Forward declarations in function prototypes, yet again

[Doug Gwyn (VLD/VMB)]

> the ENTIRE THING is the declaration, and THE ENTIRE THING must be read before
> we try to interpret any part.  The fact that this declaration happens to have
> some sub-pieces that look like other declarations is not relevant.

If you're talking about the programming language C, then the above is not
correct.  C already has identifier scope rules and a declaration grammar
that cover function declarations, and the consistency RMS was talking about
is consistency with this, not self-consistency of some postulated C-like
language that doesn't preserve these structural rules.

I must say that most of this discussion has shown no consideration of the
previous debates on this issue.  I'd like to suggest that all the previous
NCEG documents on the issue be studied before further effort is wasted.