In <200410170956(_dot_)27444(_dot_)blilly(_at_)erols(_dot_)com> Bruce Lilly
<blilly(_at_)erols(_dot_)com> writes:
On Fri October 15 2004 21:12, Hector Santos wrote:
FWIW, I agree with almost everything Hector Santos has said in this
exchange. Although I accept that "me too" postings are not generally a
good idea, I think they do have a place when there is a sustained campaign
by one poster to shout down the whole of the rest of the community.
Bruce, does this suggest not a software engineer or developer or even close
to understanding the concept? If so, I won't hold that against you thought.
Writing of MUAs as being "trained" demonstrates at best a
very odd way of thinking about MUAs; certainly not to be
expected from a software engineer or developer or anybody
"even close to understanding the concept" of MUAs.
I found that metaphor completely appropriate, and your failure to
understand it as symptomatic of the huge gulf between yourself and Real
(TM) software engineers.
It can't be done. While there are specific (RFC 2369, 2919) message
header fields that a mailing list expander MAY insert, and while a
mailing list expander SHOULD appropriately set the envelope sender
address to the mailing list maintainer's mailbox, there is no guaranteed
way that any MUA can always identify all list messages. That is because
the mailing list-specific fields are optional and because a mailing list
mailbox is indistinguishable from an individual mailbox. Mailing list-
specific fields cannot be made mandatory under IETF rules, because
imposing requirements is reserved for cases where it is necessary
to ensure interoperation and to avoid network damage (such as
severe congestion caused by retransmissions); since the list-specific
fields play no role in protocol but are merely informative, there is no
way that omitting any or all of those fields can lead to such problems.
Bullshit! The present mess that we are discussing is indeed an
interoperability problem, since there is no clear and consistent way of
conducting discussions between users on arbitrary lists using arbitrary
MUAs without being aware of the particular quirks of each user, list and
MUA and making manual corrections to allow for them. Therefore there is no
reason why an appropriate standard should not use at least SHOUL wording
as regards inserting those headers.
If the IETF mechanisms will not allow that to be done, then there is
something seriously wrong with the IETF mechanisms. But even if that is
the case, then it can (and should) be done in a BCP document.
Moreover, there is no need for an MUA to determine that a list is
involved; an author is supposed to set the Reply-To field to indicate
where responses should go if that is somewhere other than directly
to him (the author), and a respondent only needs to be able to see
the author's recommendation and to determine whether he (the
respondent) wishes to comply with that recommendation or to send
his response elsewhere. That applies for *all* messages, mailing list
or otherwise.
No, you are the only person here who believes that that solution is
adequate. Everybody else realises a need to distinguish between the
address that the author wants personal replies to go to (which may not be
his From address) and the address that the author wants list replies to go
to. You cannot convey that information with just one header.
--
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl
Email: chl(_at_)clerew(_dot_)man(_dot_)ac(_dot_)uk 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