PDA

View Full Version : RFC: slimdevices.com mailing list configuration



Healy
2003-12-05, 10:41
In message <20031205173719.GC1668 (AT) bklyn (DOT) org>, Caleb Epstein writes:
>
> I have noticed that the slimdevices.com mailing lists send out
> all postings with a Reply-To: of the list address. This
> forces all replies to go back to the entire list, even if what
> you intended was to direct a response to the poster.
>
> I content that it would be preferable to NOT do this and let
> people who want to reply on-list just do a reply-all. Simple
> reply would go back to the message originator, reply-all would
> go to the originator as well as the list.
>
> Thoughts?
>

That defeats the point of a mailing list. If you want to reply to
someone offlist, why can't you just enter their email by hand? It
only take a couple of seconds longer than clicking / typing reply.

Besides, if it's setup the way you are talking about people will get
double mails when reply to all is used.

-Healy

Got Trouble? http://trouble.nwgeeks.com/
========================================
The dirt trail to the bathroom
is screwed because of some bitchy farm boy on tatooine...
Apu is playing counter-strike.

Caleb Epstein
2003-12-05, 10:50
On Fri, Dec 05, 2003 at 05:41:06PM +0000, Healy wrote:

> That defeats the point of a mailing list. If you want to reply to
> someone offlist, why can't you just enter their email by hand? It
> only take a couple of seconds longer than clicking / typing reply.

Because the way I am suggesting, you can do it by hitting
reply, or reply-all if you want everyone to see. No
error-prone copying and pasting of email addresses required.
I think it makes more sense personally. Reply = to the person
that sent it; Reply-All = to everyone who saw the message.

I am on many mailing lists, and the slimdevices ones are the
ONLY ones I have seen that are configured this way.

> Besides, if it's setup the way you are talking about people will get
> double mails when reply to all is used.

Nope, not with the Mailman software that the fine and clued-up
folks at Slimdevices are using. It takes care of that
automagically.

--
Caleb Epstein | bklyn . org | Experience is what you get when you were
cae at | Brooklyn Dust | expecting something else.
bklyn dot org | Bunny Mfg. |

Jason Snell
2003-12-05, 11:09
Caleb Epstein wrote:

> I am on many mailing lists, and the slimdevices ones are the
> ONLY ones I have seen that are configured this way.

I have almost never seen a public, well-founded discussion mailing
list that didn't have reply-to set to the list. It used to happen,
maybe 10 years ago. But then people figured out that it wasn't
effective at all. Threads got short-circuited.

People who want to talk on a list want to talk on the list. While you
always have the option to "take it off-list," that action is a
conscious opting out from the group discussion. The reason for a
group is to have group discussion.

This is practically netiquette at this point. It's a convention. I am
dumbfounded that all your other mailing lists don't work this way,
and somewhat fascinated. I wonder what those lists are (digests?
homespun lists from a particular listadmin who doesn't like the
Reply-to header?), but I can tell you from my perspective that 99% of
the discussion lists I've been on in the past 5 years have had
reply-to set to the list itself.

-jason
--
Jason Snell / Editor in Chief, Macworld / jsnell (AT) macworld (DOT) com
415-243-3565 / AIM-iChat: MW jsnell

Caleb Epstein
2003-12-05, 11:19
On Fri, Dec 05, 2003 at 10:09:49AM -0800, Jason Snell wrote:

> This is practically netiquette at this point. It's a convention. I

Not on the mailing lists I'm familiar with.

> am dumbfounded that all your other mailing lists don't work this
> way, and somewhat fascinated. I wonder what those lists are
> (digests? homespun lists from a particular listadmin who doesn't
> like the Reply-to header?), but I can tell you from my perspective
> that 99% of the discussion lists I've been on in the past 5 years
> have had reply-to set to the list itself.

Mostly open source software development topics and (legal)
live music trading/discussion lists. Some are large, with
membership in the thousands.

I have seriously never encountered lists confiugred this way.

--
Caleb Epstein | bklyn . org | We lie loudest when we lie to ourselves.
cae at | Brooklyn Dust | -- Eric Hoffer
bklyn dot org | Bunny Mfg. |

Jason Snell
2003-12-05, 11:24
[apologies for this being hideously off-topic.]

Caleb Epstein wrote:

> Mostly open source software development topics and (legal)
> live music trading/discussion lists. Some are large, with
> membership in the thousands.

The first one is a little surprising. The second isn't quite, if it's
more trading than discussion. It sounds like you're on lists so large
they need "self-moderation" -- basically the only way to stop the
flow of mail is to force people to opt in to sending a note to the
list. (Moderated lists are another good example of not having
reply-to headers in all cases.)

I've never really been on a list like that, mostly because lists that
large are so unwieldy that I doubt I would see the appeal.

But it's an interesting perspective -- thanks for responding to my
curiosity. Still, for a list like this one, I just can't see how it
would be better to strip out the Reply-To headers. Our membership
sure isn't in the thousands, at least in terms of active posters. :-)

-jason
--
Jason Snell / Editor in Chief, Macworld / jsnell (AT) macworld (DOT) com
415-243-3565 / AIM-iChat: MW jsnell

Paul Warren
2003-12-05, 11:32
On Fri, Dec 05, 2003 at 10:24:31AM -0800, Jason Snell wrote:
> [apologies for this being hideously off-topic.]

Ever get that oh-no, not again feeling? I quote (unedited) a friend and
mailing list admin's response to this, with which i happen to agree:


Well, the two articles usually quoted in this context
are--

http://www.unicom.com/pw/reply-to-harmful.html
reply-to munging considered harmful

and

http://marc.merlins.org/~merlin/perso/reply-to-useful.html
reply-to munging considered useful

The standard of argument in these two articles, especially
the second, is fairly laughable, as you might expect.
About the only decent points in them are:

1. replacing a proper reply-to with the list's address
may hide the author's intentions, making it
impossible to reply to an individual author;

2. not replacing the reply-to with the list's address
may lead to a substantial waste of time and
bandwidth as numerous addresses `snowball' into the
recipients for each message;

3. inserting the name of the list in the reply-to
header violates the `principle of least surprise',
because people do not expect replies to go to the
list unless they have specifically asked for this
to happen by typing `g' or answering `yes' to
`reply to all recipients'.

A fourth argument which some people find convincing is--

4. RFC822 says that you should munge the reply-to
headers for `text-message teleconferencing'.

This is just plain silly. The fact that the RFC says
something doesn't mean that it is gospel truth, and it
doesn't answer the point in (1) above.


In my opinion, whatever the strength of the other
arguments, (1) clinches it. So, on that basis, I don't
anticipate changing the list to munge reply-to headers.

Paul

Jason Snell
2003-12-05, 11:47
>Ever get that oh-no, not again feeling? I quote (unedited) a friend and
>mailing list admin's response to this, with which i happen to agree:

My favorite part of these documents is this:

>Any reasonable, modern mailer provides this feature. I prefer the Elm mailer

or as your friend says:

> people do not expect replies to go to the
>list unless they have specifically asked for this
>to happen by typing `g' or answering

Uh, command-line unix mail readers are not remotely "modern." Arguing
that reply-to headers shouldn't exist because we can all hit "g" in
elm is crazy.

In fact, I don't believe I've seen a "group reply" command in any
graphical e-mail client, at least not on the Mac. And if they are
there, nobody seems to know about their existence.

I will agree that, in a perfect world, all e-mail programs would let
you opt to do a person-reply or a group-reply, and we wouldn't need
Reply-To headers. Let me know when you see a perfect world, and I'll
jump on. :-)

-jason
--
Jason Snell / Editor in Chief, Macworld / jsnell (AT) macworld (DOT) com
415-243-3565 / AIM-iChat: MW jsnell

Caleb Epstein
2003-12-05, 11:57
On Fri, Dec 05, 2003 at 06:32:23PM +0000, Paul Warren wrote:

> 1. replacing a proper reply-to with the list's address may hide the
> author's intentions, making it impossible to reply to an individual
> author;
>
> In my opinion, whatever the strength of the other
> arguments, (1) clinches it. So, on that basis, I don't
> anticipate changing the list to munge reply-to headers.

Huh? It already DOES, which is what I (seemingly alone,
annoying everyone else on the list) am lobbying against.

Anyway, I seem to be in a minority here AFAICT so I'll shut up
and remember to hit "g" when I want to reply to the poster
directly and remove the list address from the To: line. This
is IMHO horribly backwards, but its live-able.

--
Caleb Epstein | bklyn . org | QOTD:
cae at | Brooklyn Dust | The only easy way to tell a hamster
bklyn dot org | Bunny Mfg. | from a gerbil is that the gerbil has
| | more dark meat.

Caleb Epstein
2003-12-05, 12:02
On Fri, Dec 05, 2003 at 10:47:07AM -0800, Jason Snell wrote:

> > people do not expect replies to go to the
> >list unless they have specifically asked for this
> >to happen by typing `g' or answering
>
> Uh, command-line unix mail readers are not remotely "modern." Arguing
> that reply-to headers shouldn't exist because we can all hit "g" in
> elm is crazy.

<luddite>

I hate to say it, but every time I have to use a "modern" GUI
mail reading client I end up giving up after a few days of
torture and going back to simple, efficient,
standards-compliant mutt.

You shouldn't need 3D user interfaces and HTML rendering
widgets to read email.

</luddite>

> In fact, I don't believe I've seen a "group reply" command in any
> graphical e-mail client, at least not on the Mac. And if they are
> there, nobody seems to know about their existence.

Huh? You've never seen the "Reply All" button thats in EVERY
mail application ever written? The mind she boggles.

> I will agree that, in a perfect world, all e-mail programs would let
> you opt to do a person-reply or a group-reply, and we wouldn't need
> Reply-To headers. Let me know when you see a perfect world, and I'll
> jump on. :-)

One such world exists. It is the world populated by mailing
lists which do NOT add a reply-to header.

--
Caleb Epstein | bklyn . org | "No matter where you go, there you are..."
cae at | Brooklyn Dust | -- Buckaroo Banzai
bklyn dot org | Bunny Mfg. |

Roy M. Silvernail
2003-12-05, 12:42
On Friday 05 December 2003 14:02, Caleb Epstein wrote:

> <luddite>
>
> I hate to say it, but every time I have to use a "modern" GUI
> mail reading client I end up giving up after a few days of
> torture and going back to simple, efficient,
> standards-compliant mutt.

And simple, efficient, standards-compliant Mutt asks you if you want to reply
to the Reply-to: address. If you say 'no', it then offers the From: address.
I'm sorry... where was the problem here?

> You shouldn't need 3D user interfaces and HTML rendering
> widgets to read email.

Agreed, but graphical MUAs do have their appeal.

> One such world exists. It is the world populated by mailing
> lists which do NOT add a reply-to header.

And that world doesn't have SqueezeBoxen, so I ain't going!

#ifndef HUMOR_ENABLED
#include <std_smiley.h>
#endif

Jason Snell
2003-12-05, 12:51
[Okay, any further replies I have on this topic will be made by
overriding the Reply-To header and mailing people back directly.
Sorry for the list pollution. And thanks to Caleb for being civil in
all this.]

><luddite>

I give you major credit for admitting to it. :-)

>You shouldn't need 3D user interfaces and HTML rendering
> widgets to read email.

No, but the vast bulk of humanity uses interface involving icons,
windows, and menus. I haven't used a 3D interface yet (those glasses
sure look keen!), nor am I enamored with HTML rendering widgets. But
I am fond of Mr. Engelbart's little pointing-device invention and the
windowing metaphor those crazy kids at Xerox PARC developed to
replace the command-line interfaces of the 1970s.

Those who use command-line, text-based e-mail clients are to be
commended for their purity. (I used Berkeley mail for years and am
still quite skilled in the use of vi to exit text files, although I
prefer BBEdit, a tool that requires "pointing" and "clicking" from
time to time.) But this insistence that the rest of the world would
somehow be better if we stopped complaining and loved elm, pine, and
the like is laughable. Likening clicking with a mouse on a menu item
to "3D user interfaces and HTML rendering widgets" is an amusing
example of strawman methodology, but it has no basis in reality.
Genie's out of the bottle. Windowing-based e-mail clients are
probably used by 97% of all Internet e-mail users (if you count Web
mail, anyway).

>One such world exists. It is the world populated by mailing
> lists which do NOT add a reply-to header.

I find the use of reply-to-all in such situations unpleasant. Using
the "Reply" command is the common behavior, and we train ourselves
to use it, so oftentimes you reply to individuals even though you
meant to reply to the entire list. When you do "reply to all," it
usually includes the originator of the message, so they get two
copies of every reply -- one to them personally, one via the list.
(At least, that's my experience in Eudora.)

But I'll go back to my original point. People who make these
arguments by saying that we can all just press "g" in elm and be
happy aren't invalidating their argument, but they are showing a
colossal lack of perspective on how most people use e-mail. They're
not wrong -- but they are, in my opinion, really atypical.

-jason
--
Jason Snell / Editor in Chief, Macworld / jsnell (AT) macworld (DOT) com
415-243-3565 / AIM-iChat: MW jsnell

Caleb Epstein
2003-12-05, 13:05
On Fri, Dec 05, 2003 at 11:51:37AM -0800, Jason Snell wrote:

> I find the use of reply-to-all in such situations unpleasant. Using

And I find it pleasant and intiutive. I will agree to
disagree with you on this point.

> the "Reply" command is the common behavior, and we train ourselves
> to use it, so oftentimes you reply to individuals even though you
> meant to reply to the entire list. When you do "reply to all," it
> usually includes the originator of the message, so they get two
> copies of every reply -- one to them personally, one via the list.
> (At least, that's my experience in Eudora.)

Again, this doesn't happen with Mailman-controlled lists. At
least not in my experience. Slimdevices is using Mailman.

> But I'll go back to my original point. People who make these
> arguments by saying that we can all just press "g" in elm and be
> happy aren't invalidating their argument, but they are showing a
> colossal lack of perspective on how most people use e-mail. They're
> not wrong -- but they are, in my opinion, really atypical.

I still don't see how using reply-all in the case of wanting
everyone to see your message is atypical. Its how "normal"
non-mailing-list email dialogues work. Why are mailing lists
different?

--
Caleb Epstein | bklyn . org | Never commit yourself! Let someone else
cae at | Brooklyn Dust | commit you.
bklyn dot org | Bunny Mfg. |

Mark C. Langston
2003-12-05, 13:25
On Fri, Dec 05, 2003 at 03:05:41PM -0500, Caleb Epstein wrote:
> On Fri, Dec 05, 2003 at 11:51:37AM -0800, Jason Snell wrote:
>
> > I find the use of reply-to-all in such situations unpleasant. Using
>
> And I find it pleasant and intiutive. I will agree to
> disagree with you on this point.


Good MUAs know how to recognize a mailing list, and will ask you if you
meant to reply to the list, or the individual. Mutt is one such example
of a good MUA.

>
> > But I'll go back to my original point. People who make these
> > arguments by saying that we can all just press "g" in elm and be
> > happy aren't invalidating their argument, but they are showing a
> > colossal lack of perspective on how most people use e-mail. They're
> > not wrong -- but they are, in my opinion, really atypical.
>
> I still don't see how using reply-all in the case of wanting
> everyone to see your message is atypical. Its how "normal"
> non-mailing-list email dialogues work. Why are mailing lists
> different?
>


For cogent argument on the subject (which happens to reflect the
majority opinion of Them Who Build The Roads w.r.t. mailing lists), see
http://www.unicom.com/pw/reply-to-harmful.html

--
Mark C. Langston Sr. Unix SysAdmin
mark (AT) bitshift (DOT) org mark (AT) seti (DOT) org
Systems & Network Admin SETI Institute
http://bitshift.org http://www.seti.org

Roy M. Silvernail
2003-12-05, 13:28
On Fri, Dec 05, 2003 at 07:58:57PM -0000, Kevin Walsh wrote:

> The whole point of a mail list is to share article contents with
> a community of often like-minded people. If a user wants to force
> followups to be sent back privately then it's a bit like saying
> "I have a question for the list subscribers, but I don't really want
> to share the answers I get with anyone."

Indeed, the canonical Usenet example was "Please reply via email, as I
don't read this newsgroup."
--
Roy M. Silvernail is roy (AT) rant-central (DOT) com, and you're not
http://www.rant-central.com is the new scytale
Never Forget: It's Only 1's and 0's!
SpamAssassin->procmail->/dev/null->bliss

Mark C. Langston
2003-12-05, 13:41
On Fri, Dec 05, 2003 at 10:47:07AM -0800, Jason Snell wrote:
>
> Uh, command-line unix mail readers are not remotely "modern." Arguing

mutt is "modern". mh-e is "modern".

> that reply-to headers shouldn't exist because we can all hit "g" in
> elm is crazy.
>
> In fact, I don't believe I've seen a "group reply" command in any
> graphical e-mail client, at least not on the Mac. And if they are
> there, nobody seems to know about their existence.
>

Copy of Mozilla I've got running on my desktop right now (SuSE 8.2
default install): "Reply All" button sitting right there, plain as day.
This same button exists in versions of the Netscape GUI mailer going
back to at least 4.x verions on Unix and most other platforms.

Eudora has a similar, prominent button.

If I had my 12" AlBook in front of me, I'd open up Mail.App and
Entourage and locate that button/menu item in each of them for you as
well.

OE and Outlook both have a "reply all" function.



The fact that some of these mailers don't distinguish between "this
email is from a mailing list that I know about, and thus should only be
sent to the list" and "this email is from a person to a group of people,
so the reply should go to every address I can find in the headers"
signifies a problem with the MUA, not a need to insert a Reply-To.


> I will agree that, in a perfect world, all e-mail programs would let
> you opt to do a person-reply or a group-reply, and we wouldn't need
> Reply-To headers. Let me know when you see a perfect world, and I'll
> jump on. :-)


I think you misunderstand the purpose of Reply-To, and the dangers of
having an intermediary munge them.

If RFC2822 required agents to process the "Return-Path:" header, this
wouldn't be an issue. It doesn't. Therefore, Reply-To: should remain
unmolested in transit. It's not, on this list.

As a trivial proof, the Reply-To: on this message, as I send it, is set
to "vintcerf (AT) example (DOT) com". I guarantee you it arrives in your mbox with
my header contents replaced with those set by Mailman.

This is NOT a best, or even good, practice for mailing lists, for a
number of reasons, outlined in the URL I supplied in my other mail
addressing this topic.

--
Mark C. Langston Sr. Unix SysAdmin
mark (AT) bitshift (DOT) org mark (AT) seti (DOT) org
Systems & Network Admin SETI Institute
http://bitshift.org http://www.seti.org

Erv Walter
2003-12-05, 13:50
This discussion was fun for a while. Is there perhaps a mailing list about
mailing lists that you could discuss this on instead of the SlimDevices
list? If you want to lobby the SlimDevices folks, send an email to them
directly. A long drawn out discussion about this is off topic for this
list.

Thanks,
Erv

Caleb Epstein
2003-12-05, 13:52
On Fri, Dec 05, 2003 at 02:50:00PM -0600, Erv Walter wrote:

> This discussion was fun for a while. Is there perhaps a mailing
> list about mailing lists that you could discuss this on instead of
> the SlimDevices list? If you want to lobby the SlimDevices folks,
> send an email to them directly. A long drawn out discussion about
> this is off topic for this list.

Thats how it started. I emailed Dean asking him if he would
consider changing the configuration, and he asked me to
discuss it on-list to get other people's feedback.

--
Caleb Epstein | bklyn . org | "I say we take off; nuke the site from orbit.
cae at | Brooklyn Dust | It's the only way to be sure." - Corporal
bklyn dot org | Bunny Mfg. | Hicks, in "Aliens"

Erv Walter
2003-12-05, 13:54
Ok. Sorry if I jumped the gun. However, it seems like there is no
consensus. There are people and arguments on both sides of the aisle. It's
probably up to Dean and Co to just choose.

Erv

----- Original Message -----
From: "Caleb Epstein" <cae (AT) bklyn (DOT) org>
To: "SlimDevices Discussion" <discuss (AT) lists (DOT) slimdevices.com>
Sent: Friday, December 05, 2003 2:52 PM
Subject: [slim] RFC: slimdevices.com mailing list configuration


> On Fri, Dec 05, 2003 at 02:50:00PM -0600, Erv Walter wrote:
>
> > This discussion was fun for a while. Is there perhaps a mailing
> > list about mailing lists that you could discuss this on instead of
> > the SlimDevices list? If you want to lobby the SlimDevices folks,
> > send an email to them directly. A long drawn out discussion about
> > this is off topic for this list.
>
> Thats how it started. I emailed Dean asking him if he would
> consider changing the configuration, and he asked me to
> discuss it on-list to get other people's feedback.
>
> --
> Caleb Epstein | bklyn . org | "I say we take off; nuke the site from
orbit.
> cae at | Brooklyn Dust | It's the only way to be sure." - Corporal
> bklyn dot org | Bunny Mfg. | Hicks, in "Aliens"
>

Mark C. Langston
2003-12-05, 14:01
On Fri, Dec 05, 2003 at 06:32:23PM +0000, Paul Warren wrote:
>
> and
>
> http://marc.merlins.org/~merlin/perso/reply-to-useful.html
> reply-to munging considered useful

I know Marc.

I'm fairly certain I could get him to rethink and remove
that page, were I so inclined. There's no operational reason to munge
..Reply-To: headers (or any other header, for that matter, except perhaps the
..Subject: header, and even that's a bone of contention with many.)

There are multiple operational drawbacks, however. In particular,
several headaches for the list owner that are avoidable were Reply-To:
not munged.


For an example of a properly-configured mailing list, look no further
than the openbsd-* lists.

>
> This is just plain silly. The fact that the RFC says
> something doesn't mean that it is gospel truth, and it
> doesn't answer the point in (1) above.


The fact that lack of RFC-compliance is a very quick way to come up
against various incarnations of the Internet Death Penalty is usually
enough to motivate people to comply.

Non-compliant entities get ignored and routed around. This can be any
combination of troublesome, annoying, and/or expensive.


--
Mark C. Langston Sr. Unix SysAdmin
mark (AT) bitshift (DOT) org mark (AT) seti (DOT) org
Systems & Network Admin SETI Institute
http://bitshift.org http://www.seti.org

dean
2003-12-05, 15:48
On Dec 5, 2003, at 12:54 PM, Erv Walter wrote:

> Ok. Sorry if I jumped the gun. However, it seems like there is no
> consensus. There are people and arguments on both sides of the aisle.
> It's
> probably up to Dean and Co to just choose.

It is. Since there's no consensus, I'll leave it as-is for now.

Thanks for the lively discussion.

-dean

Mark C. Langston
2003-12-05, 16:54
On Fri, Dec 05, 2003 at 10:23:08PM -0000, Kevin Walsh wrote:
> it belongs. Someone said that lists like that do exist, but apart from
> "The elm appreciation society" mail list, I can't think of any project
> that would find it useful to configure its list server in such a way.
>


NANOG-L. BUGTRAQ-L (and its variants). All the OpenBSD lists. The
SUN-MANAGERS-L. All the FreeBSD lists. The Linux kernel dev
lists. All the major antispam effort mailing lists. The official
mailing lists for every major Unix-based MTA (sendmail, postfix, qmail,
exim). All the IETF working group mailing lists. The ICANN mailing
lists.


So, let's see: that covers most serious computer security
professionals, every major backbone operator and most peering and
transit providers, all of the major OSS lists, the preeminent mailing
list for the most prevalent commercial Unix, the majority of
spam-conscious and -activist postmasters, the operators of the majority
of the mail servers on the planet, and everyone involved in Internet
standards and governance.

Which lists were you referring to, exactly?

I...see.


--
Mark C. Langston Sr. Unix SysAdmin
mark (AT) bitshift (DOT) org mark (AT) seti (DOT) org
Systems & Network Admin SETI Institute
http://bitshift.org http://www.seti.org

Jason Snell
2003-12-05, 17:14
Mark C. Langston wrote:

>Which lists were you referring to, exactly?

Gee, the ones used by actual human beings?
--
Jason Snell / Editor in Chief, Macworld / jsnell (AT) macworld (DOT) com
415-243-3565 / AIM-iChat: MW jsnell

Mark C. Langston
2003-12-05, 17:45
On Fri, Dec 05, 2003 at 04:14:12PM -0800, Jason Snell wrote:
> Mark C. Langston wrote:
>
> >Which lists were you referring to, exactly?
>
> Gee, the ones used by actual human beings?


So, by implication, subscribers to the lists I just mentioned aren't
human beings?

Are you intentionally trying to alienate the majority of the technical
cognoscenti, or are you really just that polarizing on a daily basis?

If it weren't for those non-humans, you wouldn't have the convenience of
the email you're using to denigrate them.

--
Mark C. Langston Sr. Unix SysAdmin
mark (AT) bitshift (DOT) org mark (AT) seti (DOT) org
Systems & Network Admin SETI Institute
http://bitshift.org http://www.seti.org

Jason Snell
2003-12-05, 17:58
>So, by implication, subscribers to the lists I just mentioned aren't
>human beings?

So it turns out you don't have a sense of humor either.

>Are you intentionally trying to alienate the majority of the technical
>cognoscenti, or are you really just that polarizing on a daily basis?

Boy, you're quite an elitist.

You listed off a whole passel of unix sysadmin, open-source
developer, unix developer, and some high-level Internet standards
body lists. Those people are very powerful in the Internet community.
They also don't reflect the Internet community even remotely. They do
not reflect the way the Internet is used on a regular basis by
regular people discussing mundane matters like MP3 players or
favorite bands or TV shows or you-name-its.

You're right -- the unix sysadmins espouse the same things you, a
unix sysadmin, espouse! What a coincidence....

Your talking down to me and to Kevin shows me just how much you value
anyone who dares to disagree with you.

>Are you intentionally trying to alienate the majority of the technical
>cognoscenti,

All hail the technical cognoscenti! For they are the priests of our
new religion, the Internet, and without the Divine Right of Sysadmins
we should all perish from the earth!

>If it weren't for those non-humans, you wouldn't have the convenience of
>the email you're using to denigrate them.

You're right. I should kiss their feet. Because by disagreeing with
the keepers of the sacred keys, such as you, I am denigrating their
holy existence.

But hey, if you don't get it, you don't get it. After all, if NANOG-L
and the postfix mailing lists don't munge reply-to headers, it must
be the true and proper way of things!

Whatever.

-jason
--
Jason Snell / Editor in Chief, Macworld / jsnell (AT) macworld (DOT) com
415-243-3565 / AIM-iChat: MW jsnell

Mark C. Langston
2003-12-05, 18:35
On Fri, Dec 05, 2003 at 04:58:53PM -0800, Jason Snell wrote:
> >So, by implication, subscribers to the lists I just mentioned aren't
> >human beings?
>
> So it turns out you don't have a sense of humor either.

When you backhandedly and dismissively insult my profession, and that
of countless others? Nope. Not a whit.

>
> >Are you intentionally trying to alienate the majority of the technical
> >cognoscenti, or are you really just that polarizing on a daily basis?
>
> Boy, you're quite an elitist.
>

In much the same vein as saying, "exporting jobs overseas is bad for the
economy" is not racist, saying, "abiding by existing standards and
conventions whenever and wherever possible, unless there's a
demonstrably and overwhelmingly operational reason to do so" is not
elitist.

Perhaps you should shift your, "I must defend my $OS_OF_CHOICE" chip
to the arm opposite the one bearing your, "I must stick up for the
Little Guy(TM)" bleeding heart, and think this through for a second.

Reasons munging Reply-To: in transit is bad, beyond those already
mentioned here:

1) It's only a matter of time before personal and private information
is sent to the list accidentally.
2) Braindead virus filters WILL reply to this list. It's only a matter
of time.
3) Braindead spam filters WILL reply to this list. It's only a matter
of time.
4) Braindead vacation programs will create mail loops.
5) Braindead MTAs (IIS, Exchange, etc.) will create mail loops (and
they have a hard time recognizing and terminating said loops, creating a
headache for the list owner and overflowing mailboxes for subscribers).
6) Some spammer WILL joe-job the list address, creating all sorts of
mayhem for everyone subscribed; it's only a matter of time.
7) Some jerk WILL subscribe the list address to other mailing lists and
spam lists; it's only a matter of time.
8) Infected subsribers' computers WILL start spamming this list with
virus-laden messages that will irritate those of us not running
$COMPROMISED_OS, and potentially infect those that are. (Oh, and then
#2 above will kick in for added nuisance value. Enjoy!)


....and that's just off the top of my head. This list hasn't existed in
this form long enough to have experienced any of these problems yet.
But rest assured, it will.

In fact, I rather hope all of that comes to pass, just so I can
personally approach you, Jason, and see just how idealistic you remain
when the list is brought to its knees by such problems.


Is it *REALLY* worth the convenience of not having to exercise the two
braincells necessary to hit a different key or different button, or to
pay attention to the To: header before you send an email?

It may seem so now. I suspect your tune will change when the above
happens. I suspect the list will change when the above happens, much to
your chagrin.

You see, it's all hypothetical now, but each of the above WILL happen,
multiple times, unless and until the Reply-To: munging is removed. It's
inevitable -- ask anyone who's run a mailing list of any substance that
has had Reply-To: munged for any length of time. And when those things
do occur, I can assure you it will eventually surpass the level at which
the list owner is willing to tolerate (read: Require more time
perfoming list maintenance than the owner's willing/able to commit), and
the change will be made.


Some of us would rather just prevent all of the above from occurring in
the first place. AfterIn fact, I rather hope all of that comes to pass,
just so I can personally approach you, Jason, and see just how
idealistic you remain when the list is brought to its knees by such
problems.


Is it *REALLY* worth the convenience of not having to exercise the two
braincells necessary to hit a different key or different button, or to
pay attention to the To: header before you send an email?

It may seem so now. I suspect your tune will change when the above
happens. I suspect the list will change when the above happens, much to
your chagrin.

You see, it's all hypothetical now, but each of the above WILL happen,
multiple times, unless and until the Reply-To: munging is removed. It's
inevitable -- ask anyone who's run a mailing list of any substance that
has had Reply-To: munged. At some point, the headaches those things
create will surpass the tolerance of the list maintainer (read: will
require more time performing maintenance than the owner is willing or
able to commit), and the change will be forced upon you anyway.

Some of us would simply like to avoid those problems to begin with.
When I see a ball-peen hammer headed for my skull, I prefer to relocate
myself rather than wait for the hammer to strike repeatedly and then
decide how much I'm willing to put up with.



> You listed off a whole passel of unix sysadmin, open-source
> developer, unix developer, and some high-level Internet standards
> body lists. Those people are very powerful in the Internet community.
> They also don't reflect the Internet community even remotely. They do
> not reflect the way the Internet is used on a regular basis by
> regular people discussing mundane matters like MP3 players or
> favorite bands or TV shows or you-name-its.

The vast majority of anyone doing anything does not make that thing
right. It merely makes it popular. Correctness is circumscribed
entirely by the protocols and standards in use. Certain protocols and
standards allow unintended behavior. This is a two-edged sword: Some
unintended behaviors are beneficial -- ZeroConf's a good example of
building on unintended (or at least, unforseen) uses of certain
protocols.

Viruses and many network-based security holes are a good example of
malevolent unintended behaviors.


Undesirable unintended behaviors, as evolution has demonstrated
repeatedly, get weeded out. With our enormous branes, any fule kno
that smart people learn to recognize and avoid detrimental behaviors a
priori.

This, I suppose, means you're no fule. Congratulations.
(And before anyone goes off on a spelling flame tangent, pick up a book
sometime. it's a literary reference).

>
> Your talking down to me and to Kevin shows me just how much you value
> anyone who dares to disagree with you.
>

I value diagreement highly. It's how things get done, and get changed.
I have little patience for people who want to bang their head against
existing standards and insist things should be different because *they
say so*, without going through the appropriate processes to change them.

Rational people call the latter behavior "working within the system for
change". Rational people call the former behavior "idiotic, childish and
ultimately futile".


>
> All hail the technical cognoscenti! For they are the priests of our
> new religion, the Internet, and without the Divine Right of Sysadmins
> we should all perish from the earth!
>

You'd argue with physicists about whether the laws of inertia should
apply to you after a car crash, wouldn't you? Furrfu!


> >If it weren't for those non-humans, you wouldn't have the convenience of
> >the email you're using to denigrate them.
>
> You're right. I should kiss their feet. Because by disagreeing with
> the keepers of the sacred keys, such as you, I am denigrating their
> holy existence.
>

Present a workable alternative and demonstrate how it avoids the
problems with the existing, broken alternative. Then, perhaps, the
people who have so little patience for your mewling will instead listen
to and work with you.

Until then, as I mentioned to you earlier, you're simply pissing in the
breeze for the sheer pleasure of moistening your own pantleg. It
doesn't bother us per se, but don't be surprised when we point out how
silly and counterproductive it is.

I already suggested a feasible alternative: Reword the appropriate RFCs
to compel MUAs to handle Return-Path: when present, and with precedence
over Reply-To:. Then you can munge Reply-To: to your heart's content.


That's an operational solution requiring minimal changes to existing
protocols and standards and codebases. It neatly sidesteps the problems
that have been created by certain list owners ignoring the wisdom and
experience of other list owners who have been down the Reply-To: munging
path, and have learned the hard way exactly why and to what extent it's
a really bad and rather stupid idea.

But, rather than rally behind that solution, or propose one of your own,
you insist you be allowed to keep using a broken behavior, you insist
that those who know better (more often than not through ugly experience)
are not only wrong, but rude to suggest you might want to rethink your
position, and that you should be allowed to keep on doing what you're
doing, regardless of consequence.


That's just silly.


Oh, and by the way: The "average user" to whom you refer wouldn't know
a Reply-To: header from a Referrer: tag. The fact that "reply to" and
"reply to all/group" functions produce identical behavior with munged
..Reply-To:'s is a behavior instilled in the average user by misconfigured
software. It's not due to a groundswell of average users insisting that
..Reply-To:'s be munged.

And that will continue to be the case no matter how many times you
insist otherwise.


> But hey, if you don't get it, you don't get it. After all, if NANOG-L
> and the postfix mailing lists don't munge reply-to headers, it must
> be the true and proper way of things!
>

As insulting as that is, I hope you realize you've just stated, "just
because those with years upon years of running mailing lists, networks,
MTAs, and so forth, and whose hundreds upon hundreds of person-years of
experience have shown them that Certain Things Just Aren't Good
Ideas(TM)" have decided not to munge Reply-To:, doesn't mean we should
pay any attention to them!"

As I believe you said, "whatever".


When the list lay in burning ruins after repeated episodes of the
numbered items above, I hope you realize what your biased and selfish
approach has bought you.


--
Mark C. Langston Sr. Unix SysAdmin
mark (AT) bitshift (DOT) org mark (AT) seti (DOT) org
Systems & Network Admin SETI Institute
http://bitshift.org http://www.seti.org

jacobdp
2003-12-05, 19:01
On Fri, 5 Dec 2003 17:35:48 -0800, you wrote:
>When the list lay in burning ruins after repeated episodes of the
>numbered items above, I hope you realize what your biased and selfish
>approach has bought you.

Ummm....

When the list lays in burning ruins after repeated episodes of the
almost-flamewars above, I hope you BOTH realize what your hotheaded,
snide and sarcastic approaches have bought you.

For goodness sake, there's no reason to have a flame war over a
Reply-To: header!

- Jacob


--
"Terminak #3 has bad keyboard. Pkease fix."

Sean Graham
2003-12-05, 19:02
On Fri, Dec 05, 2003 at 05:35:48PM -0800, Mark C. Langston wrote:

> When you backhandedly and dismissively insult my profession, and that
> of countless others? Nope. Not a whit.

Could you ladies please take this catfight off list?

Thanks,
sean
--
Sean M. Graham - grahams (AT) gti (DOT) net
http://www.pobox.com/~grahams/

Mark C. Langston
2003-12-05, 19:27
On Fri, Dec 05, 2003 at 09:02:39PM -0500, Sean Graham wrote:
> On Fri, Dec 05, 2003 at 05:35:48PM -0800, Mark C. Langston wrote:
>
> > When you backhandedly and dismissively insult my profession, and that
> > of countless others? Nope. Not a whit.
>
> Could you ladies please take this catfight off list?
>
> Thanks,
> sean


Mark C. Langston is on vacation and won't be back until Monday, 12/8.

Apologies in advance to any mailing list owners who have
munged their Reply-To: header and are receiving this message, intended
for the original sender.

--
Mark C. Langston Sr. Unix SysAdmin
mark (AT) bitshift (DOT) org mark (AT) seti (DOT) org
Systems & Network Admin SETI Institute
http://bitshift.org http://www.seti.org

Jack Coates
2003-12-05, 22:27
On Fri, 2003-12-05 at 09:50, Caleb Epstein wrote:
....
> I am on many mailing lists, and the slimdevices ones are the
> ONLY ones I have seen that are configured this way.

so am I, and the majority set reply-to. The ones that don't, I don't
stay on long because they generate so much noise in my inbox.
Discussions stay on the mailing list, properly filtered into the folder
where I can go follow them when I'm ready.
--
Jack at Monkeynoodle Dot Org: It's A Scientific Venture...
************************************************** ********************
*"As I arose to slake my thirst, as I tried crawling from my bed, *
*I fell down flat -- I could not stagger, Nancy had me by the legs!" *
*-- Nancy Whiskey from The Snake by Shane McGowan and the Popes *
************************************************** ********************

Paul Warren
2003-12-06, 02:46
On Fri, Dec 05, 2003 at 10:47:07AM -0800, Jason Snell wrote:
> >Ever get that oh-no, not again feeling? I quote (unedited) a friend and
> >mailing list admin's response to this, with which i happen to agree:
>
> My favorite part of these documents is this:
>
> >Any reasonable, modern mailer provides this feature. I prefer the Elm mailer

This proves my point exactly. The fact that the author thinks it is
modern proves that this argument is as old as mailing lists themselves.

> or as your friend says:
>
> > people do not expect replies to go to the
> >list unless they have specifically asked for this
> >to happen by typing `g' or answering
>
> Uh, command-line unix mail readers are not remotely "modern." Arguing
> that reply-to headers shouldn't exist because we can all hit "g" in
> elm is crazy.

Sorry, that was taken out of context. This reply was written as a
response to a confirmed mutt user, hence the suggestion to use "g". The
fact that mutt is *is* a modern MUA is a completely different argument.

The point still stands. Without munged reply-to, "reply-to-all" in
your MUA replies to, err, all, whereas reply-to replies to the author.

> In fact, I don't believe I've seen a "group reply" command in any
> graphical e-mail client, at least not on the Mac. And if they are
> there, nobody seems to know about their existence.

Really? I'm not aware of a graphical e-mail client that *doesn't* have
this feature. Every one that I've ever used has an icon on the toolbar
just to the right of the "reply" button:

http://www.gnome.org/gnome-office/images/evolutionmailbig1.jpg

http://cws.internet.com/screenshots/outlookx1.html

http://www.sbutcher.pwp.blueyonder.co.uk/screenshots/mozilla2.html

http://cws.internet.com/screenshots/eudorap1.html

Mozilla is certainly available for Mac OS X. I couldn't see myself
using a mail client that does not offer this most basic of features, and
not just for mailing lists, but for engaging in the email discussions
required to do my work. That choice is essential and basic.

Paul

Ben Laurie
2003-12-06, 10:17
Mark C. Langston wrote:
> When the list lay in burning ruins after repeated episodes of the
> numbered items above, I hope you realize what your biased and selfish
> approach has bought you.

FWIW, most of the Apache mailing lists are set to reply-to the list, and
they've managed to last quite a while without collapsing.

Cheers,

Ben.

--
http://www.apache-ssl.org/ben.html http://www.thebunker.net/

"There is no limit to what a man can do or how far he can go if he
doesn't mind who gets the credit." - Robert Woodruff