Let's try an example.
Mail-Followup-To has found favor with the community. It saves time for
users. There are several independent interoperable implementations.
Mail-Followup-To is widely implemented, which I find unfortunate, because
it makes the reply problem even worse than it already was.
In the IETF, however, publication of an experimental Mail-Followup-To
spec was blocked by a small group of people who refuse to admit that
mail clients have separate reply and followup features.
This is not an accurate characterization of what happened.
Presumably the
same people would also be in charge of an IETF field-name registry.
That the same people would be in charge now as then seems unlikely to
me, because people tend to rotate out of positions of responsibility
in a few years' time. That doesn't mean the result would be different.
Am I saying that Mail-Followup-To should be standardized without review?
No. I'm simply saying that the name should be reserved, so that there's
no risk of the Mail-Followup-To deployment bumping into something else.
I personally think Mail-Followup-To is a Bad Idea. But I also think
that this specific field is worth documenting in a registry - because
there are enough implementations that support it that people need to
know what it is, and so that nobody else will choose the same name for
a different purpose.
I wouldn't object to registering Mail-Followup-To *provided* the
registration process required community review.
Keith