[TIP] list mechanics

[Ben Finney]

Titus Brown <titus at caltech.edu> writes:

> On Fri, Feb 01, 2008 at 12:27:13AM +1100, Ben Finney wrote:
> -> Instead of "reply all", the only address you need to reply to is
> -> the list address. Your email client has all the information it
> -> needs to do this for you: the header of every message from the
> -> list has a standard 'List-Post'
> -> <URL:http://www.ietf.org/rfc/rfc2369.txt> field.
> 
> while I appreciate the educational response ;)

I'm sorry that my initial interactions with this community need to go
down this pedagogic path, but it seems best to point out that this is
largely a solved problem, and what the solutions are.

> there is no way for mailer software to know whether or not you are a
> member of the list.

This is true. On many lists, the convention is "reply only to the
list, unless someone specifically asks to be included in Cc".

This makes sense, because it's a good default assumption that someone
participating often in the list is also subscribed to it (either by
signing up to the list software, or subscribing to it elsewhere as I
do), and that non-subscribing participants are the unusual case.

> Thus it cannot eliminate your address from the response. mutt, my
> mailer software, which is generally standards compliant, does indeed
> insist on adding your name on a 'group-reply' for this reason.

Mutt has the "list-reply" feature, bound to the 'L' key, for exactly
this purpose.

> More generally, [RFC 2369] isn't an "internet standard" AFAICT.

It's listed at <URL:http://www.rfc.net/std1.txt> under "Standard
Protocols". That makes it an internet standard.

There's no requirement that an MUA *must* implement a "list-reply"
feature or any other that consumes those fields, but they are the
standard way of getting at the information to do so. Since that
standard is now quite old, any MUA that purports to be good for
interacting with mailing lists is rather derelict if it does not use
those fields to aid the user.

> This link
> 
> 	http://www.metasystema.net/essays/reply-to.mhtml
> 
> does give some good reasons to go with a Reply-To header set to the
> list.

They're not good reasons, and in fact RFC 2822 (which is also a
standard) <URL:http://www.ietf.org/rfc/rfc2822.txt> makes explicit
that this is not what the 'Reply-To' field is for. More explanation of
that here:

    <URL:http://woozle.org/~neale/papers/reply-to-still-harmful.html>

> However, my lists explicitly don't have reply-to-list set, because I
> find that doing that results in an annoyingly high rate of private
> person-to-person messages being sent to the list. Were the reply-to
> set, and I wanted to reply directly to you and NOT to the list, I
> would have to go through extra steps.

Indeed, that's one of the major reasons why the list settings
shouldn't mess with Reply-To. The list is *already* doing exactly the
right thing by providing the RFC 2369 fields in every message header.

> Finally, note that python-dev is set up in exactly the same way, and
> has the same sender-duplication problem that you complain about
> above. If you can convince them to change their config, then I'll
> consider changing :)

Fortunately I'm quite convinced that Reply-To munging is the wrong
approach. Thanks for being sensible on this.

-- 
 \        "What if the Hokey Pokey IS what it's all about?" —Anonymous |
  `\                                                                   |
_o__)                                                                  |
Ben Finney