Where should the type information be 155

John,

An even cleaner solution would be to simply declare the syntaxes that are now guaranteed to make C compilers produce garbage code to at least produce a stern "warning" diagnostic indicating that the code produced is almost certainly erroneous garbage. By making this a stern "warning", legacy code (which I strongly doubt even exists) could still be run without modification. With an additional switch to force reasonable operator precedence, this would seem to solve ALL problems, though this same proposal was rejected by the C committee when it was first made, apparently only because it wasn't already included in the 100,000 erroneously preprinted copies of the draft standard.

BTW, There are a LOT of things wrong with the standards processes, and the case with C should be a learning laboratory for anyone who wonders why current standards are *SO* bad. C is certainly NOT alone in being forcibly screwed up by those in charge. A few lessons I have learned include:

1. Standards MUST include a section detailing known bugs, weaknesses, and tradeoffs. No standard should ever be accepted where this section is blank or nonexistent. Any new weaknesses not included in this list should re-open the standard for re-consideration. Certainly, the lack of this process is directly responsible for IEEE-754's many present unaddressed weaknesses.

2. Standards committee members should be personally responsible-liable for any willful wrongdoing. Of course, enforcing this would deprive many committee members of everything they own.

3. Committee leaders should be PAID and not working for special interest parties. Of course, they are already paid by the special interests that they represent. I would MUCH rather have impartial agents making the decisions.

4. Documents challenging a standard should be preserved. Willfully destroying them (as was done in the case of the C standard) should consbreastute prima facie evidence of fraud by those involved.

Note that at least in theory it might be possible to repair C if only well-motivated people were put in charge. This is NOT the case with the Internet RFC substandard standards process, where there is not only no one at the helm, but there is no longer even a helm to man! The only hope now seems to be to overlay an entirely new structure.

ANSI to its credit forced the C committee to consider our objections to the standard. However, ANSI has REALLY dropped the ball by failing to forge a standards process that works to do anything but what the special interests want.

Steve Richfie1d