[Mailman-Users] From field displayed by MS Outlook

Richard Barrett r.barrett at openinfo.co.uk
Thu Jan 22 07:52:42 CET 2004

On 22 Jan 2004, at 05:00, Rabinowitz, Ari (Exchange) wrote:

>> From: mailman-users-bounces+arimr=bear.com at python.org
>> [mailto:mailman-users-bounces+arimr=bear.com at python.org]On Behalf Of
>> Brendan Pratt
>> Sent: Wednesday, January 21, 2004 11:49 PM
>>> What is wrong with the mailman authors making their product superior
>>> enough to compensate for Microdummies poor programing and faulty RFC
>>> interpretation?

Absolutely nothing, but note that Outlook is not RFC non-compliant on 
this issue; what is displayed on the applications GUI is not covered by 
an RFC and is a matter between the developer, their conscience, and the 
end user.

> Mailman is doing the right thing, by specifying the Sender address as
> the list bounce address so that bounces go there, and the From: address
> as the author of the message.  Unfortunately, Microsoft in its latest
> versions of Outlook has decided to show both the Sender: and the From:

I am yet to be convinced that Outlook is conflating From: and Sender: 
headers. I suspect it may be conflating the From: and Return-Path: 
content but no matter; the observation about why is spot on.

> addresses in the From: field so that it is easier to notice spam with
> forged From: headers.  The two desires, to have bounces go to a 
> different
> address than replies and to display both the Sender address and the
> From: address conflict.
> There is no way for the Mailman developers to fix this, since if they
> didn't have the Sender address be the list bounce address then all of
> Mailman's bounce handling would no longer work, and the author of the
> message would get flooded with bounces that they don't care about and
> cant do anything about.
> The fix would have to be for Outlook to have an option to not display 
> the
> Sender address in some cases, which is a fix that Microsoft would have 
> to do.
> Sorry, but this isn't something that the Mailman developers can do 
> anything about.
> Ari

In principle I agree with this conclusion but in practice I think it 
may be possible to offset the problem highlighted in the original post.

The post that started this thread said that Outlook is displaying 
something like this in the "From" field of its GUI:

<mailto:listname-bounces at mailman.server.com> 
listname-bounces at mailman.server.com On behalf of 
<mailto:name at poster.domain.com> name at poster.domain.com

I suspect that what confuses/upsets/deters/induces-fear-in some Outlook 
users is the appearance of the string "bounces" in what is displayed; 
bouncing may have connotations for Outlook users which are not apparent 
to ordinary people and hence their concerns.

But suppose that what they saw displayed was instead:

<mailto:listname-sender at mailman.server.com> 
listname-sender at mailman.server.com On behalf of 
<mailto:name at poster.domain.com> name at poster.domain.com

That is to suppose that Mailman used an alias listname-sender (or some 
such) instead of listname-bounces as the return path for its bounce 
handling. After all, MM 2.0 used listname-admin and I do not remember 
this causing such angst as listname-bounces has since MM 2.1 arrived.

Would this defuse the objections being registered by Outlook users?

If so, then what I said to Jon Stethridge in a private communication, 
before he set this thread running, applies:


The best that might be otherwise be achievable would be a change in the 
alias used as the envelope sender from <listname>-bounces to something 
more palatable when it is displayed by Outlook. This would require a 
thorough trawl of the Mailman code to make sure this change did not 
lead to any undesirable side-effects.

Thus far nobody has volunteered for the task and, without a 
confirmation from Mailman's owner Barry Warsaw that he will fold it 
into the Mailman source tree, I doubt anyone is likely to commit the 
time and effort to implement and test the solution; there is also the 
issue of upgrading existing installations and the acceptability of that 
task to the site admins.

If this list's membership forged a consensus behind a manageable and 
relatively minor proposal to change Mailman, which did not compromise 
its adherence to RFCs, then maybe there is a way forward to suit most 

Richard Barrett                               http://www.openinfo.co.uk

More information about the Mailman-Users mailing list