Hi Ted, > It looks like section 4 has one or more examples in mind--can you give > example registries where this sort of allocation has been or might be > required via this stream? Sure. Random example... Consider the "LMP Object Class name space and Class type (C-Type)" registry at https://www.iana.org/assignments/lmp-parameters/lmp-parameters.xhtml#lmp-parameters-4 New assignments of Class Names can be made by Standards Action (0-127) or Expert Review (128-247). Suppose an Independent Submission asks for a code point in the 128-247 range, and suppose that the DE thinks this is OK. A value is assigned, and everyone is happy. But you'll also notice on the registry page that each Class Name code point has an associated sub-registry to track the Class Types for the Class Name. Thus, the new code point would also need a new sub-registry. I must admit that I find this corner case (and, by the way, there are plenty of examples of similar cases) to be a bit squirrelly. I would prefer if it didn't need to happen, but I think there may turn out to be cases where it does have to happen, and we have to describe it. Now, it *could* be that these cases don't actually need the creation of a new sub-registry. It could be that the code points for these cases could be contained in the RFC and that would be that. However, this would, IMHO, create a discrepancy in the registries that could cause confusion. Does that help? Thanks, Adrian -- Adrian Farrel (ISE), rfc-ise@xxxxxxxxxxxxxx
