Procedure Signature Styles and Ambiguities

Style_Table.md

	Prefix/Odin	Require Return Type	Return Named Inputs
Empty Procedure	proc()	() -> void	()
1 input, unnamed	proc(int)	(int) -> void
1 input, named	proc(x: int)	(x: int) -> void	(x: int)
1 input, 1 output	proc(int) -> int	(x: int) -> int	(x: int) -> int
2 inputs, 2 unnamed outputs	proc(int, int) -> (int, int)	(int, int) -> (int, int)	(x: int, y: int) -> (int, int)
0 inputs, 2 named outputs	proc() -> (a: int, b: int)	() -> (a: int, b: int)
2 inputs, 2 outputs, all named	proc(a, b: int) -> (c, d: int)	(a, b: int) -> (c, d: int)

Prefix/Odin Style

Advantages: Allows for any approach and any style of a procedure signature

Disadvantages: Requires an extra keyword or sigil

Require Return Type Style

Advantages: Allows for most approaches of a procedure signature

Disadvantages: Requires the concept of void in the language

Return Named Inputs

Advantages: Does not require the concept of void in the language, allow for the return value to be omitted

Disadvantages: It's worse than the other two styles in every other way

x_original.txt

()
() -> int
(int) // ambiguity
(x: int)
(x: int) -> int
(x: int) -> (int, int) // ambiguity
(int) -> (int, int) // ambiguity
(x: int) -> (x: int, y: int) // ambiguity
(x: int) -> (y, z: int) // ambiguity

Solutions:
* Prefix
* Return value
* Named inputs and not named outputs


// Prefix (best compromise with the most options)
proc()
proc() -> int
proc(int)
proc(x: int)
proc(x: int) -> int
proc(x: int) -> (int, int)
proc(int) -> (int, int)
proc(x: int) -> (y: int, z: int)
proc(x: int) -> (y, z: int)

// Return value
() -> void
() -> int
(int) -> void
(x: int) -> void
(x: int) -> int
(x: int) -> (int, int)
(int) -> (int, int)
(x: int) -> (int, int)
(x: int) -> (y: int, z: int) // possible but harder to parse
(x: int) -> (y, z: int) // possible but harder to parse


// Named inputs
()
() -> int
(x: int) -> void
(x: int) -> int
(x: int) -> (int, int)
(x: int) -> (y: int, z: int) // ambiguity
(x: int) -> (y, z: int) // ambiguity