Re: [Gnu-arch-users] Re: the dangers of no reply-to munging; Xouvert update

[Robert Anderson]

On Tue, 2003-08-19 at 03:18, Miles Bader wrote:
> Jonathan Walther <address@hidden> writes:
> > Not everyone uses mutt.  Pine, elm, Eudora, even Outlook Express are
> > common mail clients, and none of them have a concept of "group reply".
> > Email is one of those things that should "just work".
> 
> If you munge Reply-To things don't `just work.'  If they did, I
> wouldn't be arguing.
> 
> Munging Reply-To mucks with headers in a way that can make them
> inconsistent:
> 
> If a sender specifies their own Reply-To header (like I do), then either
> (1) it gets smashed by the list's Reply-To, and private replies go to
> the wrong place, or (2) the list manager is smart and _doesn't_ munge
> existing Reply-Tos -- just fine right?  Nope.  If that happens, then
> people on the list get used to the majority of Reply-Tos being munged,
> and so they get into the habit of always hitting `r' or whatever; if
> they then want to followup to me, then the reply goes to the wrong
> place, exactly the problem you said you just solved.

Sorry but I must be missing something about your argument here.  Yes, it
doesn't "do the right thing" if I want to reply to you.  So what?  It
doesn't "do the right thing" _ever_ with your scheme if what I want is
to reply to the list only.  I simply put the emphasis on the common
case, and you seem to be putting it on the non-common case.

Now, you're going to say "well that's a broken MUA problem" with respect
to the duplicate replies.  Maybe, but I'm using the default MUA on RH9
(evolution).  That doesn't seem like an outlier to me.  I can't think of
what would be more mainstream, in fact.

> That's the real problem -- enough lists do this munging that many people
> have come to expect it, and get confused when they see a list that
> doesn't do it; like a giant ball of sticky hair, this practice has
> grown, spurred on the the ignorance of users and list-masters alike.

Why is it a problem again?

> However that doesn't change the fact that doing it breaks the headers.

I don't understand what it "breaks" about the headers.  From your above
discussion it sounds like if it does not munge existing reply-to's, the
only problem is that it creates a default that _you don't like_.  Well,
I happen to like it.

> God I can't believe I'm having this argument.  Please people just bind
> `r' to `reply-to-all', then things will work.  If you're using a stupid
> mail program, switch to something better.  That's probably too much to
> ask on the lynrd-skynrd mailing list, but surely gnu-arch-users is a bit
> higher-brow?

My "joe everybody schtick" that you criticized is a reaction to both you
and Mr. Turnbull's schtick about "high brow" and "anybody worth
corresponding with" or whatever rhetorical crap that was.  That's not
much of a replacement for a real discussion, IMO.

BTW, it's Lynyrd Skynyrd.

Look, the reply-to munging seems to follow a very simple idea behind a
mailing list that makes a lot of sense.  That simple idea is that the
"list" is a robotic person who receives email replies and sends them
back out to everyone on his list.  So if I someone sends mail to the
"list guy", he then sends them to me.  When I hit "reply", why shouldn't
it go to the "list guy"?  After all, I signed up with the "list guy" to
receive his emails.  That seems perfectly natural and reasonable.  Now
you're probably going to quote RFC's and tell me "that's not how it
works."  Well, who cares - it's a perfectly natural way of thinking
about it and the resulting behavior is very natural and easy to
understand.

Bob