Whitespace is an esoteric programming language that uses only three characters:
[space]
or[tab]
or\t
(ASCII 9)[line-feed]
or\n
(ASCII 10)
Whitespace is an imperative, stack-based programming language, including features such as subroutines.
Each command in whitespace begins with an Instruction Modification Parameter (IMP).
- [space]: Stack Manipulation
- [tab][space]: Arithmetic
- [tab][tab]: Heap Access
- [tab][line-feed]: Input/Output
- [line-feed]: Flow Control
There are two types of data a command may be passed: numbers and labels.
- Numbers begin with a
[sign]
symbol. The sign symbol is either[tab]
-> negative, or[space]
-> positive. - Numbers end with a
[terminal]
symbol:[line-feed]
. - Between the sign symbol and the terminal symbol are binary digits
[space]
-> binary-0, or[tab]
-> binary-1. - A number expression
[sign][terminal]
will be treated as zero. - The expression of just
[terminal]
should throw an error. (The Haskell implementation is inconsistent about this.)
- Labels begin with any number of
[tab]
and[space]
characters. - Labels end with a terminal symbol:
[line-feed]
. - Unlike with numbers, the expression of just
[terminal]
is valid. - Labels must be unique.
- A label may be declared either before or after a command that refers to it.
As stated earlier, there commands may read data from input or write to output.
Whitespace will accept input either characters or integers. Due to the lack of an input stream mechanism, the input will be passed as a string to the interpreter function.
- Reading a character involves simply taking a character from the input stream.
- Reading an integer involves parsing a decimal or hexadecimal number from the current position of the input stream, up to and terminated by a line-feed character.
- The original implementation being in Haskell has stricter requirements for parsing an integer.
- The Javascript and Coffeescript implementations will accept any number that can be parsed by the parseInt function as a single parameter.
- The Python implementations will accept any number that can be parsed by the int function as a single parameter.
- An error should be thrown if the input ends before parsing is complete. (This is a non-issue for the Haskell implementation, as it expects user input)
- For a number, append the output string with the number's string value.
- For a character, simply append the output string with the character.
Notation: n specifies the parameter, [number]
or [label]
.
Errors should be thrown for invalid numbers, labels, and heap addresses, or if there are not enough items on the stack to complete an operation (unless otherwise specified). In addition, an error should be thrown for unclean termination.
[space]
(number): Push n onto the stack.[tab][space]
(number): Duplicate the _n_th value from the top of the stack.[tab][line-feed]
(number): Discard the top n values below the top of the stack from the stack. (For n<0 or n>=stack.length, remove everything but the top value.)[line-feed][space]
: Duplicate the top value on the stack.[line-feed][tab]
: Swap the top two value on the stack.[line-feed][line-feed]
: Discard the top value on the stack.
[space][space]
: Popa
andb
, then pushb+a
.[space][tab]
: Popa
andb
, then pushb-a
.[space][line-feed]
: Popa
andb
, then pushb*a
.[tab][space]
: Popa
andb
, then pushb/a
-
- If
a
is zero, throw an error.
- If
-
- *Note that the result is defined as the floor of the quotient.
[tab][tab]
: Popa
andb
, then pushb%a
-
- If
a
is zero, throw an error.
- If
-
- *Note that the result is defined as the remainder after division and sign (+/-) of the divisor (a).
[space]
: Popa
andb
, then storea
at heap addressb
.[tab]
: Popa
and then push the value at heap addressa
onto the stack.
[space][space]
: Pop a value off the stack and output it as a character.[space][tab]
: Pop a value off the stack and output it as a number.[tab][space]
: Read a character from input,a
, Pop a value off the stack,b
, then store the ASCII value ofa
at heap addressb
.[tab][tab]
: Read a number from input,a
, Pop a value off the stack,b
, then storea
at heap addressb
.
[space][space]
(label): Mark a location in the program with label n.[space][tab]
(label): Call a subroutine with the location specified by label n.[space][line-feed]
(label): Jump unconditionally to the position specified by label n.[tab][space]
(label): Pop a value off the stack and jump to the label specified by n if the value is zero.[tab][tab]
(label): Pop a value off the stack and jump to the label specified by n if the value is less than zero.[tab][line-feed]
: Exit a subroutine and return control to the location from which the subroutine was called.[line-feed][line-feed]
: Exit the program.
Note that in Javascript and Coffeescript, the modulus operator is implemented differently than it was in the original Whitespace interpreter, which was influenced by it being defined in Haskell. Javascript and Coffeescript also lack integer division operations. There is no difference between Whitespace and Python in the same regard. Therefore, the Javascript and Coffeescript interpreters require a little extra attention in regard to the implementation in regard to integer division and the modulus operator (See: floored division in the Wikipedia article "Modulo operation".