Eric.Burger@centigram.com
Tue, 11 Jan 2000 18:23:04 -0500
Thanks Graham and Keith for your comments.
You both hit on what was the hardest part for me to work out. I dithered
back and forth, about whether partial-delivery was a special case of DSN or
something else. In fact, my first write-up was simply an extension of DSN.
I'm not wed to it being something new. In fact, one of the reasons I tried
to make partial-delivery an extension of DSN is that I expect a simple
extension would have a much easier time of getting IETF approval than a new
message type.
The big hint that PNDN's are an extension of DSN is that 70% of the PNDN
draft is cut-and-paste from DSN (including the ABNF bug -- good catch
Graham!).
However, the following issue made me fall into the camp that PNDN's are
something different. DSN's deal with whole messages only. There is no
mechanism for identifying parts of a message that fail yet having other
parts being delivered. Either the system delivered the message or it
didn't. If a sending system has a human reading the DSN, then we have some
flexibility with extending DSN. However, we could have a sending system
that parses the DSN. However, it doesn't understand the difference between
an RFC 1894 DSN (whole message failed) and a PNDN (part of the message
succeeded). In this case, aren't we in the same situation as a system that
doesn't understand PNDN?
Keith, I'll defer to you as the DSN expert. Is there an easy way of
extending DSN without breaking it?
Graham: I've incorporated all of your "nits". Thanks for the close
reading. The only thing I saw that was up for discussion is the change of
final-recipient from MUST to MAY. I think (again, I'll defer to Keith,
since he was there) that final-recipient was mandatory because making it
available would greatly help the sender determine if the failure was an
addressing error. My position on this is the security issue of hiding
telco resources is more important (and more likely) than the end user
debugging MX records. If you think I'm being too paranoid, I can take out
this provision. Otherwise, I can put in a NOTE describing why I made the
change.
Thanks again for your review.
-- - Eric
This archive was generated by hypermail 2.0b3 on Wed Jan 12 2000 - 09:31:03 IST