At 6:14 PM -0500 10/31/04, Robert Sayre wrote:
Close PaceErrVerb. It's a nice idea that will never get done in this WG, if it's even possible at all.
Agree. If it should be done, it should be done in the W3C, and I suspect that won't happen soon.
PaceServiceError could work.
Agree.
However, we have no specified notion of Errors or Conformance, so it's tough to tell whether the proposal would be helpful.
We do have enough of a concept of errors in the format, and certainly XML has lots of specific ways to be non-compliant.
If we come up with a status quo position on conformance, then I would definitely want it in 1.0. A Dead Letter Office is better than nothing. If we decide on something better, then it could be something added later if required.
I'm more strongly in favor than that. The IETF has a long history of bad implementations of format protocols where we end up just sighing and saying, "man, there's a lot of bad programmers out there". It doesn't have to be that way. When I ran big email interop events, folks from Company A would walk over to folks from Company B with a trace of what was just sent and say "I'm sorry, but that's crap". Often, within a few hours, the folks from Company B would say "OK, let's see if this is better" and they would have fixed the bug. (Yes, they brought source code and compilers to the events; it was pretty cool.)
I fully believe that if there is an error-reporting facility, even if it is turned on only 10% of the time, it will be of great benefit to the entire Atom community. It will cause us to sigh less in five years. It will help the good developers fix their problems early, and it will help make it clearer who the good and bad developers are.
It's worth the effort.
--Paul Hoffman, Director --Internet Mail Consortium
