* Fix the section 2 intro ("this section provides") (Paul Hoffman).

* CD bit language.  Sort out what Wouter and MSJ were suggesting.  I
  think it is something like: Validators MUST copy the CD bit from
  requests to upstream queries and validators with trust anchors that
  cover a query (and could thus be expected to be able to validate the
  response) should/must set the CD on the upstream query.

Possible inclusion into the draft, but which needs more discussion:

* Clarification of the NSEC type bit map encoding.  Stuart Cheshire
  (and someone else, IIRC) stumbled on the language around the network
  order of the bitmap.

* Alfred posted a detailed note on dnssec-bis-updates-08:
  http://ops.ietf.org/lists/namedroppers/namedroppers.2009/msg00080.html

** Alfred points out the 4035 is inconsistent when discussion RRSIG
  "RRsets".  First it states that RRSIG records don't form RRsets, but
  then later talks about RRsets.

** Alfred also talks about 4035 referring normatively to RFCs that it
  Obsoletes.  He isn't suggesting that such a correction make it into
  bis-updates, however.

** Alfred suggests clarifying the special inclusion rules described in
   4035 section 3.1.