On Wed May 25 2005 22:33, Yakov Shafranovich wrote:
In this specific case, additional registered fields need to specify the
type of feedback they are used for.
That's a design issue which should probably be revisited, particularly
in light of review comments regarding "other" and "any".
By using BCP 90, there isn't
sufficient space to do that. Same goes for the DSN/MDN/MTSN fields which
need more than simple registration.
For DSN, MDN, MTSN fields, the BCP 90 registration "applicable protocol"
field could specify "DSN", "MDN", or "MTSN" as appropriate.
Any particular reason why these fields were not included in the initial
header registry? Sounds rather strange to me.
Probably nobody got around to it yet; some message header fields aren't
yet registered and HTTP fields haven't yet been registered.
That's not it at all. The issue was and is that there are already separate
registries for DSN and MDN report field extensions. There was no need for them
to use the BCP 90 registry. The IESG flatly rejected the attempt to build a
single, all-inclusive registry, and opted instead for a registry for certain
specific protocols that didn't already have then plus allowing the option for
future protocols to use the registry if they so desired.
IMO this work would be better off with a separate registry that doens't
conflate these fields with header fields in other protocols. But that's
just my position - you could try and use the BCP 90 header registry, keeping
in mind that there may be pushback since these are not, properly speaking,
header fields.
Ned