiler figures out which method to call based on the types of the arguments, so in this case the second argument appears to be optional. If you omit it, the value 1.414 is used. <lang java>System.out.println( myMethod( 97, 3.14 ) ); System.out.println( myMethod( 97 ) );</lang>

Calling a function with a variable number of arguments

This is possible if the method is defined with varargs syntax. For example: <lang java>void printAll(String... strings){

   for ( String s : strings )
       System.out.println( s );

}</lang>

The type of strings is actually a string array, but the caller just passes strings: <lang java>printAll( "Freeman" ); printAll( "Freeman", "Hardy", "Willis" );</lang>

To avoid ambiguity, only the last argument to a function can have varargs.

Calling a function with named arguments

Not directly possible, but you could simulate this (somewhat verbosely): <lang java>int myMethod( Map<String,Object> params ){

   return
      ((Integer)params.get("x")).intValue()
      + ((Integer)params.get("y")).intValue();

}</lang>

Called like this: <lang java>System.out.println( myMethod(new HashMap<String,Object>()Template:Put("x",27);put("y",52);) );</lang>

Yuk.

Using a function in statement context

If this means "use a function where a statement is expected", see all the other examples

Using a function in first-class context within an expression

Not possible - must be wrapped in a class

Obtaining the return value of a function

<lang java>int i = myMethod(x);</lang>

Distinguishing built-in functions and user-defined functions

No distinction - all methods belong to classes, and there is no real distinction between built-in and user-defined classes.

Distinguishing subroutines and functions

If the return type is void, you might consider a method as a subroutine rather than a function.

Stating whether arguments are passed by value or by reference

All arguments are passed by value, but since object variables contain a reference to an object (not the object itself), objects appear to be passed by reference. For example: <lang java>myMethod(List<String> list){

   // If I change the contents of the list here, the caller will see the change

}</lang>

Is partial application possible and how

Don't know

JavaScript

The arguments to a JavaScript function are stored in a special array-like object which does not enforce arity in any way; a function declared to take n arguments may be called with none‒and vice versa‒without raising an error.

<lang JavaScript>var foo = function() { return arguments.length }; foo() // 0 foo(1, 2, 3) // 3</lang>

Neither optional (see above) nor named arguments are supported, though the latter (and the inverse of the former) may be simulated with the use of a helper object to be queried for the existence and/or values of relevant keys. Seriously, what is "statement context"?

JavaScript functions are first-class citizens; they can be stored in variables (see above) and passed as arguments. <lang JavaScript>var squares = [1, 2, 3].map(function (n) { return n * n }); // [1, 4, 9]</lang>

Naturally, they can also be returned, thus partial application is supported. <lang JavaScript> var make_adder = function(m) {

   return function(n) { return m + n }

}; var add42 = make_adder(42); add42(10) // 52</lang>

Calling a user-defined function's toString() method returns its source verbatim; that the implementation is elided for built-ins provides a mechanism for distinguishing between the two.

<lang JavaScript>foo.toString() "function () { return arguments.length }" alert.toString() "function alert() { [native code] }"</lang>

Arguments are passed by value, but the members of collections are essentially passed by reference and thus propagate modification. <lang JavaScript>var mutate = function(victim) {

   victim[0] = null;
   victim = 42;

}; var foo = [1, 2, 3]; mutate(foo) // foo is now [null, 2, 3], not 42</lang>

jq

jq functions are pure functions that are somewhat unusual in two respects:

They are like commands in modern operating systems, in that they are parameterized filters that can accept input from the previous command and provide output to the next command if there is one in the pipeline of commands.

Functions can not only process a stream of inputs, one at a time, but each argument can also accept a stream of inputs. The outputs are then a Cartesian product of the various inputs.

In this section, we use the notation fn/N to refer to a function named "fn" with N formal parameters.

Calling a function that requires no arguments

0-arity jq functions are invoked simply by specifying their name.

Example: .

(Yes, "." is a 0-arity jq function.)

Calling a function with a fixed number of arguments

The conventional syntax is used except that ";" is the parameter separator. "," is used to construct a stream of values.

Example: range(0;100;2)

Calling a function with optional arguments

Recent versions of jq allow one to define functions with the same name but different arities, and therefore if both fn/1 and fn/2 are defined, we may say that fn requires one parameter but accepts 2. In all cases, the syntax for function invocation is the same.

Example: range(0; 10) and range(0; 10; 2)

Since jq functions can accept JSON arrays and objects, there are other ways to simulate optional arguments.

Calling a function with a variable number of arguments

See above.

Calling a function with named arguments

This is not directly supported but can be simulated by defining the function to accept JSON objects. For example, if fn were such a function, we might invoke fn like so: fn( {"required": 1, "optional": 2} ).

Using a function in statement context

The assignment to a local variable (e.g. (2*2) as $two) is similar to a statement context in that the expression as a whole does nothing to the flow of values from its input to its output.

Using a function in first-class context within an expression

jq functions cannot be assigned to variables but are otherwise "first-class" in that the composition of functions can be passed as arguments to other functions. No special syntax is required.

Example: 2 | recurse(. * .) # generate the sequence 2, 4, 16, 256, ...Obtaining the return value of a function The value (or stream of values) returned by a function is (or are) automatically available to the next function in the pipeline (e.g. sin | cos); the returned value(s) can also be assigned to a local variable (e.g. sin as $v). Distinguishing built-in functions and user-defined functions Currently there is no such distinction, but user-defined functions always have the conventional form, whereas some built-in functions have special syntax. Distinguishing subroutines and functions A jq function can be written so as never to return anything (either because it returns without generating a value or because it generates an infinite stream of values), but there is no distinctive marker associated with such functions. Stating whether arguments are passed by value or by reference Arguments are in effect passed by value. Is partial application possible and how See Currying#jq.LFECalling a function that requires no arguments: In some module, define the following: <lang lisp> (defun my-func()(: io format '"I get called with NOTHING!~n"))</lang> Then you use it like so (depending upon how you import it): <lang lisp> > (my-func) I get called with NOTHING! ok </lang> Calling a function with a fixed number of arguments: In some module, define the following: <lang lisp> (defun my-func(a b)(: io format '"I got called with ~p and ~p~n" (list a b)))</lang> Then you use it like so: <lang lisp> > (my-func '"bread" '"cheese") I got called with "bread" and "cheese" ok </lang> Calling a function with optional arguments or calling a function with a variable number of arguments:Arguments are fixed in LFE/Erlang functions. One can have a dictionary, record, or list be the function argument, and use that to achieve something like variable/optional (and named) arguments. One can define multiple functions so that it appears that one is calling a function with optional or a variable number of arguments:<lang lisp> (defmodule args(export all))(defun my-func ()(my-func () () ()))(defun my-func (a)(my-func a () ()))(defun my-func (a b)(my-func a b ()))(defun my-func (a b c)(: io format '"~p ~p ~p~n" (list a b c)))</lang> Here is some example usage: <lang lisp> > (slurp '"args.lfe")(ok args)> (my-func) [] [] [] ok > (my-func '"apple") "apple" [] [] ok > (my-func '"apple" '"banana") "apple" "banana" [] ok > (my-func '"apple" '"banana" '"cranberry") "apple" "banana" "cranberry" ok > (my-func '"apple" '"banana" '"cranberry" '"bad arg") exception error: #(unbound_func #(my-func 4)) </lang> Calling a function with named arguments:LFE/Erlang doesn't support named arguments, per se. However, by using atoms in function argument patterns (a fairly common pattern), one can achieve similar effects. One may also use records or dicts as arguments to achieve similar effects.Using a function in statement context: <lang lisp> ...(cond ((== count limit) (hit-limit-func arg-1 arg-2)) ((/= count limit) (keep-going-func count))) ...</lang> Using a function in first-class context within an expression: From the LFE REPL: <lang lisp> > (>= 0.5 (: math sin 0.5)) true </lang> Obtaining the return value of a function: There are many, many ways to assign function outputs to variables in LFE. One fairly standard way is with the (let ...) form: <lang lisp> (let ((x (: math sin 0.5)))...)</lang> Distinguishing built-in functions and user-defined functions:There is no distinction made in LFE/Erlang between functions that are built-in and those that are not. "Built-in" for LFE/Erlang usually can be figured out: if a function has the module name erlang, e.g., (: erlang list_to_integer ... )</cod>, then it's built-in. Most of the functions that come with LFE/Erlang are not even in the erlang module, but exist in other modules (e.g., io, math, etc.) and in OTP. One uses user/third-party modules in exactly the same way as one uses built-ins and modules that come with the Erlang distribution.Distinguishing subroutines and functions:One commonly made distinction between functions and subroutines is that functions return a value (or reference, etc.) and subroutines do not. By this definition, LFE/Erlang does not support the concept of a subroutine; all functions return something.Stating whether arguments are passed by value or by reference:Arguments and returns values are passed by reference in LFE/Erlang.Is partial application possible?Not explicitly. However, one can use lambdas to achieve the same effect.Liberty BASIC<lang lb> 'Call a function - Liberty BASIC 'First, function result could not be discarded ' that is, you cannot do "f(x)" as a separate statement 'Calling a function that requires no arguments res = f() 'brackets required 'Calling a function with a fixed number of arguments res = g(x) res = h(x,y) 'Calling a function with optional arguments'impossible for user-defined functions 'Some build-in functions ex. INSTR and MID$ could be called with last argument omitted'Calling a function with a variable number of arguments'impossible'Calling a function with named arguments'impossible'Using a function in statement context'impossible (see starting notice)'Using a function in first-class context within an expression'impossible'Obtaining the return value of a function res = g(x) 'Distinguishing built-in functions and user-defined functions'I would say impossible. Though built-in functions could be EVAL'ed, 'while user-defined would not be called (tries address array instead). 'Still cannot distinguish user-defined function from array.'Distinguishing subroutines and functions'then defined, subroutines and functions defined with words 'SUB and FUNCTION (case incensitive) 'Then used, function used as expression (with return value), res = g(x) 'while subroutines called with special keyword CALL and without brackets call test x, y'Stating whether arguments are passed by value or by reference'Variables passed as arguments into functions and subs are passed "by value" by default 'parameters could be passed "by reference" if formal parameter in sub/function definition uses the "byref" specifier 'Then calling a function, you can prevent pass by reference by changing variable to expression ' like x+0, x$+"" or just (x), (x$)'Is partial application possible and how'impossible</lang>Lua<lang lua>-- Lua functions accept any number of arguments; missing arguments are nil-padded, extras are dropped. function fixed (a, b, c) print(a, b, c) end fixed() --> nil nil nil fixed(1, 2, 3, 4, 5) --> 1 2 3 -- True vararg functions include a trailing ... parameter, which captures all additional arguments as a group of values. function vararg (...) print(...) end vararg(1, 2, 3, 4, 5) -- 1 2 3 4 5 -- Lua also allows dropping the parentheses if table or string literals are used as the sole argument print "some string" print { foo = "bar" } -- also serves as a form of named arguments -- First-class functions in expression context print(("this is backwards uppercase"):gsub("%w+", function (s) return s:upper():reverse() end)) -- Functions can return multiple values (including none), which can be counted via select() local iter, obj, start = ipairs { 1, 2, 3 } print(select("#", (function () end)())) --> 0 print(select("#", unpack { 1, 2, 3, 4 })) --> 4 -- Partial application function prefix (pre)return function (suf) return pre .. suf endend local prefixed = prefix "foo" print(prefixed "bar", prefixed "baz", prefixed "quux") -- nil, booleans, and numbers are always passed by value. Everything else is always passed by reference. -- There is no separate notion of subroutines -- Built-in functions are not easily distinguishable from user-defined functions </lang>Luck<lang luck>/* Calling a function that requires no arguments */ f();; /* Calling a function with a fixed number of arguments */ f(1,2);; /* Calling a function with optional argumentsNote: defining the function is cumbersome but will get easier in future versions. */f(1,2,new {default with x=3, y=4});; /* Calling a function with a variable number of arguments */ printf("%d %d %d %d":char*,2,3,4,5);; /* Calling a function with named argumentsNote: may get syntax sugar in future versions */f(1,2,new {default with x=3, y=4});; /* Using a function in statement context (what?) */ f();f();f();; /* Using a function in first-class context within an expression */ [1,2,3].map(string);; /* Obtaining the return value of a function */ let x:int = f();; /* Distinguishing built-in functions and user-defined functions */ /* Builtin function i.e. custom calling convention: */ (@ binop "==" l r);; /* User defined function i.e. normal function */ f(l)(r);; /* Distinguishing subroutines and functions: both are supported, but compiler is not aware of difference */ sub();; fun();; /* Stating whether arguments are passed by value or by reference */ f(value);; /* by value */ f(&value);; /* by pointer reference */ f(ref(value));; /* by managed reference */ /* Is partial application possible and how */ tasty_curry(a)(b)(c)(d)(e)(f)(g)(h)(i)(j)(k)(l)(m)(n)(o)(p)(q)(r)(s)(t)(u)(v)(w)(x)(y)(z);;</lang>MapleCalling a function with no arguments:<lang Maple> f()</lang> Calling a function with a fixed number of arguments:<lang Maple>f(1,sin(x), g -> int(g(t),t=0..1)</lang> Calling a function with optional arguments: <lang Maple>f(1, sin(x), g -> int(g(t),t=0..1)</lang> Calling a function with a variable number of arguments: <lang Maple>f(1, sin(x), g -> int(g(t),t=0..1)</lang> Calling a function with named arguments:<lang Maple>f(a,b,method = foo)</lang> Calling a function in a statements context:<lang Maple>f(a); f(b);</lang> Using a function in first-class context within an expression:<lang Maple>f(a) + g(b)</lang> Obtaining the return value of a function:<lang Maple> x := f(1)</lang> Distinguishing built-in functions and user-defined functions: <lang Maple>> type( op, 'builtin' );true</lang> Distinguishing subroutines and functions: There is no distinction. Stating whether arguments are passed by value or by reference: All values are passed by value. However, if an argument is a name, then it can be assigned to and, if a value is mutable (such as an array), then it can be modified by the function. This is implemented by function definition; there is no distinction when calling the function. Partial application is supported by the curry and rcurry commands.Mathematica / Wolfram LanguageCalling a function that requires no arguments: <lang Mathematica>f[]</lang> Calling a function with a fixed number of arguments: <lang Mathematica>f[1,2]</lang> Calling a function with optional arguments: <lang Mathematica>f[1,Option1->True]</lang> Calling a function with a variable number of arguments: <lang Mathematica>f[1,Option1->True] f[1,Option1->True,Option2->False]</lang> Calling a function with named arguments: <lang Mathematica>f[Option1->True,Option2->False]</lang> Using a function in statement context: <lang Mathematica>f[1,2];f[2,3]</lang> Using a function in first-class context within an expression: <lang Mathematica>(#^2)&[3];</lang> The return value of a function can be formally extracted using Return[] Built-in functions names by convention start with a capital letter. No formal distinction between subroutines and functions. Arguments can be passed by value or by reference.MATLAB / Octave<lang Matlab> % Calling a function that requires no arguments function a=foo(); a=4; end; x = foo(); % Calling a function with a fixed number of arguments function foo(a,b,c); %% function definition; end; foo(x,y,z); % Calling a function with optional arguments function foo(a,b,c);if nargin<2, b=0; end; if nargin<3, c=0; end; %% function definition; end; foo(x,y); % Calling a function with a variable number of arguments function foo(varargin);for k=1:length(varargin)arg{k} = varargin{k}; end; foo(x,y); % Calling a function with named arguments%% does not apply % Using a function in statement context%% does not apply % Using a function in first-class context within an expression % Obtaining the return value of a function function [a,b]=foo(); a=4; b='result string'; end; [x,y] = foo(); % Distinguishing built-in functions and user-defined functionsfun = 'foo'; if (exist(fun,'builtin'))printf('function %s is a builtin\n'); elseif (exist(fun,'file')) printf('function %s is user-defined\n'); elseif (exist(fun,'var')) printf('function %s is a variable\n'); else printf('%s is not a function or variable.\n'); end % Distinguishing subroutines and functions % there are only scripts and functions, any function declaration starts with the keyword function, otherwise it is a script that runs in the workspace % Stating whether arguments are passed by value or by reference % arguments are passed by value, however Matlab has delayed evaluation, such that a copy of large data structures are done only when an element is written to.</lang>Nemerle<lang Nemerle>// no arguments f() // fixed arguments def f(a, b) { ... } // as an aside, functions defined with 'def' use type inference for parameters and return types f(1, 'a') // optional arguments def f(a, b = 0) { ... } f("hello") f("goodbye", 2) f("hey", b = 2) // using the name makes more sense if there's more than one optional argument, obviously // variable number of arguments def f(params args) { ... } def g(a, b, params rest) { ... } f(1, 2, 3) // arguments should all have the same type or may be coerced to a supertype g(1.0, 2, "a", "hello") // named arguments f(a = 'a', b = 0) f(b = 0, a = 'a') f('a', b = 0) // if mixing named and unnamed args, unnamed must be first and in correct order // statement context if (f(foo) == 42)WriteLine($"$foo is the meaning to life, the universe and everything.")else WriteLine($"$foo is meaningless.") // first class function in an expression def a = numList.FoldLeft(f) // obtaining return value def a = f(3) // distinguishing built-in from user functions // N/A? // distinguishing subroutines from functions // N/A // stating whether passed by value or by reference // .NET distinguishes between value types and reference types; if a reference type is passed by reference (using ref or out), // the reference is passed by reference, which would allow a method to modify the object to which the reference refers def f(a, ref b) { ... } mutable someVar = "hey there" // doesn't make sense to pass immutable value by ref f(2, ref someVar) def g(a, out b) { ... } mutable someOtherVar // if passed by ref using 'out', the variable needn't be initialized g(2, out someOtherVar) // partial application def f(a, b) { ... } def g = f(2, _) def h = f(_, 2) def a = g(3) // equivalent to: def a = f(2, 3) def b = h(3) // equivalent to: def b = f(3, 2)</lang>NimTranslated from Python, when possible: <lang nim>proc no_args() =discardcallno_args() proc fixed_args(x, y) =echo x echo ycallsfixed_args(1, 2) # x=1, y=2 fixed_args 1, 2 # same call 1.fixed_args(2) # same call proc opt_args(x=1.0) =echo xcallsopt_args() # 1 opt_args(3.141) # 3.141 proc var_args(v: varargs[string, `$`]) =for x in v: echo xcallsvar_args(1, 2, 3) # (1, 2, 3) var_args(1, (2,3)) # (1, (2, 3)) var_args() # ()Named argumentsfixed_args(y=2, x=1) # x=1, y=2As a statementif true:no_args()proc return_something(x): int =x + 1var a = return_something(2)First-class within an expressionlet x = return_something(19) + 10 let y = 19.return_something() + 10 let z = 19.return_something + 10</lang>OCamlCalling a function that requires no arguments:<lang ocaml>f ()</lang> (In fact it is impossible to call a function without arguments, when there are no particular arguments we provide the type unit which is a type that has only one possible value. This type is mainly made for this use.)Calling a function with a fixed number of arguments:<lang ocaml>f 1 2 3</lang>Calling a function with optional arguments:For a function that has this signature: <lang ocaml>val f : ?a:int -> int -> unit</lang> here is how to call it with or without the first argument omited: <lang ocaml>f 10 f ~a:6 10</lang> Due to partial application, an optional argument always has to be followed by a non-optional argument. If the function needs no additional arguments then we use the type unit: <lang ocaml>g () g ~b:1.0 ()</lang>Calling a function with a variable number of arguments:This is not possible. The strong OCaml type system does not allow this. The OCaml programmer will instead provide the variable number of arguments in a list, an array, an enumeration, a set or any structure of this kind. (But if we really need this for a good reason, it is still possible to use a hack, like it has been done for the function Printf.printf.)Calling a function with named arguments:Named arguments are called labels. <lang ocaml>f ~arg:3</lang> If a variable has the same name than the label we can use this simpler syntax: <lang ocaml>let arg = 3 in f ~arg</lang>Using a function in statement context:<lang ocaml>(* TODO *)</lang>Using a function in first-class context within an expression:functions in OCaml are first-class citizen.Obtaining the return value of a function:<lang ocaml>let ret = f () let a, b, c = f () (* if there are several returned values given as a tuple *) let _ = f () (* if we want to ignore the returned value *) let v, _ = f () (* if we want to ignore one of the returned value *)</lang>Distinguishing built-in functions and user-defined functions:There is no difference.Distinguishing subroutines and functions:OCaml only provides functions.Stating whether arguments are passed by value or by reference:OCaml arguments are always passed by reference. OCaml is an impure functional language, for immutable variables there is no difference if the argument is passed by value or by reference, but for mutable variables the programmer should know that a function is able to modify it.How to use partial application:Just apply less arguments than the total number of arguments. With partial application, the arguments are applied in the same order than they are defined in the signature of the function, except if there are labeled arguments, then it is possible to use these labels to partially apply the arguments in any order.OforthOforth provides functions (global prodecure, without receiver) and methods (need a receiver). Oforth uses RPN notation. Arguments must be on the stack before calling a function or method. So, the same syntax is use for calling a function with or without prameters. If f is a function and c b a ares objects : <lang Oforth>c b a f</lang> will push c then b then a on the stack then call f. Calling f does not describe if f will use 1, 2 or 3 arguments (or none). Oforth adds a notation to describe parameters used by a function. It is only a way to add information about which parameters will be used by f : <lang Oforth>f(a, b, c)</lang> Intepreter will replace this second syntax by the first one. It is only "sugar"... <lang Oforth>c b a f c b f(a) c f(a, b) f(a, b, c)</lang> are the same call to function f and the interpreter will translate all of them into the first one. Which parameters are really used by f will depend on f implementation. Methods need a receiver (the object on which the method will apply and the object that will pushed on th stack when self is used into the method body). The receiver must be on the top of the stack before calling the method. If c b a and r are objects and m a method : <lang Oforth>c b a r m</lang> will call m with r as its receiver. It is also possible to use the same notation used by functions : <lang Oforth>r m(a, b, c)</lang>PARI/GPCalling a function is done in GP by writing the name of the function and the arguments, if any, in parentheses. As of version 2.5.0, function calls must use parentheses; some earlier versions allowed functions with an arity of 0 to be called without parentheses. However built-in constants (which are implicit functions of the current precision) can still be called without parentheses. Optional arguments can be skipped, leaving commas in place. Trailing commas can be dropped. Functions can be used when statements would be expected without change. <lang parigp>f(); \\ zero arguments sin(Pi/2); \\ fixed number of arguments vecsort([5,6]) != vecsort([5,6],,4) \\ optional arguments Str("gg", 1, "hh") \\ variable number of arguments call(Str, ["gg", 1, "hh"]) \\ variable number of arguments in a vector (x->x^2)(3); \\ first-class x = sin(0); \\ get function value</lang> Built-in functions are like user-defined functions in current versions. In older versions built-in functions cannot be passed as closures. Most arguments are passed by reference. Some built-in functions accept arguments (e.g., flags) that are not GENs; these are passed by value or reference depending on their C type. See the User's Guide to the PARI Library section 5.7.3, "Parser Codes".PerlThe most common syntax; simply calls the function foo on the argument(s) provided. <lang perl>foo(); # Call foo on the null list &foo(); # Ditto foo($arg1, $arg2); # Call foo on $arg1 and $arg2 &foo($arg1, $arg2); # Ditto; ignores prototypes</lang> Call foo() as a bareword. Only works after the function has been declared, which can be done normally or with the use subs pragma. <lang perl>foo;</lang> Call foo() with the current values of @_<lang perl>&foo;</lang> Call foo() with the current values of @_, discarding the previous stack frame. Not your grandfather's (harmful) goto, although the keyword can do both.<lang perl>goto &foo;</lang> For subroutines stored in references (anonymous subroutines).<lang perl>&$fooref('foo', 'bar'); &{$fooref}('foo', 'bar'); $fooref->('foo', 'bar');</lang>Perl 6Fundamentally, nearly everything you do in Perl 6 is a function call if you look hard enough. At the lowest level, a function call merely requires a reference to any kind of invokable object, and a call to its postcircumfix:<( )> method. However, there are various forms of sugar and indirection that you can use to express these function calls differently. In particular, operators are all just sugar for function calls. Calling a function that requires no arguments: <lang perl6>foo # as list operator foo() # as function foo.() # as function, explicit postfix form $ref() # as object invocation $ref.() # as object invocation, explicit postfix &foo() # as object invocation &foo.() # as object invocation, explicit postfix($name)() # as symbolic ref</lang>Calling a function with exactly one argument: <lang perl6>foo 1 # as list operator foo(1) # as named function foo.(1) # as named function, explicit postfix $ref(1) # as object invocation (must be hard ref) $ref.(1) # as object invocation, explicit postfix 1.$foo # as pseudo-method meaning $foo(1) (hard ref only) 1.$foo() # as pseudo-method meaning $foo(1) (hard ref only) 1.&foo # as pseudo-method meaning &foo(1) (is hard foo) 1.&foo() # as pseudo-method meaning &foo(1) (is hard foo) 1.foo # as method via dispatcher 1.foo() # as method via dispatcher 1."$name"() # as method via dispatcher, symbolic +1 # as operator to prefix:<+> function</lang> Method calls are included here because they do eventually dispatch to a true function via a dispatcher. However, the dispatcher in question is not going to dispatch to the same set of functions that a function call of that name would invoke. That's why there's a dispatcher, after all. Methods are declared with a different keyword, method, in Perl 6, but all that does is install the actual function into a metaclass. Once it's there, it's merely a function that expects its first argument to be the invocant object. Hence we feel justified in including method call syntax as a form of indirect function call. Operators like + also go through a dispatcher, but in this case it is multiply dispatched to all lexically scoped candidates for the function. Hence the candidate list is bound early, and the function itself can be bound early if the type is known. Perl 6 maintains a clear distinction between early-bound linguistic constructs that force Perlish semantics, and late-bound OO dispatch that puts the objects and/or classes in charge of semantics. (In any case, &foo, though being a hard ref to the function named "foo", may actually be a ref to a dispatcher to a list of candidates that, when called, makes all the candidates behave as a single unit.) Calling a function with exactly two arguments: <lang perl6>foo 1,2 # as list operator foo(1,2) # as named function foo.(1,2) # as named function, explicit postfix $ref(1,2) # as object invocation (must be hard ref) $ref.(1,2) # as object invocation, explicit postfix 1.$foo: 2 # as pseudo-method meaning $foo(1,2) (hard ref only) 1.$foo(2) # as pseudo-method meaning $foo(1,2) (hard ref only) 1.&foo: 2 # as pseudo-method meaning &foo(1,2) (is hard foo) 1.&foo(2) # as pseudo-method meaning &foo(1,2) (is hard foo) 1.foo: 2 # as method via dispatcher 1.foo(2) # as method via dispatcher 1."$name"(2) # as method via dispatcher, symbolic 1 + 2 # as operator to infix:<+> function</lang> Optional arguments don't look any different from normal arguments. The optionality is all on the binding end. Calling a function with a variable number of arguments (varargs): <lang perl6>foo @args # as list operator foo(@args) # as named function foo.(@args) # as named function, explicit postfix $ref(@args) # as object invocation (must be hard ref) $ref.(@args) # as object invocation, explicit postfix 1.$foo: @args # as pseudo-method meaning $foo(1,@args) (hard ref) 1.$foo(@args) # as pseudo-method meaning $foo(1,@args) (hard ref) 1.&foo: @args # as pseudo-method meaning &foo(1,@args) 1.&foo(@args) # as pseudo-method meaning &foo(1,@args) 1.foo: @args # as method via dispatcher 1.foo(@args) # as method via dispatcher 1."$name"(@args) # as method via dispatcher, symbolic @args X @blargs # as list infix operator to infix:<X></lang> Note: whether a function may actually be called with a variable number of arguments depends entirely on whether a signature accepts a list at that position in the argument list, but describing that is not the purpose of this task. Suffice to say that we assume here that the foo function is declared with a signature of the form (*@params). The calls above might be interpreted as having a single array argument if the signature indicates a normal parameter instead of a variadic one. What you cannot do in Perl 6 (unlike Perl 5) is pass an array as several fixed arguments. By default it must either represent a single argument, or be part of a variadic list. You can force the extra level of argument list interpolation using a prefix | however: <lang perl6>my @args = 1,2,3; foo(|@args); # equivalent to foo(1,2,3)</lang> Calling a function with named arguments: <lang perl6>foo :a, :b(4), :!c, d => "stuff" foo(:a, :b(4), :!c, d => "stuff")</lang> ...and so on. Operators may also be called with named arguments, but only colon adverbials are allowed: <lang perl6>1 + 1 :a :b(4) :!c :d("stuff") # calls infix:<+>(1,1,:a, :b(4), :!c, d => "stuff")</lang> Using a function in statement context: <lang perl6>foo(); bar(); baz(); # evaluate for side effects</lang> Using a function in first class context within an expression: <lang perl6>1 / find-a-func(1,2,3)(4,5,6) ** 2;</lang> Obtaining the return value of a function: <lang perl6>my $result = somefunc(1,2,3) + 2;</lang> There is no difference between calling builtins and user-defined functions and operators (or even control stuctures). This was a major design goal of Perl 6, and apart from a very few low-level primitives, all of Perl 6 can be written in Perl 6. There is no difference between calling subroutines and functions in Perl 6, other than that calling a function in void context that has no side effects is likely to get you a "Useless use of..." warning. And, of course, the fact that pure functions can participate in more optimizations such as constant folding. By default, arguments are passed readonly, which allows the implementation to decide whether pass-by-reference or pass-by-value is more efficient on a case-by-case basis. Explicit lvalue, reference, or copy semantics may be requested on a parameter-by-parameter basis, and the entire argument list may be processed raw if that level of control is needed.PicoLispWhen calling a funcion in PicoLisp directly (does this mean "in a statement context"?), it is always surrounded by parentheses, with or without arguments, and for any kind of arguments (evaluated or not): <lang PicoLisp>(foo) (bar 1 'arg 2 'mumble)</lang> When a function is used in a "first class context" (e.g. passed to another function), then it is not yet called. It is simply used. Technically, a function can be either a number (a built-in function) or a list (a Lisp-level function) in PicoLisp): <lang PicoLisp>(mapc println Lst) # The value of 'printlin' is a number (apply '((A B C) (foo (+ A (* B C)))) (3 5 7)) # A list is passed</lang> Any argument to a function may be evaluated or not, depending on the function. For example, 'setq' evaluates every second argument <lang PicoLisp>(setq A (+ 3 4) B (* 3 4))</lang> i.e. the first argument 'A' is not evaluated, the second evaluates to 7, 'B' is not evaluated, then the fourth evaluates to 12.PythonUnder the hood all Python function/method parameters are named. All arguments can be passed as name=value pairs or as a dictionary containing such pairs using the myfunc(**key_args) (apply over dictionary) syntax). One can also "apply" a function over a sequence of arguments using the syntax: myfunc(*args) as noted in comments below. Parameters can be mixed so long parameters with default values (optional arguments) follow any "positional" (required) parameters, and catchall parameter (*args) follow those, and any "keyword arguments' parameter" is last. (Any function can only have up to one "catchall" or *args' parameter and up to one "keyword args" **kwargs parameter). <lang python>def no_args():passcallno_args() def fixed_args(x, y):print('x=%r, y=%r' % (x, y))callfixed_args(1, 2) # x=1, y=2Can also called them using the parameter names, in either order:fixed_args(y=2, x=1)Can also "apply" fixed_args() to a sequence:myargs=(1,2) # tuple fixed_args(*myargs) def opt_args(x=1):print(x)callsopt_args() # 1 opt_args(3.141) # 3.141 def var_args(*v):print(v)callsvar_args(1, 2, 3) # (1, 2, 3) var_args(1, (2,3)) # (1, (2, 3)) var_args() # ()Named argumentsfixed_args(y=2, x=1) # x=1, y=2As a statementif 1:no_args()First-class within an expressionassert no_args() is None def return_something():return 1x = return_something() def is_builtin(x): print(x.__name__ in dir(__builtins__))callsis_builtin(pow) # True is_builtin(is_builtin) # FalseVery liberal function definitiondef takes_anything(*args, **kwargs):for each in args: print(each) for key, value in sorted(kwargs.items()): print("%s:%s" % (key, value)) # Passing those to another, wrapped, function: wrapped_fn(*args, **kwargs) # (Function being wrapped can have any parameter list # ... that doesn't have to match this prototype)A subroutine is merely a function that has no explicit return statement and will return None.Python uses "Call by Object Reference". See, for example, http://www.python-course.eu/passing_arguments.phpFor partial function application see: http://rosettacode.org/wiki/Partial_function_application#Python</lang>RTranslated from Python, when possible. <lang rsplus>### Calling a function that requires no arguments no_args <- function() NULL no_args()Calling a function with a fixed number of argumentsfixed_args <- function(x, y) print(paste("x=", x, ", y=", y, sep="")) fixed_args(1, 2) # x=1, y=2 fixed_args(y=2, x=1) # y=1, x=2Calling a function with optional argumentsopt_args <- function(x=1) x opt_args() # x=1 opt_args(3.141) # x=3.141Calling a function with a variable number of argumentsvar_args <- function(...) print(list(...)) var_args(1, 2, 3) var_args(1, c(2,3)) var_args()Calling a function with named argumentsfixed_args(y=2, x=1) # x=1, y=2Using a function in statement contextif (TRUE) no_args()Using a function in first-class context within an expressionprint(no_args)Obtaining the return value of a functionreturn_something <- function() 1 x <- return_something() xDistinguishing built-in functions and user-defined functions Not easily possible. See http://cran.r-project.org/doc/manuals/R-ints.html#g_t_002eInternal-vs-_002ePrimitive for details.Distinguishing subroutines and functions No such distinction.Stating whether arguments are passed by value or by reference Pass by value.Is partial application possible and how Yes, see http://rosettacode.org/wiki/Partial_function_application#R</lang>Racket<lang Racket>lang racketCalling a function that requires no arguments(foo)Calling a function with a fixed number of arguments(foo 1 2 3)Calling a function with optional arguments Calling a function with a variable number of arguments(foo 1 2 3) ; same in both casesCalling a function with named arguments(foo 1 2 #:x 3) ; using #:keywords for the namesUsing a function in statement context Using a function in first-class context within an expression Obtaining the return value of a function -> Makes no sense for Racket, as well as most other functional PLsDistinguishing built-in functions and user-defined functions(primitive? foo)but this is mostly useless, since most of Racket is implemented in itselfDistinguishing subroutines and functions -> No difference, though `!' is an idiomatic suffix for names of side-effect functions, and they usually return (void)Stating whether arguments are passed by value or by reference-> Always by value, but it's possible to implement languages with other argument passing styles, including passing arguments by reference (eg, there is "#lang algol60")Is partial application possible and how(curry foo 1 2) ; later apply this on 3 (λ(x) (foo 1 2 x)) ; a direct way of doing the same </lang>REXXversion 1<lang rexx>/*REXX program to demonstrate various methods of calling a REXX function*/ /*┌────────────────────────────────────────────────────────────────────┐│ Calling a function that REQUIRES no arguments. │ │ │ │ In the REXX language, there is no way to require the caller to not │ │ pass arguments, but the programmer can check if any arguments were │ │ (or weren't) passed. │ └────────────────────────────────────────────────────────────────────┘*/yr=yearFunc() say 'year=' yr exit yearFunc: procedure if arg()\==0 then call sayErr "SomeFunc function won't accept arguments." return left(date('Sorted'),3) /*┌────────────────────────────────────────────────────────────────────┐│ Calling a function with a fixed number of arguments. │ │ │ │ I take this to mean that the function requires a fixed number of │ │ arguments. As above, REXX doesn't enforce calling (or invoking) │ │ a (any) function with a certain number of arguments, but the │ │ programmer can check if the correct number of arguments have been │ │ specified (or not). │ └────────────────────────────────────────────────────────────────────┘*/ggg=FourFunc(12,abc,6+q,zz%2,'da 5th disagreement') say 'ggg squared=' ggg**2 exit FourFunc: procedure; parse arg a1,a2,a3; a4=arg(4) /*another way get a4*/ if arg()\==4 then docall sayErr "FourFunc function requires 4 arguments," call sayErr "but instead it found" arg() 'arguments.' exit 13 endreturn a1+a2+a3+a4 /*┌────────────────────────────────────────────────────────────────────┐│ Calling a function with optional arguments. │ │ │ │ Note that not passing an argument isn't the same as passing a null │ │ argument (a REXX variable whose value is length zero). │ └────────────────────────────────────────────────────────────────────┘*/x=12; w=x/2; y=x**2; z=x//7 /* z is x modulo seven.*/ say 'sum of w, x, y, & z=' SumIt(w,x,y,,z) /*pass 5 args, 4th is null*/ exit SumIt: procedure; sum=0do j=1 for arg() if arg(j,'E') then sum=sum+arg(j) /*the Jth arg may have been omitted*/ endreturn sum /*┌────────────────────────────────────────────────────────────────────┐│ Calling a function with a variable number of arguments. │ │ │ │ This situation isn't any different then the previous example. │ │ It's up to the programmer to code how to utilize the arguments. │ └────────────────────────────────────────────────────────────────────┘*//*┌────────────────────────────────────────────────────────────────────┐│ Calling a function with named arguments. │ │ │ │ REXX allows almost anything to be passed, so the following is one │ │ way this can be accomplished. │ └────────────────────────────────────────────────────────────────────┘*/what=parserFunc('name=Luna',"gravity=.1654",'moon=yes') say 'name=' common.name gr=common.gr say 'gravity=' gr exit parseFunc: procedure expose common.do j=1 for arg() parse var arg(j) name '=' val upper name call value 'COMMON.'name,val endreturn arg() /*┌────────────────────────────────────────────────────────────────────┐│ Calling a function in statement context. │ │ │ │ REXX allows functions to be called (invoked) two ways, the first │ │ example (above) is calling a function in statement context. │ └────────────────────────────────────────────────────────────────────┘*//*┌────────────────────────────────────────────────────────────────────┐│ Calling a function in within an expression. │ │ │ │ This is a variant of the first example. │ └────────────────────────────────────────────────────────────────────┘*/yr=yearFunc()+20 say 'two decades from now, the year will be:' yr exit /*┌────────────────────────────────────────────────────────────────────┐│ Obtaining the return value of a function. │ │ │ │ There are two ways to get the (return) value of a function. │ └────────────────────────────────────────────────────────────────────┘*/currYear=yearFunc() say 'the current year is' currYear call yearFunc say 'the current year is' result /*┌────────────────────────────────────────────────────────────────────┐│ Distinguishing built-in functions and user-defined functions. │ │ │ │ One objective of the REXX language is to allow the user to use any │ │ function (or subroutine) name whether or not there is a built-in │ │ function with the same name (there isn't a penality for this). │ └────────────────────────────────────────────────────────────────────┘*/qqq=date() /*number of real dates that Bob was on. */ say "Bob's been out" qqq 'times.' www='DATE'('USA') /*returns date in format mm/dd/yyy */ exit /*any function in quotes is external. */ date: return 4 /*┌────────────────────────────────────────────────────────────────────┐│ Distinguishing subroutines and functions. │ │ │ │ There is no programatic difference between subroutines and │ │ functions if the subroutine returns a value (which effectively │ │ makes it a function). REXX allows you to call a function as if │ │ it were a subroutine. │ └────────────────────────────────────────────────────────────────────┘*//*┌────────────────────────────────────────────────────────────────────┐│ In REXX, all arguments are passed by value, never by name, but it │ │ is possible to accomplish this if the variable's name is passed │ │ and the subroutine/function could use the built-in-function VALUE │ │ to retrieve the variable's value. │ └────────────────────────────────────────────────────────────────────┘*//*┌────────────────────────────────────────────────────────────────────┐│ In the REXX language, partial application is possible, depending │ │ how partial application is defined; I prefer the 1st definition (as│ │ (as per the "discussion" for "Partial Function Application" task: │ │ 1. The "syntactic sugar" that allows one to write (some examples│ │ are: map (f 7 9) [1..9] │ │ or: map(f(7,_,9),{1,...,9}) │ └────────────────────────────────────────────────────────────────────┘*/</lang>version 2<lang rexx>/* REXX ***************************************************************29.07.2013 Walter Pachl trying to address the task concisely f1 Calling a function that requires no arguments f2 Calling a function with a fixed number of arguments f3 Calling a function with optional arguments f4 Calling a function with a variable number of arguments f5 Calling a function with named arguments f6 Using a function in statement context f7 Using a function within an expression f8 Obtaining the return value of a function f8(...) is replaced by the returned value call f8 ... returned value is in special vatiable RESULT f9 Distinguishing built-in functions and user-defined functions bif is enforced by using its name quoted in uppercase fa,fb Distinguishing subroutines and functions Stating whether arguments are passed by value or by reference Arguments are passed by value ooRexx supports passing by reference (Use Arg instruction) Is partial application possible and how no ideas /say f1() Say f2(1,2,3) say f2(1,2,3,4) say f3(1,,,4) Say f4(1,2) Say f4(1,2,3) a=4700; b=11; Say f5('A','B') f6() /* returned value is used as command */ x=f7()**2 call f8 1,2; Say result '=' f8(1,2) f9: Say 'DATE'('S') date() call fa 11,22; Say result '=' fa(1,,2) /* the second comma above is for line continuation */Signal On Syntax Call fb 1,2 x=fb(1,2) Exit f1: Return 'f1 doesnt need an argument' f2: If arg()=3 ThenReturn 'f2: Sum of 3 arguments:' arg(1)+arg(2)+arg(3) Else Return 'f2: Invalid invocation:' arg() 'arguments. Needed: 3'f3: sum=0do i=1 To arg() If arg(i,'E')=0 Then Say 'f3: Argument' i 'omitted' Else sum=sum+arg(i) End Return 'f3 sum=' sumf4: sum=0; Do i=1 To arg(); sum=sum+arg(i); EndReturn 'f4: Sum of' arg() 'arguments is' sumf5: Parse Arg p1,p2Say 'f5: Argument 1 ('p1') contains' value(p1) Say 'f5: Argument 2 ('p2') contains' value(p2) Return 'f5: sum='value(p1)+value(p2)f6: Say 'f6: dir ft.rex'Return 'dir ft.rex'f7: Say 'f7 returns 7'Return 7f8: Say 'f8 returns arg(1)+arg(2)'Return arg(1)+arg(2)date: Say 'date is my date function'Return translate('ef/gh/abcd','DATE'('S'),'abcdefgh')fa: Say 'fa returns arg(1)+arg(2)'Return arg(1)+arg(2)fb: Say 'fb:' arg(1)','arg(2)ReturnSyntax:Say 'Syntax raised in line' sigl Say sourceline(sigl) Say 'rc='rc '('errortext(rc)')' If sigl=39 Then Say 'fb cannot be invoked as function (it does not return a value' Exit</lang>Output:f1 doesn't need an argument f2: Sum of 3 arguments: 6 f2: Invalid invocation: 4 arguments. Needed: 3 f3: Argument 2 omitted f3: Argument 3 omitted f3 sum= 5 f4: Sum of 2 arguments is 3 f4: Sum of 3 arguments is 6 f5: Argument 1 (A) contains 4700 f5: Argument 2 (B) contains 11 f5: sum=4711 f6: dir ft.rex Datenträger in Laufwerk D: ist DATA Volumeseriennummer: B0F8-F2C3 Verzeichnis von D:\ 29.07.2013 00:33 2.661 ft.rex 1 Datei(en), 2.661 Bytes 0 Verzeichnis(se), 251.050.979.328 Bytes frei f7 returns 7 f8 returns arg(1)+arg(2) f8 returns arg(1)+arg(2) 3 = 3 date is my date function 20130729 07/29/2013 fa returns arg(1)+arg(2) fa returns arg(1)+arg(2) 33 = 3 fb: 1,2 fb: 1,2 Syntax raised in line 39 x=fb(1,2) rc=44 (Function or message did not return data) fb cannot be invoked as function (it does not return a value)RubyRuby does not have functions, but Ruby classes have "methods" which are equivalent. The parentheses around the arguments are optional (definition and call). A method returns the last expression that was evaluated in the body of the method. The return keyword can be used to make it explicit that a method returns a value.Calling a function that requires no arguments<lang ruby>def foo() p "foo" endfoo #=> "foo" foo() #=> "foo"</lang>Calling a function with a fixed number of arguments<lang ruby>def foo arg; p arg end # one argumentfoo(1) #=> 1 foo "1" #=> "1" foo [0,1,2] #=> [0, 1, 2] (one Array)</lang>Calling a function with optional arguments<lang ruby>def foo(x=0, y=x, flag=true) p [x,y,flag] endfoo #=> [0, 0, true] foo(1) #=> [1, 1, true] foo(1,2) #=> [1, 2, true] foo 1,2,false #=> [1, 2, false]</lang>Calling a function with a variable number of arguments<lang ruby>def foo(*args) p args endfoo #=> [] foo(1,2,3,4,5) #=> [1, 2, 3, 4, 5]</lang>Calling a function with named arguments<lang ruby>def foo(id:0, name:"", age:0) p [id, name, age] endfoo(age:22, name:"Tom") #=> [0, "Tom", 22]</lang>Using a function in statement context?Using a function in first-class context within an expressionThe method is not a first-class function. However, there is Proc object. See First-class functions#RubyObtaining the return value of a function<lang ruby>def foo(a,b) a + b endbar = foo 10,20 p bar #=> 30 p foo("abc","def") #=> "abcdef"return multiple valuesdef sum_and_product(a,b) return a+b,a*b end x,y = sum_and_product(3,5) p x #=> 8 p y #=> 15</lang>Distinguishing built-in functions and user-defined functionsThere is no distinction.Distinguishing subroutines and functionsSubroutine and function don't exist at ruby. It is only a method that there is. The Kernel module is included by class Object, so its methods are available in every Ruby object. These methods are called without a receiver and thus can be called in functional form.<lang ruby>puts "OK!" # Kernel#putsraise "Error input" # Kernel#raise Integer("123") # Kernel#Integer rand(6) # Kernel#rand throw(:exit) # Kernel#throwmethod which can be seen like a reserved word.attr_accessor # Module#attr_accessor include # Module#include private # Module#private require # Kernel#require loop { } # Kernel#loop</lang>Stating whether arguments are passed by value or by referencepassed by value of object reference.Is partial application possible and howHowever something similar can be done, see Partial function application#RubyOthersBlock Argument: The block argument sends a closure from the calling scope to the method. The block argument is always last when sending a message to a method. A block is sent to a method using do ... end or { ... }. <lang ruby>class Arraydef sum(init=0, &blk) if blk inject(init){|s, n| s + blk.call(n)} else inject(init){|s, n| s + n} end endend ary = [1,2,3,4,5] p ary.sum #=> 15 p ary.sum(){|n| (-n).to_s} #=> "-1-2-3-4-5" p (ary.sum do |n| n * n end) #=> 55</lang>Splat operator: You can turn an Array into an argument list with * (or splat) operator. <lang ruby>def foo(a,b,c) p [a,b,c] endargs = [1,2,3] foo *args #=> [1, 2, 3] args = [1,2] foo(0,*args) #=> [0, 1, 2]</lang>Syntax sugar: In Ruby, many operators are actually method calls. <lang ruby># return value substancei = 3 p 1 + i #=> 4 1.+(i) p i < 5 #=> true i.<(5) p 2 ** i #=> 8 2.**(i) p -i #=> -3 i.-@() a = [1,2,3] p a[0] #=> 1 a.[](0) a[2] = "0" # a.[]=(2,"0") p a << 5 #=> [1, 2, "0", 5] a.<<(5) p a & [4,2] #=> [2] a.&([4,2]) p "abcde"[1..3] #=> "bcd" "abcde".[](1..3) p "%2d %4s" % [1,"xyz"] #=> " 1 xyz" "%2d %4s".%([1,"xyz"])</lang>Method call which was displayed in the comment is usable actually.ScalaLibrary: Scala<lang Scala>def ??? = throw new NotImplementedError // placeholder for implementation of hypothetical methods def myFunction0() = ??? myFunction0() // function invoked with empty parameter list myFunction0 // function invoked with empty parameter list omitted def myFunction = ??? myFunction // function invoked with no arguments or empty arg list /* myFunction() */ // error: does not take parameters def myFunction1(x: String) = ??? myFunction1("foobar") // function invoked with single argument myFunction1 { "foobar" } // function invoked with single argument provided by a block// (a block of code within {}'s' evaluates to the result of its last expression)def myFunction2(first: Int, second: String) = ??? val b = "foobar" myFunction2(6, b) // function with two arguments def multipleArgLists(first: Int)(second: Int, third: String) = ??? multipleArgLists(42)(17, "foobar") // function with three arguments in two argument lists def myOptionalParam(required: Int, optional: Int = 42) = ??? myOptionalParam(1) // function with optional param myOptionalParam(1, 2) // function with optional param provided def allParamsOptional(firstOpt: Int = 42, secondOpt: String = "foobar") = ??? allParamsOptional() // function with all optional args /* allParamsOptional */ // error: missing arguments for method allParamsOptional;// follow with `_' if you want to treat it as a partially applied functiondef sum[Int](values: Int*) = values.foldLeft(0)((a, b) => a + b) sum(1, 2, 3) // function accepting variable arguments as literal val values = List(1, 2, 3) sum(values: _*) // function acception variable arguments from collection sum() // function accepting empty variable arguments def mult(firstValue: Int, otherValues: Int*) = otherValues.foldLeft(firstValue)((a, b) => a * b) mult(1, 2, 3) // function with non-empty variable arguments myOptionalParam(required = 1) // function called with named arguments (all functions have named arguments) myFunction2(second = "foo", first = 1) // function with re-ordered named arguments mult(firstValue = 1, otherValues = 2, 3) // function with named variable argument as literal val otherValues = Seq(2, 3) mult(1, otherValues = otherValues: _*) // function with named variable argument from collection val result = myFunction0() // function called in an expression context myFunction0() // function called in statement context /* myOptionalParam(optional = 1, 2) */ // error: positional after named argument. def transform[In, Out](initial: In)(transformation: In => Out) = transformation(initial) val result = transform(42)(x => x * x) // function in first-class context within an expression def divide(top: Double, bottom: Double) = top / bottom val div = (divide _) // partial application -- defer application of entire arg list val halve = divide(_: Double, 2) // partial application -- defer application of some arguments class Foo(var value: Int) def incFoo(foo: Foo) = foo.value += 1 // function showing AnyRef's are passed by reference /* def incInt(i: Int) = i += 1 */ // error: += is not a member of Int// (All arguments are passed by reference, but reassignment // or setter must be defined on a type or a field // (respectively) in order to modify its value.)// No distinction between built-in functions and user-defined functions // No distinction between subroutines and functions</lang>Seed7Seed7 provides two kinds of subroutines: proc, which has no return value, and func, which has a return value. The return value of a func must be used by the caller (e.g. assigned to a variable). If you don't want do deal with the return value, use a proc instead.Seed7 supports call-by-value, call-by-reference, and call-by-name parameters. Programmers are free to specify the desired parameter passing mechanism. The most used parameter passing mechanism is 'in'. Depending on the type 'in' specifies call-by-value (for integer, float, ...) or call-by-reference (for string, array, ...). It is prohibited to write something to an 'in' parameter.All parameters are positional.There are no differences between between calling built-in vs. user defined functions.<lang seed7>env := environment; # Call a function that requires no arguments.env := environment(); # Alternative possibility to call of a function with no arguments. cmp := compare(i, j); # Call a function with a fixed number of arguments.</lang>There are no optional arguments, but a similar effect can be achieved with overloading.<lang seed7>write(aFile, "asdf"); # Variant of write with a parameter to specify a file.write("asdf"); # Variant of write which writes to the file OUT.</lang>Seed7 does not support functions with a variable number of arguments. But a function argument can be an array with as many values as you want:<lang seed7>const func integer: sum (in array integer: intElems) is funcresult var integer: sum is 0; local var integer: element is 0; begin for element range intElems do sum +:= element; end for; end func;s := sum([] (1, 2, 3)); # Use an aggregate to generate an array. t := sum([] (2, 3, 5, 7));</lang>Concatenation operators can be used to concatenate arguments. This solution is used to provide the write function:<lang seed7>write("Nr: " <& num); # Use operators to concatenate arguments.</lang>The procedure ignore can be used to ignore a return value.<lang seed7>ignore(getln(IN)); # Using a function in statement context (ignore the result).</lang>Call-by-name parameters use a function in first-class context. The function doMap from the examples section of the Seed7 homepage uses a given expression to modify the elements of an array:<lang seed7>seq := doMap([](1, 2, 4, 6, 10, 12, 16), x, succ(x));</lang>SmalltalkWhere f is a closure and arguments is an array of values for f to operate on. <lang smalltalk>f valueWithArguments: arguments.</lang>Swift<lang Swift>// call a function with no args noArgs() // call a function with one arg with no external name oneArgUnnamed(1) // call a function with one arg with external name oneArgNamed(arg: 1) // call a function with two args with no external names twoArgsUnnamed(1, 2) // call a function with two args and external names twoArgsNamed(arg1: 1, arg2: 2) // call a function with an optional arg // with arg optionalArguments(arg: 1) // without optionalArguments() // defaults to 0 // function that takes another function as arg funcArg(noArgs) // variadic function variadic(opts: "foo", "bar") // getting a return value let foo = returnString() // getting a bunch of return values let (foo, bar, baz) = returnSomeValues() // getting a bunch of return values, discarding second returned value let (foo, _, baz) = returnSomeValues()</lang>Tcl<lang tcl>aCallToACommandWithNoArguments aCallToACommandWithOne argument aCallToACommandWith arbitrarily many arguments aCallToACommandWith {*}$manyArgumentsComingFromAListInAVariable aCallToACommandWith -oneNamed argument -andAnother namedArgument aCallToACommandWith theNameOfAnotherCommand aCallToOneCommand [withTheResultOfAnother]</lang> Tcl does differentiate between functions and other types of commands in expressions: <lang tcl>expr {func() + [cmd]} expr {func(1,2,3} + [cmd a b c]}</lang> However, there are no deep differences between the two: functions are translated into commands that are called in a particular namespace (thus foo() becomes tcl::mathfunc::foo). There are no differences in usage between built-in commands and user-defined ones, and parameters are passed to commands by value conceptually (and read-only reference in the implementation).UNIX ShellIn the shell, there are no argument specifications for functions. Functions obtain their arguments using the positional parameter facilities and functions are simply called by name followed by any arguments that are to be passed: <lang sh>sayhello # Call a function in statement context with no arguments multiply 3 4 # Call a function in statement context with two arguments</lang> The shell does not support the use of named parameters. There is no lookahead in the shell, so functions cannot be called until their definition has been run.XSLT<lang xml><?xml version="1.0" encoding="UTF-8"?> <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0"><xsl:output method="xml" indent="yes"/> <xsl:template match="/"> <demo> <xsl:call-template name="xpath-function-demos"/> <xsl:call-template name="xslt-template-demos"/> </demo> </xsl:template> <xsl:template name="xpath-function-demos"> <line>This test is <xsl:if test="false()">NOT</xsl:if> OK.</line> <line>'haystack' does <xsl:if test="not(starts-with('haystack', 'hay'))">NOT</xsl:if> start with 'hay'.</line> <line>'<xsl:value-of select="substring('haystack', 1, 3)"/>' = 'hay'</line> <line>'<xsl:value-of select="substring('haystack', 4)"/>' = 'stack'</line> <line>'<xsl:value-of select="concat('abcd', 'efgh')"/>' = 'abcdefgh'</line> <line>'<xsl:value-of select="concat('ij', 'kl', 'mn', 'op')"/>' = 'ijklmnop'</line> <line>The number of root elements in the input document is <xsl:value-of select="count(/*)"/> (should be 1).</line> <line foo="{concat('Hello, ', 'Hello, ', 'Hello')}!">See attribute.</line> </xsl:template> <xsl:template name="xslt-template-demos"> <xsl:call-template name="nullary-demo"/> <xsl:apply-templates select="/*" mode="nullary-context-demo-mode"/> <working note="When all parameters are supplied"> <xsl:call-template name="ternary-demo"> <xsl:with-param name="a" select="4"/> <xsl:with-param name="b">3</xsl:with-param> <xsl:with-param name="c" select="2 + 3"/> </xsl:call-template> </working> <broken note="When the third parameter 'c' is omitted"> <xsl:call-template name="ternary-demo"> <xsl:with-param name="a" select="4"/> <xsl:with-param name="b">3</xsl:with-param> </xsl:call-template> </broken> <working name="When all parameters are supplied"> <xsl:call-template name="binary-or-ternary-demo"> <xsl:with-param name="a" select="4"/> <xsl:with-param name="b" select="3"/> <xsl:with-param name="c" select="5"/> </xsl:call-template> </working> <working name="When 'a' and 'b' are supplied but 'c' is defaulted to 0"> <xsl:call-template name="binary-or-ternary-demo"> <xsl:with-param name="a" select="4"/> <xsl:with-param name="b" select="3"/> </xsl:call-template> </working> <xsl:variable name="lesser-demo-result"> <xsl:call-template name="lesser-value"> <xsl:with-param name="a"> <xsl:call-template name="lesser-value"> <xsl:with-param name="a" select="34"/> <xsl:with-param name="b" select="78"/> </xsl:call-template> </xsl:with-param> <xsl:with-param name="b"> <xsl:call-template name="lesser-value"> <xsl:with-param name="a" select="12"/> <xsl:with-param name="b" select="56"/> </xsl:call-template> </xsl:with-param> </xsl:call-template> </xsl:variable> <line> <xsl:value-of select="concat('And the answer, which should be 12, is ', $lesser-demo-result, ', of course.')"/> </line> </xsl:template> <xsl:template match="/*" mode="nullary-context-demo-mode"> <xsl:call-template name="nullary-context-demo"/> </xsl:template> <xsl:template name="nullary-demo"> <line>No parameters needed here!</line> </xsl:template> <xsl:template name="nullary-context-demo"> <xsl:for-each select="self::*"> <line>The context element here is named "<xsl:value-of select="local-name()"/>"</line> </xsl:for-each> </xsl:template> <xsl:template name="ternary-demo"> <xsl:param name="a"/> <xsl:param name="b"/> <xsl:param name="c"/> <line>(<xsl:value-of select="$a"/> * <xsl:value-of select="$b"/>) + <xsl:value-of select="$c"/> = <xsl:value-of select="($a * $b) + $c"/></line> </xsl:template> <xsl:template name="binary-or-ternary-demo"> <xsl:param name="a"/> <xsl:param name="b"/> <xsl:param name="c" select="0"/> <line>(<xsl:value-of select="$a"/> * <xsl:value-of select="$b"/>) + <xsl:value-of select="$c"/> = <xsl:value-of select="($a * $b) + $c"/></line> </xsl:template> <xsl:template name="lesser-value"> <xsl:param name="a"/> <xsl:param name="b"/> <xsl:choose> <xsl:when test="number($a) < number($b)"> <xsl:value-of select="$a"/> </xsl:when> <xsl:otherwise> <xsl:value-of select="$b"/> </xsl:otherwise> </xsl:choose> </xsl:template></xsl:stylesheet> </lang>zklThe syntax and semantics of function calls is the always the same: name/object(parameters). All calls are varargs, it is up to the callee to do default/optional parameter handling (but that is hidden from the programmer). No named parameters. Pass by reference or value, depending. Using f has a function, method or object: <lang zkl>f(); f(1,2,3,4); fcn f(a=1){}() // define and call f, which gets a set to 1 fcn{vm.arglist}(1,2,3,4) // arglist is L(1,2,3,4) fcn{a1:=vm.nthArg(1)}(1,2,3) // a1 == 2 (f() == True); (f() and 1 or 2) if (f()) println() f(f) // pass f to itself s:=f() fcn{}.isType(self.fcn) //True fcn{}.len.isType(self.fcn) //False, len is a Method</lang> Partial application is done with the .fp* methods or the 'wrap keyword <lang zkl> fcn(a,b,c).fp(1)() // call function with a always set to 1 fcn(a,b,c).fp1(2,3)() // call function with b & c always set to 2 & 3 fcn(a,b,c,d).fpN(3,5)() // call function with d always set to 5 fcn{vm.arglist}.fpN(3,66)(1,2,3,4,5) //-->L(1,2,3,66,4,5) fcn{}.fpM("01-",5) // use a mask to select parameters// 1 is supplied, 0 is get upon call, - is chop arglistfcn{vm.arglist}.fpM("01-",66)(1,2,3,4) //-->L(1,66) a:=5; f('wrap(b){a+b}) // 'wrap is syntactic sugar for .fpN// to create a lexical closure --> f(fcn(b,a){a+b}.fpN(1,a))</lang>ZX Spectrum BasicOn the ZX Spectrum, functions and subroutines are separate entities. A function is limited to being a single expression that generates a return value. Statements are not allowed within a function. A subroutine can perform input and output and can contain statements. <lang zxbasic>10 REM functions cannot be called in statement context 20 PRINT FN a(5): REM The function is used in first class context. Arguments are not named 30 PRINT FN b(): REM Here we call a function that has no arguments 40 REM subroutines cannot be passed parameters, however variables are global 50 LET n=1: REM This variable will be visible to the called subroutine 60 GO SUB 1000: REM subroutines are called by line number and do not have names 70 REM subroutines do not return a value, but we can see any variables it defined 80 REM subroutines cannot be used in first class context 90 REM builtin functions are used in first class context, and do not need the FN keyword prefix 100 PRINT SIN(50): REM here we pass a parameter to a builtin function 110 PRINT RND(): REM here we use a builtin function without parameters 120 RANDOMIZE: REM statements are not functions and cannot be used in first class context.</lang>