Peter,
Peter J. Holzer wrote:
* Envelope:
I find the usage of the term envelope for the combination of the SMTP
envelope and part of the header very unusual and somewhat confusing,
despite the explanation in 4.1.1. (especially since it is first used
in 2.2.3). I also don't really see a reason for it: Why not keep
envelope and header separate and merely note that MTAs may add some
information to the header?
The header contains two very different types of information. One is between
the originator and the recipient, such as TO and Subject and Date. The other
type is part of the handling information associated with message transfer.
If the latter is not part of the envelope, then what is? Entries such as
Return-path underscore this class of information within the header. As do
Received header fields.
In other words, the "header" is a syntactic construct within the message, but
it contains information that divides into two broad semantic categories.
* 3.3.1 Message-ID:
"it is associated with the RFC2822.From field, ..." How so? I don't
see any conceptual association between the From and the Message-ID
field: While some MUAs may use (parts of) the From field in generating
the Message-ID, others ignore it.
The discussion about Message-ID attempts to characterize it as distinguishing
among different messages. The difference among messages is primarily due to
choices made by the authors. That's the From field.
Or, at least, that's the theory behind the statement...
* 4.1 Message Data:
"Message Filtering (SIEVE)". It might be worth pointing out that while
SIEVE is the only filtering language specified in an RFC, other such
languages are in use.
A similar qualifier would apply to virtually every mechanism described in the
document. While true, I'm not sure how helpful it would be. For simplicity
and focus, the paper limits itself to discussing standards-based mechanisms
that have gained substantial use in the open Internet.
* 4.1.4 Identity references in a message:
RFC2822.BCC: "An MUA will typically make separate postings for TO and
CC recpients, versus BCC recipients". Is this really typical? In my
experience, the message is normally only posted once, with the BCC
header removed, and all recipients in the envelope.
I have not tracked the latest dominant behavior. I know that MUAs have a long
history of using a couple of different schemes. If folks tell me to remove
the "typically", I'm glad to do it. (I might do it anyhow, but at this point,
I'd rather not make any change that isn't due to consensus pressure or
identifying a plain error.
* 5.1 Aliasing:
"It resubmits the message, replacing the envelope address, on behalf
of the mailbox address that was listed in the envelope". It would be
clearer if the sentence read "... replacing the envelope recipient
address ...".
ack.
* 5.3 Mailing Lists
"... can be viewed as an elaboration of the Re-Director role"
Since the previous section is called "Re-Sending" and not
"Re-Directing", that probably should be "... of the Re-Sender role".
ack. I'll review the text.
Thanks!
d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net