1) The culprit isn’t the parens, it’s the comma. A Tuple is defined as a comma-separated list of items. The parens are not specifically needed. With that said, the precedence of the “,” operator is fairly low, so the expression “foo a,b” is taken as a 2-tuple, the first part being the value of “foo a” and the second being “b”. However if you wrote it with parens, “foo (a, b)”, this applies the function “foo” to the 2-tuple “a, b”.

2) Technically, when you call a .NET Framework function (“foo”) that takes (say) three arguments, the F# interface requires that you pass a *single* parameter, which in this case would be a 3-tuple. Hence the parens are required (as per the previous point). So the difference is that in F# you need the parens so that the statement parses the way you want it to, whereas in C# they’re necessary simply because that’s a syntactic requirement of the language.

2a) I’ve never tried it, but presumably you could rewrite “foo(a, b, c)” as two statements, “let args = a, b, c” followed by “foo args” (and with no parens at all).

Re your point about single pass compilation, I recently asked Don Syme about making multiple source passes to improve the type inference process. His reply was “Yes, it’s possible to do multi-pass type inference. There are also single-pass variations that generate a finite set of constraints.

However these approaches tend to give bad error messages and poor intellisense results in a visual editor.
“

Thank you for submitting this cool story – Trackback from DotNetShoutout…