First of all, cool things down a bit. Second, people will continue to do
double-replies, and talking about wit will not help. If you want to
continue discussing mailing list design, this is outside the scope of
Brad Knowles wrote:
At 4:12 PM +0100 2003/10/23, Tom Thomson wrote:
Some lists are designed to
the originator of the message; the design of such a list will be broken
instantly by a decision not to munge the Reply-To header.
Those lists are very uncommon, and are not the audience that is
being addressed by the previously mentioned page. If you want to set up
an anonymizing list, then obviously you would have some additional
requirements that you would need to meet.
the most commonly intended reply action is to send the reply to the list
without a second copy to the originator: as the most common action, it
should be achievable with a single mouse-click or key-stroke
In an ideal world, that would be easily achievable.
This is not an ideal world.
all the MUAs out there will not allow you to do that if the list
doesn't munge the Reply-To header.
So we break something for many users for whom there is no
alternative, simply to make it easier for other users to do the
"default" action, and for whom if they don't want the default action
they're usually also screwed?
This sounds like an extremely brain-damaged idea.
For other lists the most commonly
required reply action is a reply to the list moderator - again, Reply-To
header munging is required to make this the default action of most MUAs.
Again, not the most common type of list, and for which the
previously quoted page is not attempting to address.
Each list has its own requirements, the standardised
RFC2822 is not adequate to cover all these requirements without header
Agreed -- you can't use header munging to conform to the issues and
requirements of all mailing lists, especially not with all MUAs (since
many are severely broken).
Therefore, you should make no attempt to use header munging for this
What I want out of an MUA is to be able to chose between reply to
originator, reply to list, reply to list with copy to originator,
list with copy to cc reipients of the original message, .... quite a few
Get a better MUA. Don't try to fix this problem within the mailing
The headers provided by RFC2822 are not adequate to permit
to offer all these options in a simple and straightforward usable
Sounds like you need to write an RFC to detail all these issues and
lay out the suggestions for MUA authors.
So I think this list would probably be
by adding the munging
That's assuming that no one on the list (currently or in the future)
is/will be using an MUA that is known to be severely broken, in such a
way that they are completely unable to reply to any message other than
what is specified in the "Reply-To:" header.
Not a valid assumption.
- biut I don't think it's worth arguing about;
Then why are we having this argument?
do think is worth arguing about is someone claiming that as a general
principal such munging should not be done: the RFCs make no attenmpt to
dictate to me (or anyone else) what features I should put into
I supply to my users, and I don't like to see Chip claiming they do and
neither do I like to see Brad quoting him as if he were holy writ.
His page hasn't been updated to make reference to the latest RFCs,
but then it doesn't really need to. The issues with regards to the
broken MUAs haven't changed, and many sites are still running the same
ten-year old software that was at the heart of the problem originally,
which lead to Chip writing that page.
If the problem hasn't changed, it doesn't make a whole lot of sense
to try to update the description of the problem, or recommendations on
how to avoid tickling the problem and making it even worse.
I make reference to this page because it's the best collection I've
seen yet for the various problems and the ways to deal with them.
Would you prefer that I write all this up as an RFC and really set
things in stone?
Asrg mailing list