Declaring variables inside for loop initialization sections is indeed C99-only. I wouldn't do it if I were you.

double just has more precision (usually). Oftentimes, a float is 4 bytes and a double is 8. There used to be a speed penalty for using floating-point numbers, which is why float exists, but there's not really any reason not to use double any more.

[edit] *hides even further away* [/edit]

>*hides even further away*
There's nothing hidden there. Obfuscated, perhaps. Quite often, hard to read code, it is associated with cleverness, when the opposite it is true.

An abstraction layer is the only place where "hiding" some "goring details" of a program should be used. And interesting enough, it is for the sake of simplicity.