Wither Self

[Joachim Durchholz]

Joe Armstrong wrote:
> 
>     Shouldn't there be a Self in funs?
> 
>     IMHO writing:
> 
> 	Fac = fun(0) -> 1;	
> 		 (N) -> N*Self(N-1) end
>  
>     is just teensy weensy bit simpler than:
> 
> 	Fact = fun(X) -> 
>             G = fun(0, F)  -> 1; 
>                    (N, F) -> N*F(N-1,F) 
>                 end, 
>             G(X, G) 
>           end.

That's nice, but it doesn't handle mutual recursion.
Personally, I'd prefer a scope construct that first introduces a set of 
names, then binds definitions to them.
Syntactically, this need not have an ugly syntax; I'm thinking along 
lines like these:

	define
		Fac = fun/1
			0 -> 1;
			N -> N * Fac (N - 1)
		end
		some_constant = 2;
		Even = fun/1
			0 -> True;
			N -> odd (N - 1)
		end
		Odd = fun/1
			1 -> False;
			N -> Even (N - 1)
		end
	end

(I know this is very different from the current fun syntax, more a "how 
I'd like it if I could designed it scratch" thing.)

Just my 2c.

Regards,
Jo