On Fri, 27 May 2005, Bruce Lilly wrote:
Kmail and Evolution provide options for directing responses "To
List", in addition to other places (e.g. "To Sender", "To Author"),
not a "context".
It's "the context in which the user wishes to reply". Tortured
terminology maybe.
I don't know -- maybe quirky developers (ones who like medium gray
text on a dark gray background).
Damn their souls. :)
Next question: Why do quite a few MUAs provide a "list reply"
function?
Because the most likely reason for it being set are:
- The author's preferred 'receive' mailbox is not their From
(ie to indicate preference for their personal mailbox)
Makes no sense -- for replies "to author" the From field mailbox can
(and should) be set accordingly.
Ok.
Umm, look at this message for example (and note past messages as
well). That specific case (along with a couple of others) is
specifically mentioned in RFC 724, which has been around for close
to three decades, and its successors.
I've only ever once seen anyone set "Reply-To: <list>", and they
eventually got burned by the risk mentioned below. I've never seen
anyone set Reply-To to both their own mailbox and the list mailbox
(until you replied to me, and I tried it with one or two of my
mails).
No, it's a problem with your proposal; Reply-To:
list-foo(_at_)example(_dot_)com works fine to specify responses to the
(including) list (possibly in addition to other mailboxes). That
removes any ambiguity about which list responses are supposed to be
sent to.
No it doesn't work fine.
It absolutely is not fine to set Reply-To == list mailbox because of
the *definite* risk of MUAs sending 'private' mails to the Reply-To.
(I've seen it happen countless few times both on lists which munged
Reply-To and with someone who set it themselves.).
If Reply-To is supposed to the answer, then at a minimum there should
be a sanctioned clarification of how implementations should deal with
it.
However, even if such a clarification document were released, you
still *wont* see much uptake in use of Reply-To to solve the problem
I (and a minority like me) have because it wont alleviate the risk
until /all/ implementations behave better.
Are you saying that the problems around use of Reply-To and lists
with deployed bad implementations we apparently are not to consider?
"Reply-To /should/ be the answer, therefore it is the answer - real
world be damned"?
A new header, with restricted and unambigious scope, to solve only a
very specific problem around replying to list emails (something
common enough that MUA implementors feel is worthwhile enough to add
an extra button to their UI for - even Apple, who dont like UI
clutter) is out of the question, because a header which is widely
misinterpreted /should/ work, if only everyone fixed their software?
:)
Or have I missed some key point in the discussion?
regards,
--
Paul Jakma paul(_at_)clubi(_dot_)ie paul(_at_)jakma(_dot_)org
Key ID: 64A2FF6A
Fortune:
You! What PLANET is this!
-- McCoy, "The City on the Edge of Forever", stardate 3134.0