<< Previous Message Main Index Next Message >>
<< Previous Message in Thread This Month Next Message in Thread >>
Date   : Fri, 23 Feb 2007 06:31:32 +0000 (GMT)
From   : debounce@... (Greg Cook)
Subject: Mailing list headers

On 23 Feb 2007 00:27:43 +0000, Jonathan Graham Harston
<jgh@...> wrote:

> >Message-ID: <45DAF8F0.7050701@...>
>  
> Andrew Benham <adsb@...> wrote:
> > On your email client, you would hit the "reply to sender"
> > button to send a reply to just me, and the "reply to all
> > recipients" button to send a reply to everyone.
>  
> Hold on... people keep mentioning "reply to sender" and "reply to
> all recipients" buttons. I only have "reply" which prepares a
> message to the reply-to address if present, or the from address if
> reply-to="", starting "{poster in from header} wrote:"
>  
> What is the "reply to /all/ recipients" supposed to do? The
> email client receives "From: fred", "Reply-to: jim". What does
> reply to /all/ do? Reply to /both/ the from and the reply-to?
> Wouldn't that be reply to /both/? Where does the /all/ come from,
> and how on earth does the email client know who else the /server/
> has sent copies of the message to?

Okay.

From: Fred
Reply-To: Jim
To: Sheila, Hazel
Cc: Arthur

If Sheila presses "Reply to Sender" she will send to Jim (the address
to which Fred wants replies sent; RFC 2822, s. 3.6.2, 3rd para.) 
"Reply to All" will send to Jim, CC-ing Hazel and Arthur.

If Hazel is a non-munging mailing list and Fred and Eros are
subscribers, they both receive the original, and by pressing "Reply to
All", Eros will send To: Jim;  Cc: Sheila, Arthur, Hazel. Hazel then
delivers a blind copy to Fred and Eros himself.  Without support for
RFC 2919, Eros's mailer cannot know that Hazel is a list, that Fred is
a member and will get the reply, or therefore that he would have a
duplicate if the Reply-To: address were also Fred.

With RFC 2919 compliance, Eros's mailer can deduce that Hazel is a
list, and that Eros is a member.  Assuming that only members can post,
it can infer that Fred subscribes too.  But again it could not tell if
the other recipients were members, or lists themselves.

> > If the list was configured with a "Reply-To:" header with
> > the list's address, then you would use "reply to sender"
> > to reply to all recipients, and "reply to all recipients"
>  
> No.... I would use "reply" to reply to where the message came
> from, ie the list, and I would use "reply to author (only)" to
> reply to the person who originally sent the message to the list.

A potentially useful option, I'll admit.

> Programatically, "reply" sends to [[if (reply-to="") {from} else
> {reply-to}]] and "reply to author" sends to {from}.
>  
> > You no longer have a button to just reply to the originator
> > of the message, so you've lost useful functionality.
>  
> "reply to author"...?
>  
> > And if you put the list's address in a "Reply-To:" header,
> > what do you do if the original message had its own "Reply-To:"
> > header ?  Discard that 'Reply-To', so now you really can't
>  
> Yes. After all, it's a post to a list, not a post to a person. 

Ah. Then munging makes the list awkward in a mixed environment of
personal and list addresses.  

> If
> it so wishes, the /server/ can track /posters/' reply-to address
> as the address that that poster should be send posts to, but that
> sort of information is better managed as part of the subscriber
> information.

Supposing Fred sends:

From: Fred
Reply-To: Jim
To: Sheila, Hazel
Cc: Arthur

Now Hazel is a munging list, and Fred and Eros get:

From: Fred
Reply-To: Hazel
To: Sheila, Hazel
Cc: Arthur

Their "Reply" buttons now default to Hazel (while everyone else's has
Jim), and "Reply to All" goes to Hazel, with carbon-copies to Sheila
and Arthur.  "Reply to Author" would go directly to Fred, which while
probably okay, was not Fred's preference on this occasion.

In other words, the list now interferes with *Fred's* capture of
responses.  (Thread-based Reply-To tracking, whereby Jim is Bcc'ed into
just this thread, would solve this but be an extra burden on the
server.)

> >     http://www.unicom.com/pw/reply-to-harmful.html
>  
> That talks about reply-to munging consider harmful. I agree.
> Munging the reply-to so that it does not point back to the list is
> harmful.

To mung means to alter.  Setting something in the first place is not
munging it.  So if you mean a server altering the Reply-To value to
other than its intake address, then I agree.  A list maintainer would 
be careless to mis-set this.  Is it harmful if the poster does not
point Reply-To to the list, or if the server does not mung it? It
depends on the content of the post.

Again, there's the issue of mailing-list-as-shared-address-book versus
mailing-list-as-forum.  While the above was written according to the
former, opinion here and elsewhere favours the forum approach, and I
would go along with that, but with reservations that the quality may
suffer.

Greg Cook
debounce@...
http://homepages.tesco.net/~rainstorm/



       
       
               
___________________________________________________________ 
New Yahoo! Mail is the ultimate force in competitive emailing. Find out more
at the Yahoo! Mail Championships. Plus: play games and win prizes. 
http://uk.rd.yahoo.com/evt=44106/*http://mail.yahoo.net/uk 
<< Previous Message Main Index Next Message >>
<< Previous Message in Thread This Month Next Message in Thread >>