[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: bugs + rant + constructive criticism (long)

On Thu, Jan 04, 2001 at 10:43:05AM -0800, Philip Brown wrote:
> [ Craig Sanders writes ]
> > On Wed, Jan 03, 2001 at 11:26:25AM -0800, Philip Brown wrote:
> > > And in the case of the debian mailing lists, you should "reply to" the
> > > list.
> > some replies should go to the list, and some replies should be private.
> > it's up to the person writing the reply to make that decision, not the
> > list software.
> But the primary point of a mailing list is for discussion ON THE LIST.
> Do you want to disagree with that?

yes, of course i will disagree with that because it's wrong.

there are as many primary purposes of mailing lists as there are mailing

> So headers should be optimized for group discussion.
> Replying to individuals is a secondary function.

that's for the person writing the reply to decide, not you and not the
list operator.

> >setting reply-to back to the list just makes it difficult (or in
> >some cases impossible) to reply privately. ... .... however reply-to
> >munging by list software does have the serious disadvantage of
> >replacing any Reply-To header created by the original author of a
> >message.
> So what, if the mailing software rewrites From: to have any reply-to
> information from the original sender? Then the information is still
> available.

so your proposed solution to the evils of header munging is to do more
header munging in an attempt to correct the problem?  what a great idea!

there are better solutions, including the Mail-Followup-To header which
has been documented and in use for years. all decent MUAs should have
implemented it by now. if they haven't then they are inadequate.

> > the Reply-To header exists for the *person* who originally sent the
> > message to be able to direct replies to their preferred destination. it
> > is not there so that mailing lists can screw with it.
> So your argument is
>  "A mailing list is not a person, so it can't use reply-to:".
> Bad argument.

no, my argument is that if a list sets the reply-to header then it makes it
difficult (or impossible in some cases) to reply privately to the author of
a message, so it should not do it.

not munging the Reply-To header also follows the principle of doing
least harm. if a reply is accidentally sent privately, no harm is done -
just forward the message to the list if it matters.  OTOH if a private
reply is accidentally sent to the list (which happens all too often when
Reply-To points back to the list) then harm is done with no possibility
of undoing it because confidential or embarrasing material is now posted
(and probably archived) publicly.

setting Reply-To back to the list is brain-damaged and moronic behaviour
on the part of the list operator.

> [http://faqs.org/rfcs/rfc822.html]
> rfc822, section 4.4.3, EXPLICITLY MENTIONS 
> "text message teleconferencing" groups 
>   (eg mailing lists) as potential users of the reply-to header,
> expressly for the purpose of having "reply" direct email to the list!!!!

yes, and the people who wrote that (and many others) have since
realised and stated that it is broken and causes many more problems
than it solves. that's why there are other documents recommending the
Followup-To and Mail-Followup-To and other headers.

this has been discussed many times over many years.  you're not adding
anything new to the discussion.

> And finally, example A.3.3 EXPLICITLY shows that "reply-to" is NOT
> exclusively for "who wrote the message". It is for "Where do you want
> replies to normally go to"

what's your point?

it is for the person who wrote the message to direct where replies are
to go.


craig sanders

Reply to: