Keith Moore wrote:
There are fewer problems with lists that only insert reply-to
if it isn't there already, but some problems still remain.
(see http://www.cs.utk.edu/~moore/opinions/reply-problem-list.html ,
items B3, B4, and B5)
Bearing in mind the context of the discussion:
1. a hypothetical list consensus for reply-to-list as default
implemented by
2. list expansion adds a Reply-To field pointing to the list
submission mailbox iff there is no Reply-To field in the
message
Let's look at those:
B3. If a message is sent to multiple recipients, one
or more of which is a mailing list, and one of
those mailing lists adds a Reply-To field,
members of that mailing list have difficulty
replying to all recipients of the original
message. This thwarts cross-list discussions,
which are sometimes desirable.
If the sender doesn't set Reply-To and the lists don't
set Reply-To, then a reply-all reaches all recipients
plus the author(s) plus and other original message recipients.
Whereas a plain "reply" would go only to the author(s).
In the situation posited, some list sets Reply-To;
presumably To and Cc are not corrupted. Therefore a
reply-to-all will still go to all recipients plus the
author(s) plus any other original recipients, with no
particular difficulty involved. A plain "reply", however
will only go to the one list that set Reply-To. If the
sender recognizes that multiple lists may be involved, he
can forestall the situation by including a Reply-To
field that points to all lists (or to the highest
precedence list which includes sublists).
B4. If a mailing list leaves an existing Reply-To
field in place, but supplies a Reply-To field
if one is missing, there's no way for the
recipient's UA to know whether the field was
added by the list or not. The results can be
surprising.
(If Reply-To points to the list and the recipient
presses a button labeled "reply to author",
there's a good chance he doesn't want the
response to go to the list.)
Agreed that failing a comment or some other indication
that a list expanded added a Reply-To field -- "no way"
overstates the case -- a user can't tell where the Reply-To
field came from. Why should he need to know or care?
As for the hypothetical UA that has a button labeled
"reply to author" but which uses the Reply-To field when
it is pressed, that UA is broken; the From field -- not
Reply-To -- indicates the message author(s). Reply-To
is not and never has been guaranteed to point to the
message author. Now if that hypothetical "reply to
author" button uses the From field, which is what a
button with such a label ought to do, B4 as a "problem"
vanishes in a puff of smoke. Likewise if the label
says simply "reply" and Reply-To is used -- that is how
a reply is normally constructed.
B5. If a message is sent to a mailing list that
adds or mungs Reply-To, and the author of
that message is not on the mailing list, any
replies that go just to Reply-To won't go to
the author. Even replies that go to
Reply-To+To+Cc won't go to the author of the
subject message unless the author has included
himself in the To or Cc list.
Such an author who does not want replies other than to
the list will be happy as a pig in manure. An author
who wants only personal replies can add a Reply-To
field -- remember, we're talking about a list expander
that does *not* add Reply-To if one exists -- and an
author who wants both list and personal replies can
set Reply-To to include list and personal addresses if
he wishes that to be the case for a plain "reply", or
he can -- as you suggest -- simply put his address in
To or Cc (leaving out Reply-To or setting it to point
to the list mailbox) if he wants plain replies to go
to the list and reply-all replies to go to the list
plus his mailbox.
Now it may be the case that some UAs might not make it
easy to set Reply-To. That may warrant some advice to
UA authors. The UAs that I tend to use have no serious
problems in that regard; most have the ability to set a
default Reply-To field on a per-"identity" basis and
provide an easy-to-use means of switching identities.
Some have the ability to set the default Reply-To field
on a per-folder basis without changing "identity". The
ones with GUIs have the ability to copy-and-paste
between address fields, so starting with reply-to-all,
it is simple enough to copy the list mailbox and paste
it over Reply-To whether or not there is a default.
That leaves non-GUI MUAs with no means to set a default
Reply-To field on a per-folder or per-identity basis or
which otherwise make it difficult to switch identities.
The non-GUI UAs that I am familiar with that are likely
to be used by anybody that receives a significant amount
of mail all have at least some sort of "address book"
that can be used to select the appropriate list mailbox
for a Reply-To field, once it has been entered one time
in the address book. So while it's possible that somewhere
there's a UA that makes setting Reply-To difficult, I
don't recall having used it (and if I did use it, the lack
of such basic functionality would have been a good reason
to have moved on to something else).