ietf-822
[Top] [All Lists]

Re: Understanding response protocols

2004-09-14 09:42:48

D. J. Bernstein wrote:

Specifically: A huge number of today's messages have a suboptimal reply
address in From, the desired reply address in Reply-To,

As suggested by Keith, and as discussed in my earlier message, the
proposal under discussion is based on individual responses going
to the mailbox(es) specified in the From field, when the respondent
specifically wishes to direct a response to the original messages
author(s).  Default responses would go to the addresses specified
in Reply-To, whether that is because the author prefers that set
of addresses because of some "optimization" issue, because of
delegation, because he feels that default responses should go to
one or more list exploders, etc. -- and if he wishes to indicate
a reason, he can use a standard Comments field, or he can embed
the addresses in a named group with a descriptive phrase, or he
can use a parenthesized comment (although the latter are
deprecated in standard address fields).

and the desired
followup addresses in To etc.

No. The To field indicates original recipients, not "followup"
addresses.

 What happens when these messages are
interpreted according to your protocol? Answer: not only do replies go
to a suboptimal address, but also followups are completely misdirected.

Wrong. Default responses go to wherever the original message
sender set Reply-To to point -- and as noted above that could
be an "optimized" mailbox, a list expander, a delegated person
or group of people, etc., depending on what the original message
author has determined to be appropriate for the case at hand (and
Keith has noted that the non-standard, undefined "Mail-Followup-To"
isn't appropriate for all circumstances, or for all authors [in
addition to the fact that it is not supported by most UAs and
MTAs]).

``But the desired reply address could go into From, leaving Reply-To
empty!'' you scream. Yes, it _could_ go into From, but in _reality_ a
huge number of today's messages have it in Reply-To.

I'm not screaming.  And yes, the premise is that the mailbox(es)
for direct responses to the message authors (as distinct from
default responses) would be placed in the original message's From
field.

Moving it means
changing and redeploying a huge number of programs, including homegrown
mail scripts that haven't been touched for years.

Whereas none of any of the other proposals to date would require
any changes whatsoever, right? :-)

Only _after_ that's
done would it be safe to start using the name Reply-To for something
else---in particular, for MUAs to switch from To+Cc+(Reply-To||From) to
Reply-To||(To+Cc+From) for followups.

No, default responses would go to Reply-To? Reply-To : From, which
is supported by essentially all UAs.  Wide responses that cover
original message recipients, which as several participants in this
discussion have noted would involve editing by the respondent,
could be initialized via a "reply-all" function as now, and
(Reply-To? Reply-to : From) + To + Cc -- as now used by most MUAs --
is not unreasonable; Reply-To + From + To + Cc might be a tad
better.  Responses directly to the original message authors (From
field) are already supported by several MUAs and is a feature of
several proposals discussed here.

I'm not saying that your protocol is inherently broken. I'm just saying
that it's disastrously incompatible with current practice.

I believe you're wrong about that. You certainly haven't presented
any convincing argument to support that assertion.

For comparison: Mail-Followup-To avoids all the costs of that first
transition step.

Nonsense. First it has to be properly defined. Then the existing
standards that specify use of Reply-To? Reply-To : From would
have to be replaced. MUAs would have to have support added, not
only to use the field for some sorts of responses, but also simply
to display the field (otherwise, an unsuspecting user will be
confounded when a response goes to a place which isn't visibly
specified anywhere). MTAs that modify address fields would have
to be modified.

It's safe for MUAs to switch to Mail-Followup-To||...
for followups; many MUAs have done this already,

Many more (i.e. most) have not done so.

saving time for their
users.

Spare us the marketing BS.

The use of a _new name_ makes the Mail-Followup-To protocol much
less expensive to deploy than the protocol you're advocating.

No, as noted above it requires many changes -- to existing protocol
standards, to all MUAs, to MTAs, etc.

P.S. I pointed all this out, although not in as much detail, in my first
``Understanding response protocols'' message. This discussion would make
more progress if you started paying attention.

That was of the form "I want...", and didn't address the issues
brought up in recent discussion.