Re: Understanding response protocols

["Charles Lindsey" <[e-mail]>]

In <[e-mail]> Bruce Lilly <[e-mail]> writes:

>Charles Lindsey wrote:
>> 
>>>And MTAs.
>> 
>> 
>> Why MTAs? MSAs perhaps.

>Some MTAs rewrite addresses. For example, a border MTA may rewrite
>[e-mail] as [e-mail] in order to hide internal host names.
>If such addresses are rewritten in the envelope and in standard
>address fields, then any new address field that is intended to be
>used by recipients should be similarly processed.

The proper solution for that is for the author to configure his MUA to
always use [e-mail] as his From/Reply-To address (most MUAs provide
that config option, which would probably also be automatically used within
MFT is the agent was smart enough to be MFT aware.

But I agree that the sysadmin in charge to the border might also configure
his MTA to do that rewrite (it is indeed commonly done), in which case he
would have to apply it to all the various headers we are talking about. I
have no doubt sendmail.cf could be made to do it.

>>>>Looks like we have to choose some least-harmful
>>>>alternative. Reply-To is what we have at the moment, and it is clearly not
>>>>working.
>> 
>> 
>>>It works fine for me...
>> 
>> 
>> But not for anybody else, apparently.

>It seems to work fine for others that use it also.

Nobody on this list is using it except for you.

>> We are discussing a suggestion that it might be introduced as a solution
>> to the problem. You do not need ABNF and a full draft (though they would
>> be required eventually). But, if you insist, that you may consult
>> ftp://ftp.dsv.su.se/users/jpalme/draft-ietf-drums-mail-followup-to-00b.txt.
>> Oddly, that draft does not actually specify any ABNF, but you may assume
>> it intended syntax similar to the Reply-To field in RFC 2822.

>Actually there is a syntax specification (however it omits the
>colon which delimits field name from field body...).  You also
>seem to be very confused about address fields; you claim that
>the syntax in Jacob's 5-year-old draft is intended to be like
>that of the Reply-To field, but Jacob's draft specifies a
>mailbox-list whereas Reply-To uses an address list.  That
>draft would make wide responses difficult at best, since it
>deprecates such wide responses.  There are provisions that are
>impractical for implementation.

Yes, that draft was written in the middle of DRUMS, before the final form
of RFC 2822 was fixed. If we wanted to follow that route, then there are
several things in that draft I would like to change.

But its most interesting feature was that it took it for granted that an
MUA would know when it was doing a "group reply", and that implies that
some form of "Reply-to-List" button was envisaged.

-- 
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: [phone] Fax: [phone]   Web: http://www.cs.man.ac.uk/~chl
Email: [e-mail]      Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9      Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5