ietf-822
[Top] [All Lists]

Re: Understanding response protocols

2004-09-22 07:17:54

Philip Hazel <ph10(_at_)cus(_dot_)cam(_dot_)ac(_dot_)uk> writes:
On Wed, 22 Sep 2004, Arnt Gulbrandsen wrote:

True, but I thought we were discussing a problem that arises when a mailing
list *is* used for a discussion. 

As is happening here ;-) !
Perhaps we could analyse the headers of various resonses and 
see which ones fit the models being discussed.

In that situation, it does become more like
USENET, and users have similar expectations of the two.

One of the USENET concepts which has translated quite well to mail is 
the use of the References: header to thread mails.


Some users do. Not all.

Yes. Agreed. Sorry. I should have said "some users".

Some people want mail to behave like mail, whether a list is involved or not.
Others want mailing lists to behave like usenet. 

Part of the problem is it is not easy to identify when a list is involved.
I (automatically) file incoming mail based on list it comes from 
when I can. Do do so I have to match various "headers":

         $mail->get('X-Mailing-List-Name') ||
         $mail->get('List-ID')  ||
         $mail->get('List-Name')  ||
         $mail->get('X-List')  ||
         $mail->get('X-listname') ||

and I was missing the 'List-ID' used by this list till recently.

If there was a clearly defined header that said "you got this mail from 
the list <whatever(_at_)somewhere>" it would be trivial to add a 
"Reply to list" button to a MUA.

Note that 'List-ID' for this list is List-ID: <ietf-822.imc.org>
with no @ which would make it hard to use if this list 
had been called (say) ietf(_dot_)822(_at_)imc(_dot_)org

Sending _just_ to the list would probably suit most discussion type lists.
List+Author would cover a few more cases where author may not be 
subscribed.


At present, I use "reply to all" for virtually every message to which I 
respond, whether it comes from a mailing list or not, though of course 
it doesn't always do the right thing, and if I remember, I edit the 
recipients manually sometimes when I know people don't want personal 
copies. There are, of course, some exceptions, for which I just use
"reply", but these are pretty rare for me.

Likewise.


On the other side of the fence, I just quietly delete the duplicate
copies I receive because I know it isn't easy for the posters to find 
out that I don't actually want them.

My home grown MUA quietly supresses duplicates, when it can, based 
in Message-Id. So I have a pet hate of those mailers which Cc me 
and a mailing list and use different Message-Id for the two (or more)
copies.

As I use a MUA I wrote it does what I want it to do (provided I can 
be bothered to change it), so I am most interested in defining 
headers that allow "smart" tools to do what their users want.

To this end MFT is a useful addition to the available information,
but getting concensus on how you identify the mailing list in
the To Cc forest is probably more useful.
 


Philip