In article <199603281233.GAA16242@solutions.solon.com>,
Peter Seebach  <seebs@solon.com> wrote:
> I would argue that there's something fundementally wrong with it *in
> a Unix-like environment*, which is that the entire system is assuming
> that a process exits with some kind of status, and declaring main with
> no return type strikes me as dishonest.

But your making the assumption that main's return and the process exit
value must somehow be connected. That's certainly a sensible way for
things to be arranged, but it doesn't have to be that way. Some might
say exit from the program is a system related interaction and should
require a system call, and say falling off the end of main, or returning
from main, is sloppy. But I suppose it's irrelevant, because the
standard we have now says things are the way we are, and the C9X
comittee probably doesn't care very much what any of us think.