It seems some people mistakenly believe that C is a high level language :-)
In fact in the hands of an experienced programmer C can be a structured
assembler. That is the derogatory term for it that most structured language
bigots (Pascal, M2, etc) use. Take for example a routine to toggle a bit
on and off. If you code it:

#define PORTB *((char *)(0x6))
main() {
        while (1) {
                PORTB ^= 1;
        }

And what you get is nearly identical to what you would write by hand. Now
you could code it "stupidly" as in:

void poke(a, b) {
        *(char *)a = b;
}

main() {
        char    a;

        a = 1;
        while (1) {
                if (a == 1) {
                        poke(6, 1);
                        a = 0;
                } else {
                        poke(6, 0);
                        a = 1;
                }
        }
}

You would take up lots of instructions and it would look as if C was a
very inefficient language to write in.

Basically C can be written like assembler. And people who use C in embedded
applications use it that way. Unfortunately most programming texts that use
C, teach it as though it is a much more sophisticated language :-). C was
written originally as a "structured assembler" for the PDP-11, so that folks
at Bell Labs could write the UNIX system code easily and recognize what was
going on just by looking at what they had written. This is a step up from
raw assembler and the BLISS compiler was either non-existant or not powerful
enough to do the same sort of thing.

The bottom line is that it makes perfect sense to have a C compiler for
PIC chips. However, this doesn't mean that you should expect to write anything
more complicated than you currently can, just that the source to what you
write will be easier to read later. Also, in my opinion, if you are on the
absolute edge of performance or space requirements, you'll probably end up
in assembler anyway.

--Chuck