ietf-822
[Top] [All Lists]

Re: Understanding response protocols

2004-09-13 09:13:24


In <20040909221713(_dot_)0eb23aab(_dot_)moore(_at_)cs(_dot_)utk(_dot_)edu> 
Keith Moore <moore(_at_)cs(_dot_)utk(_dot_)edu> writes:

There really is a conflict sometimes between what the sender thinks
should happen in the event of a reply and what the recipient wants to
happen with the reply he composes.  In that circumstance, the recipient
needs to win. 

Yes, but that is the exceptional case. Normally, what the sender asks for
should be what the sender gets. Ultimately, if the replier wants to do
something special, then he can do it, but that is not the commn case and
therefore it should not be what happens if the replier just presses one of
the buttons on offer. If the replier wants to do such special things, then
let him cut and paste.

I think you have it sort of backwards.  The exceptional case is
where the sender asks for something different than "reply all".  The
recipient's MUA should take pains to make the recipient aware that
the sender has requested a different kind of reply, while still giving
the recipient the ability to easily choose between what the sender
asked for, the normal "reply all" behavior, or something else.  
And "cut and paste" is just too cumbersome.

I think we are in agreement that user agents should uniformly
implement two reply functions (along with being able to edit the
reply address list);

- reply to author 
- wide reply (more or less "reply to everyone")

No we are not in such agreement. Firstly, if by "reply to author" you mean
"reply to From", then that is certainly not correct 

it certainly _is_ correct, presuming the author used From correctly.

(a replier that wants
to override the Reply-To should be made to jump through some extra hoop,
just to be usre he knows what he is doing).

no, a replier should be made aware that Reply-To was set so that he can
make the choice of whom to reply to in light of the knowledge that the
sender has requested that he do something unusual.  but Reply-To 
should probably be deprecated in favor of something that specifies
the *complete* recipient list rather than just a replacement for From.

And secondly, there are still some who would like an explicit "reply to
list" - "reply to all" is just too wide and should be reserved for the
cases where "all" is really, really what you want.

I don't really have a problem with that, I just think it's of marginal
utility.

Keith