It's important to understand that the author sets followups. Let me
emphasize once again that reply-to munging is harmful; the question is
how to fix the message duplication _without_ reply-to munging.
In light of this understanding, let's review Keith's complaints:
+ The author of the subject message isn't always on the mailing list,
in which case he misses the replies/followups.
No, he doesn't miss anything, because he puts his own address into the
followup field along with the list address.
Of course, he doesn't have to bother with a followup field in this case,
since the default followup targets will work properly.
If he _is_ on the list, he should supply a followup field without his
address. His MUA can do this automatically if it knows that he's on the
list; some MUAs already support the necessary database.
+ A single message can be posted to multiple mailing lists, in which
case followups should arguably go to all lists.
So the user puts all the lists into the followup field.
MUAs configured as above will do the right thing automatically.
When replies
from my list postings go to moore(_at_)cs(_dot_)utk(_dot_)edu, I can respond
more
quickly.
So leave your address in the followup field. You have control.
(Btw, why is ietf-822 so slow? It doesn't look good for a mailing list
to take half an hour to distribute a message---especially when it's a
mailing list for mail implementors!)
And a followup-to like header is certainly not "the" only available
solution. For instance, UAs could suppress duplicate messages.
Suppressing duplicates is fine, if you can afford to do it securely, but
it solves only half the problem.
After a long sequence of followups---let's say a sequence of 500
messages from different contributors---there are, with current defaults,
501 recipient addresses in the header. Software breaks. Humans can no
longer read the header. Each new message takes ten times as much archive
space as it should. Every contributor is at risk of being a permanent
recipient of a neverending discussion; there is no escape.
Of course, people don't let the header grow so much. They trim it down
to just the list address. But this introduces some of the same problems
as reply-to munging; for example, it destroys cross-posted discussions.
With a followup field, the solution is trivial: the MUA copies the
message's followup field into subsequent followups by default.
Even if you only have reply, sometimes you want to reply to author
only, sometimes to author and all recipients, sometimes to author and
particular recipients, etc.
So what? Are you going to argue that we shouldn't have Reply-To since
people sometimes want followups instead?
Add a followup-to
(wide-reply-to, group-reply-to whatever) field to the mix and it just
makes things more confusing.
Explanation, please. Who exactly would be confused?
Users will have a button for replies, just as they do now.
Users will have a button for followups, just as they do now.
They can control replies with Reply-To, just as they do now.
They can control followups with the new field. This is much simpler than
the current situation.
What we haven't had is
agreement that such proposals, if adopted, would do more good than
harm.
Explanation, please. How exactly would a followup field hurt anyone?
We've already had proposals for a name.
I know. I don't think ``Wide-Reply-To'' will work. As you can easily
verify, ``wide reply'' is meaningless for a typical user. (``Does that
have something to do with the window size?'') Even when the MUA has a
``wide reply'' feature, most users have no clue what it does.
``Group reply'' is somewhat better (and apparently more common in MUAs),
but still not immediately clear to most users.
Any other suggestions?
---Dan
Let your users manage their own mailing lists. http://pobox.com/~djb/qmail.html