Icon

Appeared in:
1977
Paradigm:
Typing discipline:
File extensions:
.icn
Versions and implementations (Collapse all | Expand all):
Programming language

Icon is a high-level general-purpose programming language suited for tasks that involve strings and data structures.

Icon was created in 1977 at the University of Arisona. It has objectives and capabilities similar to those of SNOBOL, but doesn’t really resemble it. Note that the language was developed much earlier than graphical user interfaces were invented, let alone became popular, so the name is not associated with this meaning of word “icon”.

Icon is a procedural language with strong dynamic typing. Its key features are:

  • mechanism of expression evaluation which uses goal-oriented evaluation and backtracking.
  • generators — expressions which produce more than one result. All expressions and imperative constructs of the language are potential generators. A special control structure every-do allows to process all results returned by a generator.
  • string scanning: a mechanism of processing strings in a convenient way by binding a string to an expression and scanning it with string matching functions.
  • support for several types of data structures: lists (linear collection which allows access both by element position and by stack/queue-style functions), sets (collections without any particular ordering) and tables (associative lookup collections).
  • high-level window-based graphic facilities.
  • automated garbage collection.
  • integers with arbitrary precision.
  • the original language is not object-oriented; in 1996 an object-oriented extension was developed which late became Unicon.

Elements of syntax:

Inline comments #
Variable assignment := or <- (reversable at backtracking)
Block { ... }
Physical (shallow) equality = (numbers) or == (strings)
Physical (shallow) inequality ~= (numbers) or ~== (strings)
Deep equality ===
Deep inequality ~===
Comparison < > <= >= << >> <<= >>=
Function definition procedure f(p1, p2, ...) ... end
Function call f(a, b, ...)
Function call with no parameters f()
Sequence ;
If - then if condition then ...
If - then - else if condition then ... else ...
Loop forever loop ...
While condition do while condition do ...
For each value in a numeric range, 1 increment every i := 1 to 10 do ...
For each value in a numeric range, 1 decrement every i := 10 to 1 by -1 do ...

Icon logo
Icon logo

Examples:

Hello, World!:

Example for versions iconc 9.4
procedure main ()
   write ("Hello, world!")
end

Factorial:

Example for versions iconc 9.4

This example uses recursive factorial definition. 0 to 16 is a generator which returns all integers between 0 and 16, inclusive. every ... do allows to fetch every result returned by the generator and process it (calculate factorial and perform write).

procedure factorial (n)
   if n = 0 then
      return 1
   else if n > 0 then
      return n * factorial (n - 1)
end

procedure main ()
   local i
   every i := 0 to 16 do
      write (i, "! = ", factorial (i))
end

Factorial:

Example for versions iconc 9.4

This example uses iterative factorial definition written in short form. every j *= 1 to i multiplies j by all values returned by the generator 1 to i. The expression in braces calculates factorial by multiplying j by all values between 1 and i and returns it. The outer expression calculates and prints factorials of all numbers returned by the generator 0 to 16.

procedure main ()
   local i, j
   every write (i := 0 to 16, "! = ", { j := 1; every j *:= 1 to i; j })
end

Fibonacci numbers:

Example for versions iconc 9.4

This example uses recursive calculation of Fibonacci numbers with memoization.

global fib_memo

procedure fib (n)
   if n >= 0 then
      return ((/fib_memo [n] := fib (n - 2) + fib (n - 1)) | fib_memo [n])
end

procedure main ()
   local i
   fib_memo := table ()
   fib_memo [0] := 0; fib_memo [1] := 1
   every i := 1 to 16 do {
      writes (fib (i) || ", ")
   }
   write("...")
end

Quadratic equation:

Example for versions iconc 9.4
procedure main () 
    A := read();
    if A = 0 then {
        write("Not a quadratic equation.");
        return;
    }
    B := read();
    C := read();
    D := B*B - 4*A*C;
    if D = 0 then {
        write("x = ", -B/2/A);
        return;
    }
    if D > 0 then {
        write("x1 = ", (-B+sqrt(D))/2/A);
        write("x2 = ", (-B-sqrt(D))/2/A);
    } else {
        write("x1 = (", -B/2/A, ",", sqrt(-D)/2/A, ")");
        write("x2 = (", -B/2/A, ",", -sqrt(-D)/2/A, ")");
    }
end

CamelCase:

Example for versions iconc 9.4

First of all the program reads the string to process and adds a space to its end (|| is concatenation operator). After this text variable is scanned. ? is an operator which binds a string to the expression, so that all string matching functions in the expression are performed on this string.

Commands ReFind and ReMatch from regular expressions library regexp both find all sequences of symbols which match the given regex, but ReFind returns the index of sequence start, and ReMatch — the index of the first character after sequence end. In one iteration ReFind finds the start of the next sequence of non-letter characters. Command tab moves current position pointer to this position and returns part of the string from previous position of the pointer to the new position — a word. After this the word is converted to proper case and concatenated to the resulting string. *word is a function which returns length of the string. map replaces all characters of its first argument which are present in its second argument with corresponding characters from its third argument (in this case — switch case of characters; &lcase and &ucase are built-in constants which contain lowercase and uppercase alphabet). Finally, one more call of tab moves the pointer past the sequence of non-letter characters to the start of next word.

link regexp

procedure main () 
    text := read() || " ";
    cc := "";
    text ? {
        while j := ReFind("[^a-zA-Z]+") do {
            word := tab(j);
            cc ||:= map(word[1],&lcase,&ucase) || map(word[2:*word+1],&ucase,&lcase);
            tab(ReMatch("[^a-zA-Z]+"));
        }
    }
    write (cc);
end