I don't see why there is an apriori requirement to be compatible with
X.400 error reports. A *lot* more people out there run Novell MHS.
The best solution is to determine what approach satisifies the minimal
set of requirements and is the least invasive. Compatibility with other
systems is nice, but hardly a requirement for a successful solution.
Well, OK. Perhaps so. The reason I brought up the existing definitions
in 1148 is because they *exist* now. That is, there's already a wheel
shaped object (it might, however, be a bit oblong... which is the
nature of the OSI beast) so why work on reinventing one? The simplest
path, IMO, is to adopt the existing stuff.
I haven't studied what the requirements are and whether this path is
terribly invasive, etc etc etc. You might well be right, or you might
be overreacting. Either way it is worth some study. Which, of course,
the IETF-ACK group has done ...