On Fri, Jul 10, 2009 at 09:13:39AM +0100, Robert Jones wrote:
So have I entirely missed the point of tuples?
The point is that tuple<int,int> is the same type as any other tuple<int,int>. With ad-hoc structures, struct A {int x, y; }; struct B {int x, y; }; the two are completely different types, so you can't write f.ex., a=b, whereas you could do that with tuples. So tuples help to minimize the number of ad-hoc structures in your program, and enable you to write code that is more generic.. [i.e., you can pass tuple<int,int> from one subsystem of your program to another with no problems, whereas it'd be more harder to pass A from subsys1 to subsys2 expecting B. you *could* implement implicit conversion operators from A to B and vice-versa, but that sounds like calling for trouble..] + they predefine a number of useful operations for you, such as lexicographic comparison, I/O operators, etc. With fusion maps, you get to use names, instead of indexes. (fusion maps, in effect, replace a.x syntax with at_key<x>(a) syntax.. somewhat more verbose, with the net win of the extra functionality..)