Hugo Haas <[EMAIL PROTECTED]> wrote on Tue, 27 Jun 2000:
> Hi.
Hello!
> I read a lot about Mail-Followup-To and Reply-To recently (including in
> the mutt-users archive) and my conclusion is that:
> - Mail-Followup-To is not a standard and is supported by very few MUA's.
Well, it's not a standard in the way that it would have been defined in
a RFC. It's also not a standed in the way that there are a lot of
popular MUAs which support it (ie. "industry standard" sort of). It
*is* a standard in a way that there exists specific documentation on the
use of it with definitions. (http://cr.yp.to/proto/replyto.html)
The only MUA I know of which supports MFT is Mutt. Unfortunately.
This means that users of other MUAs should campaign for adding such a
feature to their clients, because it is unlikely to get added otherwise.
:-( Except maybe if it was included in some RFC...?
qmail (a MTA) does support MFT passively -- it can add the header to
outgoing emails, but this only works for messages added locally with
qmail-inject, and it doesn't mean that whatever MUA is used on such
system will then respond to MFTs.
> - Reply-To should be able to do the right thing, even if some
> implementations are forcing people to use this field.
This is not quite true.
Reply-To doesn't specify whether a message reply is "discussion group
followup" or a "private reply" -- there is a distintion between the two,
and this distinction is what causes the problems with Reply-To.
I'm also not aware of whether there is any specified way to have
Reply-To set to more than one address. You can either have multiple
Reply-To headers, one address per header, or you can have multiple
addresses in one header. I think that in either case, the behaviour
of MUAs is unspecified, so you may and will get random results. Does
anyone know more about this?
> Of course, mailing-lists adding a Reply-To header would break that, but
> anyway there is no perfect solution.
There is a perfect solution, which is MFT, it's just that the other MUAs
don't support it. :-) Get them to support it, and the problem goes
away. If they don't start to support it, the problems inherent in
playing with Reply-To will remain. It's that simple.
Most of the (discussion) mailing lists I'm on go through the "should
Reply-To be set to the list or sender?" talk at least once, some go
through it every now and then. This is proof enough that Reply-To
isn't the right solution to mailing list issues, there's always some
problems if you try to make it be the list-discussion indication
method.
Modern MUAs, except for Mutt (that I know of) simple do not deal with
mailing lists intelligently, they are only intended for private email
use. Just like the original mail specifics. I find it rather curious
that nobody (well, the MUA authors and the users of those MUAs) actually
perceives this as a serious problem...
Yes, at the moment it doesn't look like MFT is widely supported.
However, IMHO it's *clearly* the best solution to the problem. Just
because it might be difficult to get the right solution adopted because
of legacy behaviour, that doesn't mean we shouldn't *try* -- if we don't
try, then some things will *never* get fixed, will they?
Oh well, ranting about this on mutt-users is pretty much preaching to
the choir, since Mutt already does support MFT. It's the *other* MUAs
we should be concerned about, and the developers/users of those don't
read mutt-users.
To conclude, yes you shouldn't expect MFT to work yet except on lists
where the majority of people use Mutt as the MUA. That doesn't mean
that you should abandon it though!
Hugo Haas <[EMAIL PROTECTED]> wrote on Thu, 29 Jun 2000:
> I don't think that it's possible to do this with send-hooks (or if it
> is, I'd be very interested to see how to make send-hook interact with
> the From, To, Cc and the value of lists and subscribe), so it would
> require some new code.
You can't have it done automagically, like MFT is done now. You can
match against the current From header with ~f, you can match against
sending it to a subscribed (?) mailing list with ~l, but you can't
construct similar sets of addresses as MFT generation does with .muttrc
commands alone. So yes, it would require some new code.
Regards,
Mikko
--
// Mikko H�nninen, aka. Wizzu // [EMAIL PROTECTED] // http://www.iki.fi/wiz/
// The Corrs list maintainer // net.freak // DALnet IRC operator /
// Interests: roleplaying, Linux, the Net, fantasy & scifi, the Corrs /
Warning: The electrons relaying this email travel at extremely high speed.
