Errors-to: owner-drums@CS.UTK.EDU
Message-Id: <199508252150.RAA25455@CS.UTK.EDU>
To: drums@CS.UTK.EDU
From: "Roger Fajman" <RAF@CU.NIH.GOV>
Date:     Fri, 25 Aug 1995  17:49:53 EDT
Subject:  Re:  From the Chair:  the Reply-To issue

> I want to close out this issue within the next few days.
>
> My estimate of what we have rough concensus on is as follows:
>
> [ ] reply-to is primarily intended to allow the original
>     author of the message to state his preference as to where
>     replies to this message should be sent
>
> [ ] reply-to MAY be added by a mailing list to encourage readers to
>     reply to the list and/or the orignal author.
>
> [ ] reply-to MAY be used by a sender to encourage replies to go to
>     her primary mailbox, though it's probably better to use From
>     for this purpose if the sender's UA supports doing so.
>
> (All of the above appear to be consistent with RFC 822, section 4.4.3;
> however, the use of the From field could be considered a slight change.)

That all sounds good to me.

> [ ] while a UA SHOULD respect reply-to when generating replies,
>     in practice, recipients can reply to anyone they want to.

I'm not sure what you mean by "respect Reply-to".  I like a UA
that let's me easily reply to any or all of the addresses in a
message, without reentering them.  I think that it would be good
for such a UA to somehow indicate that Reply-to is the preferred
choice of the sender (show it in bold, or whatever).  I think that
UAs should be encouraged to somehow confirm with the user where
a reply is going to be sent (e.g., by making them select from a
menu of addresses, rather than having Reply and Reply-to-all
and users having to remember what each function really does.

> I don't yet see concensus on:
>
> [ ] whether or not lists should be discouraged from munging reply-to

I don't think that consensus on this is likely to happen.  How about
just laying out in neutral language what the consequences of doing
it and not doing it are, without taking a position?

> [ ] whether reply-to is intended to allow the original author
>     of a message to state his preference as to where replies
>     to HIM should be sent (note contrast from above)

I think that's an allowable use.  How is this different from the third
item in your first list above?  Anyway, how would restrictions in what
a sender puts in Reply-to be enforced?  People will put what they want
there and a program is not going to be able to second guess them, nor
should it as long as the addresses are valid.