[Mailman-Developers] Reply-To: handling

J C Lawrence claw@kanga.nu
Fri, 19 Oct 2001 22:11:49 -0700

On Fri, 19 Oct 2001 21:30:30 -0700 
Chuq Von Rospach <chuqui@plaidworks.com> wrote:
> On 10/19/01 8:50 PM, "J C Lawrence" <claw@2wire.com> wrote:

>> I agree, except that as a moderator who regularly moves threads
>> about, having the abilty to coerce, or minimally force initiation
>> of the thread move as a Good and Useful Thing.

> Fair 'nuff. Stylistic differences, neither is right or wrong. I do
> it differently.


>>> What about the case where a list is not coerced reply-to, but
>>> one fo the subscribers feels it should be, so he coerces
>>> reply-to covertly, which is propogated out and through the list.

>> Which is actually orthogonal to the case under discussion and is
>> thus a red herring.

> I'm not sure I agree, but I won't push the issue. I think the MLM
> should be consistent to the end-user, which means it does the same
> thing based on its configuration, no matter what the original
> poster sets in their headers.  That means headers like reply-to
> much be stripped before processing, not carried through, or you
> hae inconsistent variations of those headers.

Two cases:  

  1) Reply-To munging off:
       -- Users can fully control post disposition, which may not
       involve the list at all, or may be the list (thus preventing
       simple private replies).  They can control whether simple
       replies go to the list, don't go to the list, go only to the
       list, etc.

  2) Reply-To munging on:

       -- With Reply-To removal:

         All replies go back to the list ONLY.  There is no way for
         simple replies to go anywhere else.

       -- With Reply-To extension:
         All posts go back to the list and may ALSO go an address
         the previous poster specified.  In both cases simple
         replies are guaranteed to go the list.

I'm of the general mind that the two cases (munging on and off) need
separate handling.

>> What you seem to be asking for is reply-to stripping for lists
>> that don't munge reply-to.  I can see the reasoning, but also see
>> considerable danger/pain in that direction.

> Exactly. And I don't see danger/pain, I see consistency of
> operations.


The pain bits are the standard/good uses of Reply-To as a dupe
prevention technique, use as an address re-writing, use for thread
redirection, etc.  Depending on your lists and their demographics,
these are variously useful/useless.

However, conflating the problems on non-reply-to munging lists and
the problems of reply-to munging lists seems a mistake.  They are
distinct problems with different problems related to feature/header

I see value in two options for NON-reply-to munging lists:

  1) Strip all poster-set Reply-To headers.

  2) Strip all poster-set Reply-To headers which point at the list.

I don't think I'd use them, but then my lists have very different
public to your lists Chuq.

> Which, I guess, makes me the hobgoblin of small mindedness.....

Hardly.  You're arguing for simplicity and transparency of interface
for the end-user, and we all well know, end-users are hob goblins.
I'm hardly going to argue with that.

J C Lawrence
---------(*)                Satan, oscillate my metallic sonatas.
claw@kanga.nu               He lived as a devil, eh?
http://www.kanga.nu/~claw/  Evil is a name of a foeman, as I live.