I would think that the issue is 'is there consensus to remove
something that is currently in the spec' - i.e., the removal
of an established feature needs strong support, the default is
to keep it.
(note that the above observation is not a comment on the
specific question, it is comment on proper process in the IETF)
As I was reading these documents again, it is not clear to me that they
cover the situation.
Trying to resume Tony's appeal in one sentence, I would say "The subject
of the so-called consensus was not clear, therefore the so-called
consensus shall be voided".
I failed to find text that would define what could be appropriate matter
for a consensus call and what could not.
For example, it is common practice to call for consensus on IDs or on
what to do with an ID. In a meeting, it is common practice to call for
consensus on a slide being presented on the screen or on an agenda item.
Does it appear to you that a clear definition of what is appropriate
matter for a consensus call could be an addition to a revised RFC 2418?