🚀 go-pugleaf

To distill the essence of a recent thread from comp.lang.c --

I raised a question about the term "obsolete" being used in regard to
the storage class specifier "auto". It is my meek contention that,
though the use of the auto keyword may certainly be redundant, I
believe there's nothing in the Standard nor the Rationale that either
shows it to be obsolete, heading toward obsolescence or disparaged.
(I must come clean and shamefully confess that I *occasionally* use
the auto keyword for symmetry with the static storage class specifier
in certain rare contexts).

As the debates over the merits of "auto" raged on, one of the more
respected members of the c.l.c community, Dan.Pop@mail.cern.ch put
forth the proposition:

> It's [auto] obsolete because it's useless, not because the standard
> says so. It's in the standard for a single purpose: to avoid breaking
> existing code.

Certainly my intent is not to engage Mr. Pop in argument (which
readers of c.l.c know is sure to be a no-win situation for his hapless
challenger), I don't know that I can agree completely with his
assertion that the avoidance of breaking existing code is the *single*
purpose of auto being in the Standard.

I have resigned myself to thinking that Mr. Pop, as always, is probably
quite correct in his statement -- although I was personally unable
to locate "avoid breaking existing code" used with regard to auto in
either of the aforementioned documents. Moreover, it *seems* that the
Committee's inclusion of the unary + operator *may* similarily be used
to justify an explicit "auto" for reasons of symmetry [Rationale 3.3.3.3]
amongst the other more worthy storage class specifiers.

So...to the readers of comp.std.c (and perhaps esteemed members of
the Committee who may be hereabouts), I humbly ask: Is there any
value whatsoever to the auto keyword short of lingering around merely
for the benefit of existing code? (If I find that auto is indeed
useless or obsolete, not only will I fire up a sed script to banish
it from my current corpus of code -- I hereby swear [or affirm] on my
tattered copy of K&R2 that this second class citizen of keywords
will be most unwelcome in my future C efforts. Amen.)

-- 
=============================================================================
          Bob Nelson: Dallas, Texas, U.S.A.  -  bnelson@netcom.com
      Linux for fun, M$ for $$$...and the NFL for what really counts!
=============================================================================