Re: [META] proper use of mailing list (was RE: Capturing current handler name in a string variable)
Re: [META] proper use of mailing list (was RE: Capturing current handler name in a string variable)
- Subject: Re: [META] proper use of mailing list (was RE: Capturing current handler name in a string variable)
- From: Chuq Von Rospach <email@hidden>
- Date: Sun, 30 Dec 2001 22:55:35 -0800
(apologies in advance to lumping replies to multiple questions in one
response, but I'm trying to keep the noise on the list down to a dull roar.
And please: if you wish to discuss this further, please talk to me
privately. Take it off the list, and don't annoy the other users by
continuing this in public... -- chuq)
On 12/30/01 12:25 PM, "Paul Berkowitz" <email@hidden> wrote:
>
All this is very true, Chuq. But I think it's safe to say that just about
>
everyone who is Replying to All is using it as a device to reply to the list
>
without having to copy and paste the list's email address every time,
In my experience running these things, no, that's not a safe assumption.
>
So you should probably count "sightings" of such strings of multiple
>
recipients and cc's as evidence of a desire for the alternate "Reply to
>
List" behavior, in addition to what you found were small numbers in favor of
>
it.
Even better: I survey the users. I don't guess, when I'm trying to find out
what users want/need, I do formal surveys. That way, among other things, you
don't bias the numbers with your own preconceptions of what you think the
users might want, and you don't bias the numbers by only counting the noisy
ones that pop up in the mail lists and ask for stuff. Not to say that's not
sometimes an appropriate thing to do (but in all honesty, the way it ought
to be done is not on the list, but by talking to the admins directly, since
these thigns, as you notice, tend to create off-topic, noisy meta
discussions that get in the way of the real reason these lists exist, and 95
times out of a hundred, the answer is already in a FAQ, or already something
we've answered and can easily give the answer again privately without the
chaos and angst)
So I don't need to try to guess what the users are thinking. I've gone out
and asked them, multiple times over the years. And the vast majority of
users DON'T WANT reply-to coerced on mailing lists. I've found that to be a
consistent result of surveys I've taken.
>
So we really are stuck with the annoyance. Yes, it's a minor annoyance,
You consider it an annoyance. Most people want it that way. Should we change
it to please the "noisy minority"? I don't believe that's the right thning
to do.
>
I do
>
find it hard to believe that most people prefer to reply privately - I just
>
don't believe that.
I've taken over a dozen surveys over the years. I've never had even close to
a majority preferring it. It's almost always 12% or under.
>
have your reasons, so that's OK. But it doesn't hurt for people to become
>
aware of the annoyance they might be causing some other people.
Yes it does, when it gets in the way of the list doing it's business. And
this list is busy enough without adding to the noise with meta-discussions
about how the list ought to run. That's not making people aware of the
annoyance of this: it's simply annoying to most of the users on the list. I
know, because every time one of these meta discussions break out, they email
me. I get lots more requests to shut up these discussions than I do to
implement what these discussions ask for. Honest.
>
still happen, and I'm prepared to be fairly tolerant as Chuq suggested. it
>
would certainly be nice if every email client had clever ways of dealing
>
with mailing list's different behaviors.
The problem here is that the list-header setup is hosed. List admins try to
make reply-to do too many things, and none of them are really correct.
The problem is there is a trinary operation here: reply-to-author,
reply-to-list, reply-to-all. And a binary data field. So admins argue
endlessly about the "best" way to set reply-to, when the reality is whatever
you do, it's wrong in some way or another.
That's one reason why the new RFC for list-* headers (see full headers on
any of lists.apple.com's messages....) was done, with the list-post header.
The idea is to build in support in the headers so mail clients can create an
interface to do things RIGHT. When mail-clients support those headers
properly, there WILL be a "post to list" option, as well as "unsubscribe",
"help" and other options available, so we don't have to keep trying to futz
things to do things "less wrong". The list server supports it, now it's up
to the clients to come on board. The real answer is to argue with your mail
client authors to do that properly, not try to fake it with reply-to.
On 12/30/01 3:28 PM, "Sander Tekelenburg" <email@hidden> wrote:
>
At 12:25 -0800 UTC, on 30-12-2001, Paul Berkowitz wrote:
>
> [...] You can't really set up a filter for rejecting duplicate subjects
>
> and senders, as you were advising Sander,
Depends on your mail system. Procmail does it pretty easily, by filtering on
duplicates of the message-id. Again, my argument is: if you want this, and
your mail client can't, then yell at the client authors to do it, don't try
to coerce the mail server into hacking around what is in reality a
user-specific preference. Server-based settings should be for things that
are global in nature, not things that some users set and force other users
to put up with, because, well, no matter how you set them, some users will
be pissed at the result. My job, as admin, is to figure out what's 'best'
for the overall user population and list operations on a global/macro sense.
Then, each user should customize their client environment to make it look
the way they want.
>
Anyway, I prefer prevention ;) Without actually saying so, it seems Chuq
>
suggested we use a Reply-To header if we want to 'force' (there's no such
>
thign on the Net) replies be sent to the list.
No, I don't indirectly support that. And while, in all honesty, the current
list server allows that to work, don't expect that to continue indefinitely,
and it's explicitly Not Supported. That's been a point of discussion among
the list server developers for mailman, and I expect that to be changed in a
future release, because as I've pointed out to them, it opens up a list
server to being used as an attack device on a user as a mailbomb delivery
system, and it also creates serious user-interface consistency problems that
confuse the heck out of users who aren't mail-header experts (which is,
these days, almost all users) -- because lists suddenly start working
differently based on which author is posting, and that creates lots of
confusion among users. Bad karma.
On 12/30/01 3:28 PM, "Sander Tekelenburg" <email@hidden> wrote:
>
> Then you should teach your client to strip duplicates out or your incoming
>
> mail stream.
>
>
What do you do when you break your leg - use pain killers[*], or try to get
>
it fixed?
Wrong analogy, actually. If I break my leg, I don't tell other people to fix
it by sticking cotton in their ears to get away from my screaming. I get the
leg fixed.
>
> list?" of course not. And you'd ignore them, most likely.
>
>
Actually, no. I'm usually more polite than that. I would point out *why* I do
>
not honour such a request.
But you'd still not honor it. So why do you expect others to honor yours? If
you know, going in, that if someone asked it of you, you wouldn't do it, why
did you make the same demand of others? Isn't that a bit selfish?
>
Isn't that the whole point of a mailing list like this - informing each
>
other? Then why should that only apply to AppleScript?
This is an Applescript list. Not a "how to run mail lists" or "how to use
mail lists" list. It's not the purpose of this list to tell users how to use
this list (or other lists). Especially when you're not the admin of the list
and can't set policy on the list. Admins HATE it when users run behind their
back and try to enforce policies they aren't empowered to enact in the first
place. Hint hint. (grin)
>
Sounds a bit like saying that if some people prefer to drive 200mph through
>
urban areas the rest of us should just have to deal with that. Not my ideal
>
society.
No, it's more a case of everyone sitting in the park and enjoying the
sunlight, and not attempting to tell the couple at the next table to stop
playing bridge, because the appropriate card game in this park is canasta.
You know what happens when you have a group of relative strangers in the
same space, and one of them decides to tell the others how things ought to
happen? Fights. That's why in the real world, people generally don't do
that: they call the authority instead. But online, some folks simply decide
they can become the authorities, and that's where a good chunk of these
list-fights start. And that's why we take a dim view of folks playing "topic
cop" when they have no authority to....
>
I just meant to draw people's attention to how their method affects others.
In all honesty, though, you shouldn't have. You should have gone to the
admins, and not attempted to force your preferences on everyone else. That
was, frankly, like walking into the park and turning your personal radio up
really loud, so that everyone could hear it. If you felt people were doing
something wrong, talking to the admins is the way to handle it: because the
admin has the authority to MAKE people fix things -- or to tell you that
things are working the way the folks in charge want it to work, and save us
yet another round of these meta-discussions on the list.
There's a right way and wrong way to do things. This was not the right way.
>
In my experience, on the Net people are often quite unaware of how what they
>
send/publish shows up at the other end.
And others don't seem to be aware that just because they want something some
way, doesn't mean it'll happen. Or are aware of how they're viewed when they
stand up and try to force their personal preferences on everyone else...
Now, please: let's take this back offline. If you want to discuss this
further, please do it with me privately, and let the list get back to work.
Chuq