Jutta Degener (jutta@pobox.com)
Tue, 12 Jun 2001 14:54:56 -0700
[Greg V. wants to make DSN/MDN a Message-Context]
> My initial thought was that DSN is orthogonal to context, that is, a message
> may be both a failure and a pager message, but upon further reflection, I
> believe a failure is a context in and of itself, sent in response to a
> message of any context, and treated primarily as a failure, not as a message
> in the original context.
My knee-jerk reaction is to disagree; I think that on an abstract application
level, a failure is primarily an attribute of the message that failed. (It is
not a "fax message" because a fax message failed, but it is the negative
confirmation coming out of your fax machine, or a voice announcement after
the voice mail you sent bounced, or ...)
Two other arguments against:
* The current draft begins:
> This document describes a mechanism to allow senders of an Internet
> mail message to convey the message's contextual information. Taking
> account of this information, the receiving user agent (UA) can make
> decisions that improve message presentation for the user in the
> context the sender and receiver expects.
MDNs and DSNs are usually not sent by a human sender. There are no
expectations. There's just an e-mail system that can't deliver an
email message.
* MDNs and DSNs are already completely and reliably distinguishable by
Content-Type. This mechanism would be at best superfluous.
If there is a need to change things, I would prefer to change the draft
to more strongly point out that the Context's aim is _not_ to tag _all_
possible e-mail messages with the expectations of their human sender,
but to merely express those expectations where they obviously exist.
Jutta
This archive was generated by hypermail 2.0b3 on Wed Jun 13 2001 - 00:53:50 IDT