On Fri, 17 Oct 2003, Bob Braden wrote:
> Unfortunately that seems to fall outside what we would consider to be
> an erratum. It is a substantive change to the document. If it's
> important, you could publish 1-page containing it. There is no
> rule that RFCs have to be longer than 100 pages ;-)
Hi Bob -
I don't believe that this is a substantive change to the document, since
it has always been a requirement that the mailbox state not be ambiguous.
See, for example, section 5.5.
Unfortunately, it is sometimes non-obvious when ambiguity problems come
up. Section 5.5 was added to the specification to deal with one set of
ambiguity problems.
The situation in which recent messages are expunged but the server fails
to update the recent count is another type of ambiguity. I discovered
that a commercial server had this problem (the UW and Cyrus servers do
not). When the vendor's attention was called to the problem, they
immediately agreed that it should be fixed.
However, I also determined that RFC 3501's text for the EXPUNGE command
and response did not warn about this situation. This is definitely a flaw
in RFC 3501.
I believe that this qualifies as an erratum, since it clarifies an aspect
of the protocol that is otherwise non-obvious but does not change the
protocol. It is possible that my current erratum text suggests more than
just an erratum, and needs to be revised. If so, can you suggest a
change?
Please let me know whether we can get this and the other errata in
ftp://ftp.cac.washington.edu/mail/rfc3501-errata
as an rfc-editor.org errata for 3501, or if the ruling on appeal is still
that the 1-page RFC is needed.
Thanks!
-- Mark --
http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.
