Two points are missing:
1) Undue cost for making a *bis revision
Strongly agree with this. It's already too difficult for IETF to revise
documents. There is too much tendency for IESG to insist on rewrites
that raise this cost far out of proportion to the value gained in
exchange for the extra work and the delay - and often, IMHO, the
rewrites actually harm the documents.
We really need to be making it easier to revise documents in IETF so
that we can do a better job of keeping them current. It's one thing for
IESG to hold up a revision because of some newly-discovered bug that
harms interoperability or clarity of the document, but it seems like a
waste of precious resources to hold up a revision merely to insist on
some new editorial convention.
2) Protecting the installed base.
agree with this also, for reasons stated in a separate message.