Wearing my "I am not an implementor" hat:
1) Most people seem to like the idea of an Atom error reporting facility, although some don't because they think it will never get used.
2) It is not clear whether or not this should be Atom-specific. Even the erstwhile W3C liaison to the IETF suggests that maybe the IETF making a generic change to HTTP for all error reporting is OK.
3) The reasons for wanting to reuse POST or invent a new verb are deeply religious, where "religion" is defined as a mixture of "my family has always done it this way", "bad things just might happen if you do it a different way", and "I'll ignore the fact that the other way has worked just fine in different environments". That certainly resembles how religion (and national politics...) works where I'm from.
Wearing my co-chair hat:
Given the vehemence of the arguments of the last day, I propose that consensus on *how* to do error reporting is not possible in the short term. Feel free to argue the "new verb" question ad nauseam (minus the pre-adolescent comments like calling people "mental"), but maybe don't expect your arguments to settle anything right now. And there are other open issues in the work rotation.
--Paul Hoffman, Director --Internet Mail Consortium
