FWIW my overall take on this is in line with Dave's.
I also think EKR and Larry made good points.
Basically, I'll just go with whatever xml2rfc produces
at any given moment.
Lastly, I can't really find the energy to get exercised
by any of this, and wonder if interminable discussions
like this are part of the reason folks keep away from all
the well meaning attempts to revise IETF processes.
Ietf mailing list