Path: utzoo!mnetor!uunet!husc6!cmcl2!nrl-cmf!ames!amdahl!pyramid!prls!philabs!micomvax!ray
From: ray@micomvax.UUCP (Ray Dunn)
Newsgroups: comp.lang.c
Subject: Re: Why I won't use ANSI C
Message-ID: <959@micomvax.UUCP>
Date: 29 Mar 88 17:53:26 GMT
References: <7563@elsie.UUCP> <7022@brl-smoke.ARPA> <7564@elsie.UUCP> <7290@brl-smoke.ARPA> <671@cresswell.quintus.UUCP> <933@micomvax.UUCP> <7436@brl-smoke.ARPA>
Reply-To: ray@micomvax.UUCP (Ray Dunn)
Organization: Philips Electronics Ltd. (TDS - Montreal) St. Laurent P.Q., Canada
Lines: 55

In article <7436@brl-smoke.ARPA> gwyn@brl.arpa (Doug Gwyn) replies to a
posting of mine:
 >......
 >The one real advantage that participating in X3J11 has given me is that
               ^^^^^^^^^
 >I've heard all the arguments for changing = to := and the other such
 >ideas that keep popping up in this newsgroup many times before.  Some
 >topics, such as = vs. ==, have probably been sparked by magazine articles,
 >since the same old arguments keep getting repeated.

I always suspected Doug was a masochist!! (:-)(:-)!!

To be fair to your audience Doug, you often post only conclusions, which
appear to be opinions, when a little piece of the arguments considered by
X3J11 would lend a lot of credence to the posting and might *sigh* suppress
some of the net.traffic.  E.G.:

 >...
 >By the way, = vs. == was indeed considered by X3J11, but what can we do
 >about it?  If it were changed, the result could not reasonably be called
 >"C".  In fact, as I recall, most of us don't even think it should be
 >changed if it were feasible to do so.  The only approach that is at all
 >compatible with a large body of C code (but by no means all) would be
 >to introduce Boolean expressions and require that the control conditions
 >of if(), while(), for(), do..while() be Boolean expressions.  That would
 >solve the
 >	if ( a = 0 )
 >problem.  But it would break so much existing code that I for one am
 >sure that no such "C" standard [that requires a diagnostic when the
 >Boolean-expression constraint is violated] would be generally accepted. 

Now if only Doug had posted this paragraph 2 or was it 3, months ago, when
the discussion started it would, I think, have suppressed much of the
lunatic fringe (O.K. - include me in that if you will (:-)).  We would not
have had the "this isn't a problem" and "*I* don't do this so how can
*anyone* want to change..." postings.

Doug clearly states:

a) The =/== question was thought serious enough to be considered by X3J11.
b) Various approaches to solve the problem were explored and rejected.

There is a *BIG* difference between saying "not a problem" and "solutions
are difficult or contradict other mandates etc".

I don't fully agree that a switch or pragma enabled boolean extension to the
langauge is not possible to enable *future* code to be more fireproof,
however I'm sure that that has all been fully discussed at X3J11 (:-).

Anyone think we should open a new newsgroup to address exensions to the
language in general, and X3J11 discussions in particular??  The total
traffic here has got out of hand!


Ray Dunn.  ..{philabs, mnetor, musocs}!micomvax!ray