![](https://secure.gravatar.com/avatar/920e6c5fc32991f18ad3ebaeff6ed8ad.jpg?s=120&d=mm&r=g)
On 18 Jun 2008, at 20:59, adsarebad-at-......... wrote:
At least the advantage of this solution is that you set the transport
for an entire domain to be this script, and then don’t need to have
the Mailman email addresses explicitly stated in your Postfix config.
If you want to have both regular and Mailman addresses for the same
domain, then I am afraid you will have to explicitly have the Mailman
addresses in your Postfix config. Just, they should go in the
appropriate section (addresses for virtual domains instead of aliases).
I am unsure what you actually wanted help with in your original
letter, i.e. was it a convenient way to do “newlist” after you setup
the multiple Mailman installs? Was it routing the different addresses
to the proper installs? etc.
![](https://secure.gravatar.com/avatar/36c620309a75f21b0a535d27e25d5c73.jpg?s=120&d=mm&r=g)
Well I finished my setup. I have a Multi-Mailman install set up and working. I have also wrote up a little writeup on how I did it. If someone who knows mailman would look at it and tell me if I have done something wrong that would be nice. Its at http://skagitattic.com/mailman.html
Thanks!
Also it seems clicking reply on a msg here does a off list reply. So all my previous reply`s were done off list.
Mark Sapiro wrote:
Thanks for the idea of editing Postfix.py! However I could not seem to make it work without.
fun@domain.com fun_domain@localhost
in virtual_mailman
Allan Odgaard wrote:
I was not able to get my mail to mailman without doing my own aliases.
Mark Sapiro wrote:
That is a good idea however I was done with my editing of Postfix.py by then. Now that its working I don`t want to spend the time to make a script to do what I want.
![](https://secure.gravatar.com/avatar/920e6c5fc32991f18ad3ebaeff6ed8ad.jpg?s=120&d=mm&r=g)
On 19 Jun 2008, at 08:38, adsarebad-at-......... wrote:
It is the default for new Mailman installs and the admin UI even has:
Where are replies to list messages directed?
Poster is strongly recommended for most mailing lists
I never understood why they have this recommendation or even default
setting and it has always bothered me when lists I subscribe to does
not change away from the default (fortunately majority of lists _do_
change it).
The problem is actually twofold:
- I routinely forget to reply to the list, and
- I get private replies which I _think_ are meant for the list, but
I feel bad about following up to the list incase the poster really
wanted it to be private (plus list members lose half the conversation
when only the list is cc’ed on every second letter).
So what exactly is the reason for this recommendation (and default)?
And where do we lobby for a change? :)
![](https://secure.gravatar.com/avatar/56f108518d7ee2544412cc80978e3182.jpg?s=120&d=mm&r=g)
Allan Odgaard wrote:
The reasoning, including links to a couple of articles pro and con on the subject, is explained on the page linked as "(Details for reply_goes_to_list)". You have said in another thread that these links contribute to excessive verbosity of the labels, but they often link to useful supplemental information.
[...]
So what exactly is the reason for this recommendation (and default)?
The principles of least harm and least surprise.
And where do we lobby for a change? :)
You can continue this thread, or you can just search the archives of this (and maybe the mailman-developers) list and see if you have something new to add.
Note that the Mailman developers have strong opinions about this, and it would take a new and compelling argument to change it.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
![](https://secure.gravatar.com/avatar/920e6c5fc32991f18ad3ebaeff6ed8ad.jpg?s=120&d=mm&r=g)
On 19 Jun 2008, at 15:53, Mark Sapiro wrote:
I don’t dispute that. I am sure there are lots of useful info there,
but it drowns in the excessive verbosity. Take the reply-to UI in
question:
Reply-To: header munging
Should any existing Reply-To: header
found in the original message be
stripped? If so, this will be done (o) No ( ) Yes
regardless of whether an explict
Reply-To: header is added by Mailman or
not. (Edit first_strip_reply_to)
Where are replies to list messages
directed? Poster is strongly recommended (o) Poster ( ) This list
for most mailing lists. (Details for ( ) Explicit address
reply_goes_to_list)
Explicit Reply-To: header. (Details for
[ ] reply_to_address)
That’s almost 100 words and the link that explains why the Poster
option is recommended bears the mundane title of “(Details for
reply_goes_to_list)” and is part of a label which is really 3 sentences.
Here is how I would change all of the above:
List replies go to: (o) Poster (recommended, more info)
( ) The list
[x] Strip existing Reply-To header
( ) Other: [ ]
The grouping should make it clear how the settings relate to each
other, so no need for long explanations about the explicit Reply-To
address etc.¹
Less text means better chance the user will read/grasp it. I only
added one “more info” link and put it right next to the ‘recommended’
text, so it should be clear from context that it will elaborate on why
this is the recommended setting.
¹ The third setting is dependent on list replies going to the list and
while it indicates that the first setting can be used even for replies
sent to poster, that does IMHO not make much sense.
![](https://secure.gravatar.com/avatar/56f108518d7ee2544412cc80978e3182.jpg?s=120&d=mm&r=g)
Allan Odgaard wrote:
I really don't want to get into a discussion of the web UI here. I am not directly involved in the changes that are going on, and I think the wiki <http://wiki.list.org/x/RoBE> and the mailman-developers list are more appropriate for this discussion.
However, I certainly agree that your suggested improvement above is more readable and user friendly except that it doesn't correctly capture the interrelated settings.
The first_strip_reply_to setting is independent and while it doesn't make sense for 'poster', it certainly does for 'other'
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
![](https://secure.gravatar.com/avatar/334b870d5b26878a79b2dc4cfcc500bc.jpg?s=120&d=mm&r=g)
Allan Odgaard writes:
[ repetitive lobbying removed ]
Please stop lobbying, period, and move these discussions off this list. We heard you the first time, you've been told the appropriate venue. Now this is just noise interfering with helping users with their everyday problems.
If you really want to push any of your proposals forward, submit patches. Nobody is going to preapprove a verbal description of a patch that you *might* write. (No offense to you, but my experience as a project leader is that about 2/3 of the patches that somebody is absolutely positively definitely going to write are never submitted.)
I have some comments on your proposals which I am absolutely positively definitely going to write assuming the proposals ever get submitted in an appropriate venue.
![](https://secure.gravatar.com/avatar/efd26422479c7fa3f0f080b07b6986bb.jpg?s=120&d=mm&r=g)
Allan Odgaard wrote:
See FAQ 3.48 at <http://wiki.list.org/x/44A9>.
There are far stronger arguments that have been made for your view, and they are countered by even stronger arguments on our side.
So what exactly is the reason for this recommendation (and default)?
See FAQ 3.48.
And where do we lobby for a change? :)
Everything that could possibly be argued on either side of this issue has already been said, years ago. There are no new arguments to be made.
Unless all of the current Mailman developers go away and all of the replacements feel the same way you do, this decision is not ever going to change.
-- Brad Knowles <brad@python.org> Member of the Python.org Postmaster Team & Co-Moderator of the mailman-users and mailman-developers mailing lists
![](https://secure.gravatar.com/avatar/efd26422479c7fa3f0f080b07b6986bb.jpg?s=120&d=mm&r=g)
adsarebad@gmail.com wrote:
Also it seems clicking reply on a msg here does a off list reply. So all my previous reply`s were done off list.
Correct. The Mailman developers feel that forcing all replies to go back to the list causes much more harm than good, see FAQ 3.48 at <http://wiki.list.org/x/44A9>.
-- Brad Knowles <brad@python.org> Member of the Python.org Postmaster Team & Co-Moderator of the mailman-users and mailman-developers mailing lists
![](https://secure.gravatar.com/avatar/374ba076bead84905266162abc728fb8.jpg?s=120&d=mm&r=g)
On 6/19/2008, Brad Knowles (brad@python.org) wrote:
Whats the big deal anyway? If you want lists configured to reply to the list, just set it that way. What difference does it make what the default is?
Until the vast majority of the most popular mail clients have a proper 'Reply-To-List' function (TBird doesn't, although it has an extension that tries to do it, it isn't very good at it) discussion lists should definitely (imnsho) be configured to reply to list, and I always change the ones I manage to do so.
--
Best regards,
Charles
![](https://secure.gravatar.com/avatar/efd26422479c7fa3f0f080b07b6986bb.jpg?s=120&d=mm&r=g)
Charles Marcus wrote:
The point is that there are lots of MUAs out there that are broken, and if you screw with the Reply-To: header, they are completely and totally unable to change who the reply is sent to. This is how private information gets exposed on public lists, with consequences ranging from just being personally embarassing, to getting you fired, to actually being life-threatening in some cases.
Do you really want to be responsible for something that could get someone killed?
Then you must not have read the FAQ I referenced.
-- Brad Knowles <brad@python.org> Member of the Python.org Postmaster Team & Co-Moderator of the mailman-users and mailman-developers mailing lists
![](https://secure.gravatar.com/avatar/920e6c5fc32991f18ad3ebaeff6ed8ad.jpg?s=120&d=mm&r=g)
For users who have their mail go through procmail the problem can be
fixed by a rule like the following. But in light of Brads letter, I
feel it is necessary to add a disclaimer saying that if this setting
gets you killed or fired, I cannot be held responsible! :p
:0 fw
* ^TO_mailman-users@python.org|^List-id: .*<mailman-
users.python.org> | formail -i "Reply-To: mailman-users@python.org"
A more fancy solution could extract the list address from the list- post header (and target all list letters w/o a reply-to).
I use an anonymous remailer so personally wouldn’t work for me, plus
when people reply to me and cc the list, there are no list headers in
that message, even though it is a list message (one of the reasons
“reply to poster” sucks).
Btw: does anyone know any of these many MUAs that Brad speak of? I
can’t help but wonder how users of these can participate in list
conversation with Mailman’s default setting, given their inability to
control where their letters are sent to.
On 19 Jun 2008, at 23:23, Brad Knowles wrote:
![](https://secure.gravatar.com/avatar/374ba076bead84905266162abc728fb8.jpg?s=120&d=mm&r=g)
You replied on list, so I will too... but I will not respond to further arguments about it, since neither of us are likely to change our mind.
On 6/19/2008, Brad Knowles (brad@python.org) wrote:
The point is that there are lots of MUAs out there that are broken,
I didn't intend to open up this debate, because I *have* read - and *understand* the arguments on both sides. *Ideally*, I agree with you and the Mailman devs... *realistically*, though, most discussion lists - for most people, using the most popular mail clients - operate much more smoothly when Reply-to munging is implemented.
*My* point was simply pointing out that there *is a preference setting* in the Mailman GUI for changing this, so if someone *wants* to change it, they obviously *can* - so what difference does the *default* make?
and if you screw with the Reply-To: header, they are completely and totally unable to change who the reply is sent to.
Sorry, but this isn't true for any mail client I've ever used... ever heard of copy/cut/paste? Yeah, it requires some manual labor, rather than clicking a button, but it can still be done.
In fact, I must do this a little of this when participating *on this list*. First, I have to hit 'Reply All' to get the list address in the CC field, then I change the 'CC' to 'To', then *delete* the *posters* email address so that they don't get a duplicate - which, by the way, I notice you don't have the courtesy to do.
I wouldn't have to do any of that, if the Reply-to was set to the list.
But without being specific - like, what specific MUA's have this problem
- its kind of hard to argue.
If someone is dumb enough to send information of such a nature without actually *looking* at where it is going, then yeah, they might actually be required to pay the consequences...
Lol! Thanks, I needed that...
Then you must not have read the FAQ I referenced.
I did... I just believe that it is *ideally* correct, but *realistically* incorrect, due to the reality of limitations in both the most popular mail clients *and* the *behavior* of most people on discussion lists (they don't know the difference, and more importantly *don't care*).
--
Best regards,
Charles
![](https://secure.gravatar.com/avatar/56f108518d7ee2544412cc80978e3182.jpg?s=120&d=mm&r=g)
Charles Marcus wrote:
There is also an mm_cfg.py setting to override the default for new lists, but the fact remains that many people, even competent people, are busy and don't have time to research every setting and option, so they tend to go with defaults until there is a problem. That's why the choice of default is important to many people.
<snip>
It is not a matter of courtesy. There are many arguments for not editing out the addiditional recipients including: that person in the loop.
- one or more may be a reply-to that is not a list member address but an address the poster wants the mail to go to, and similarly for Ccs.
- the poster may be a digest member and Ccing back to the poster keeps
Note that Mailman will remove addresses from Cc: if they are list members with 'nodupes' set.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
![](https://secure.gravatar.com/avatar/efd26422479c7fa3f0f080b07b6986bb.jpg?s=120&d=mm&r=g)
Charles Marcus wrote:
You replied on list, so I will too...
Not really a very wise idea. When I use my @python.org e-mail address (which I very rarely do), I speak from a position of a certain amount of authority on the subject.
If you want to try to have a civilized private conversation with us on the subject, we can probably find a way to do that.
But if you continue to aggressively and publicly challenge us on our own mailing list over a topics that were considered dead and buried years ago, the only likely outcomes are ones that you probably will not like.
In your experience. And how many decades have you been specializing in Internet e-mail systems administration?
How many of the screwed-up MUAs that we talk about have you actually encountered? None?
So, it is out of ignorance that you say everyone should completely and totally ignore our advice?
Do you really want to use ignorance as your only defense?
The default makes a huge difference. Ask Microsoft.
In my experience, over 90% never change the defaults on the software they install, whether you're talking about our mailing list management software or anything else.
If those defaults are not secure, then they are not secure -- again, go talk to Microsoft.
There are lots of admins out there who are capable of doing an "apt-get mailman" (or whatever), and not much beyond that. Therefore, we have to be extra careful in terms of what is enabled or disabled by default.
There are battered womens shelters who use our software, and some of those battered women literally do have stalkers coming after them. There are dissident groups in authoritarian countries that use our software, and some of those dissident groups really do have the secret police coming after them.
These kinds of things are always in the backs of our minds as we develop and maintain our software, and while we won't necessarily leave out certain features of our software just because it could be dangerous if mis-configured, we certainly do keep in mind the fact that we should probably ship those features disabled by default.
But as strongly as we hold our views that Reply-To: munging should not be done, we do acknowledge that there are certain limited circumstances where it might potentially be acceptable to do this kind of munging -- like when a company is running an internal discussion list and they want to force all their employees to keep all replies on the list. And if those employees screw up and post sensitive private information on the list, then the only thing at risk is their jobs, and the jobs of any others who might have also been accidentally exposed.
So, we allow people like you to choose to configure your software differently. Why can you not accept that we choose to configure the default for this option to be disabled?
Do you really want to take that fascist approach with us, where we hold an opinion but we allow you to speak, but you do not in return allow us the same courtesy?
Just because you have not encountered something does not mean it does not exist. There are those of us who've actually been around for a while who have seen all sorts of seriously weird crap.
And your claim that you've never run into this weird crap is not proof that this weird crap does not and cannot exist.
Then you be the one to try to explain that to their surviving family members. Go back to the top of this response and re-read the part where I talked about who some of our customers are for this software.
And no, as much as I might like to, I'm not going to provide those family members with firearms, so that they can demonstrate to you the danger that their dear departed one faced.
I did... I just believe that it is *ideally* correct, but *realistically* incorrect,
I'm glad you feel privileged that you are the only person on the planet who should be allowed to define what reality is.
-- Brad Knowles <brad@python.org> Member of the Python.org Postmaster Team & Co-Moderator of the mailman-users and mailman-developers mailing lists
![](https://secure.gravatar.com/avatar/374ba076bead84905266162abc728fb8.jpg?s=120&d=mm&r=g)
On 6/20/2008 1:27 PM, Brad Knowles wrote:
Charles Marcus wrote:
You replied on list, so I will too...
Not really a very wise idea.
Really. Well, I had started to reply point by point until it became obvious that you are confused about something...
Read the following carefully...
I am *not* arguing for changing the default - whatever gave you that idea?????
In fact, I was pointing out to the person that *was* arguing for changing it that, since the devs had graciously provided the ability to *change* this setting, that the default *did not really matter*.
Are you seriously suggesting that the mere fact that you posted from an @python.org address is supposed to mean something to the average list participant here? I rarely even look at the original posters address, and even if I had noticed it, it wouldn't have meant anything to me other than you have an email address @python.org.
That said, if this really was a subtle way for you to intend to be speaking from a position of authority, then maybe you should read things a little more carefully before you threaten someone on a public list from such a position.
Sheesh.
--
Best regards,
Charles
![](https://secure.gravatar.com/avatar/efd26422479c7fa3f0f080b07b6986bb.jpg?s=120&d=mm&r=g)
Charles Marcus wrote:
I am *not* arguing for changing the default - whatever gave you that idea?????
You may say that, but then virtually everything else you wrote was in strong, even violent, support of changing the default.
So, what am I supposed to believe -- the tiny little disclaimer, or the vast bulk of what you write?
And you don't ever look at the signature block, either?
Do you drive on the roads like you drive your computer? By not looking or paying any attention whatsoever to any of the signs or signals? Does your insurance company know about this?
It wasn't a threat. It was a statement of fact -- continued arguing with us on a subject considered long since dead is likely to have results that you're probably not going to like.
And that's the last I'm going to say on this subject. Anything further on this subject will be action, and no words involved.
-- Brad Knowles <brad@python.org> Member of the Python.org Postmaster Team & Co-Moderator of the mailman-users and mailman-developers mailing lists
![](https://secure.gravatar.com/avatar/374ba076bead84905266162abc728fb8.jpg?s=120&d=mm&r=g)
On 6/20/2008, Brad Knowles (brad@python.org) wrote:
I was talking about your cute little remark about giving someone a FIREARM so that they could 'come demonstrate to me the danger their deear departed one faced'.
If thats not a threat of VIOLENCE then I don't know what is.
That chip on your shoulder is getting pretty heavy, isn't it?
--
Best regards,
Charles
![](https://secure.gravatar.com/avatar/efd26422479c7fa3f0f080b07b6986bb.jpg?s=120&d=mm&r=g)
Charles Marcus wrote:
I said I wouldn't do it. So there's no threat there.
That chip on your shoulder is getting pretty heavy, isn't it?
Ad hominem attacks are not going to get you anywhere, at least not anywhere you like.
-- Brad Knowles <brad@python.org> Member of the Python.org Postmaster Team & Co-Moderator of the mailman-users and mailman-developers mailing lists
![](https://secure.gravatar.com/avatar/b476c22cb2e32de59459a302eea18a45.jpg?s=120&d=mm&r=g)
Brad Knowles wrote:
For a real-world example: We run several thousand lists here at our university (not all in Mailman) and we had a class list set up with "replies go to list" (the professor/owner had inadvertantly changed that setting from the default). One night, the professor sent out a message to the class list, giving them some instructions or something. Well one student on the list didn't realize this and thought the prof had sent the message only to her, basically targetting her (I guess she didn't look at the headers). She replied with a pretty personal message back to the prof, and it went to the entire class (180 students). She was so embarrassed and humiliated the next day to know that entire class had read her personal message and were talking about her, that she dropped the class.
This is basically why you want people to specifically address the list, if they mean to send to it. Sure it was her fault, but it's wiser to use the "strongly recommended" defaults than to try to teach a revolving group of 15,000 students an email lesson.
- jim -
![](https://secure.gravatar.com/avatar/36c620309a75f21b0a535d27e25d5c73.jpg?s=120&d=mm&r=g)
adsarebad@gmail.com wrote:
I just wanted to say I did not mean to start a header war here. What I meant was I should have looked more closely where my reply was going and fixed it. I had not really even thought about the question of changing headers until now as my only lists are announcement only, so the reply should always go back to the sender. Also a quick question to the side the believes we should not do header mangling. After reading the FAQ and the links it pointed to, I noticed the counter reply talked about RFC 822. If that document defines email standards then isn't that the correct way to do it?
Anyway sorry about starting this whole thing.
![](https://secure.gravatar.com/avatar/36c620309a75f21b0a535d27e25d5c73.jpg?s=120&d=mm&r=g)
Well I finished my setup. I have a Multi-Mailman install set up and working. I have also wrote up a little writeup on how I did it. If someone who knows mailman would look at it and tell me if I have done something wrong that would be nice. Its at http://skagitattic.com/mailman.html
Thanks!
Also it seems clicking reply on a msg here does a off list reply. So all my previous reply`s were done off list.
Mark Sapiro wrote:
Thanks for the idea of editing Postfix.py! However I could not seem to make it work without.
fun@domain.com fun_domain@localhost
in virtual_mailman
Allan Odgaard wrote:
I was not able to get my mail to mailman without doing my own aliases.
Mark Sapiro wrote:
That is a good idea however I was done with my editing of Postfix.py by then. Now that its working I don`t want to spend the time to make a script to do what I want.
![](https://secure.gravatar.com/avatar/920e6c5fc32991f18ad3ebaeff6ed8ad.jpg?s=120&d=mm&r=g)
On 19 Jun 2008, at 08:38, adsarebad-at-......... wrote:
It is the default for new Mailman installs and the admin UI even has:
Where are replies to list messages directed?
Poster is strongly recommended for most mailing lists
I never understood why they have this recommendation or even default
setting and it has always bothered me when lists I subscribe to does
not change away from the default (fortunately majority of lists _do_
change it).
The problem is actually twofold:
- I routinely forget to reply to the list, and
- I get private replies which I _think_ are meant for the list, but
I feel bad about following up to the list incase the poster really
wanted it to be private (plus list members lose half the conversation
when only the list is cc’ed on every second letter).
So what exactly is the reason for this recommendation (and default)?
And where do we lobby for a change? :)
![](https://secure.gravatar.com/avatar/56f108518d7ee2544412cc80978e3182.jpg?s=120&d=mm&r=g)
Allan Odgaard wrote:
The reasoning, including links to a couple of articles pro and con on the subject, is explained on the page linked as "(Details for reply_goes_to_list)". You have said in another thread that these links contribute to excessive verbosity of the labels, but they often link to useful supplemental information.
[...]
So what exactly is the reason for this recommendation (and default)?
The principles of least harm and least surprise.
And where do we lobby for a change? :)
You can continue this thread, or you can just search the archives of this (and maybe the mailman-developers) list and see if you have something new to add.
Note that the Mailman developers have strong opinions about this, and it would take a new and compelling argument to change it.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
![](https://secure.gravatar.com/avatar/920e6c5fc32991f18ad3ebaeff6ed8ad.jpg?s=120&d=mm&r=g)
On 19 Jun 2008, at 15:53, Mark Sapiro wrote:
I don’t dispute that. I am sure there are lots of useful info there,
but it drowns in the excessive verbosity. Take the reply-to UI in
question:
Reply-To: header munging
Should any existing Reply-To: header
found in the original message be
stripped? If so, this will be done (o) No ( ) Yes
regardless of whether an explict
Reply-To: header is added by Mailman or
not. (Edit first_strip_reply_to)
Where are replies to list messages
directed? Poster is strongly recommended (o) Poster ( ) This list
for most mailing lists. (Details for ( ) Explicit address
reply_goes_to_list)
Explicit Reply-To: header. (Details for
[ ] reply_to_address)
That’s almost 100 words and the link that explains why the Poster
option is recommended bears the mundane title of “(Details for
reply_goes_to_list)” and is part of a label which is really 3 sentences.
Here is how I would change all of the above:
List replies go to: (o) Poster (recommended, more info)
( ) The list
[x] Strip existing Reply-To header
( ) Other: [ ]
The grouping should make it clear how the settings relate to each
other, so no need for long explanations about the explicit Reply-To
address etc.¹
Less text means better chance the user will read/grasp it. I only
added one “more info” link and put it right next to the ‘recommended’
text, so it should be clear from context that it will elaborate on why
this is the recommended setting.
¹ The third setting is dependent on list replies going to the list and
while it indicates that the first setting can be used even for replies
sent to poster, that does IMHO not make much sense.
![](https://secure.gravatar.com/avatar/56f108518d7ee2544412cc80978e3182.jpg?s=120&d=mm&r=g)
Allan Odgaard wrote:
I really don't want to get into a discussion of the web UI here. I am not directly involved in the changes that are going on, and I think the wiki <http://wiki.list.org/x/RoBE> and the mailman-developers list are more appropriate for this discussion.
However, I certainly agree that your suggested improvement above is more readable and user friendly except that it doesn't correctly capture the interrelated settings.
The first_strip_reply_to setting is independent and while it doesn't make sense for 'poster', it certainly does for 'other'
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
![](https://secure.gravatar.com/avatar/334b870d5b26878a79b2dc4cfcc500bc.jpg?s=120&d=mm&r=g)
Allan Odgaard writes:
[ repetitive lobbying removed ]
Please stop lobbying, period, and move these discussions off this list. We heard you the first time, you've been told the appropriate venue. Now this is just noise interfering with helping users with their everyday problems.
If you really want to push any of your proposals forward, submit patches. Nobody is going to preapprove a verbal description of a patch that you *might* write. (No offense to you, but my experience as a project leader is that about 2/3 of the patches that somebody is absolutely positively definitely going to write are never submitted.)
I have some comments on your proposals which I am absolutely positively definitely going to write assuming the proposals ever get submitted in an appropriate venue.
![](https://secure.gravatar.com/avatar/efd26422479c7fa3f0f080b07b6986bb.jpg?s=120&d=mm&r=g)
Allan Odgaard wrote:
See FAQ 3.48 at <http://wiki.list.org/x/44A9>.
There are far stronger arguments that have been made for your view, and they are countered by even stronger arguments on our side.
So what exactly is the reason for this recommendation (and default)?
See FAQ 3.48.
And where do we lobby for a change? :)
Everything that could possibly be argued on either side of this issue has already been said, years ago. There are no new arguments to be made.
Unless all of the current Mailman developers go away and all of the replacements feel the same way you do, this decision is not ever going to change.
-- Brad Knowles <brad@python.org> Member of the Python.org Postmaster Team & Co-Moderator of the mailman-users and mailman-developers mailing lists
![](https://secure.gravatar.com/avatar/efd26422479c7fa3f0f080b07b6986bb.jpg?s=120&d=mm&r=g)
adsarebad@gmail.com wrote:
Also it seems clicking reply on a msg here does a off list reply. So all my previous reply`s were done off list.
Correct. The Mailman developers feel that forcing all replies to go back to the list causes much more harm than good, see FAQ 3.48 at <http://wiki.list.org/x/44A9>.
-- Brad Knowles <brad@python.org> Member of the Python.org Postmaster Team & Co-Moderator of the mailman-users and mailman-developers mailing lists
![](https://secure.gravatar.com/avatar/374ba076bead84905266162abc728fb8.jpg?s=120&d=mm&r=g)
On 6/19/2008, Brad Knowles (brad@python.org) wrote:
Whats the big deal anyway? If you want lists configured to reply to the list, just set it that way. What difference does it make what the default is?
Until the vast majority of the most popular mail clients have a proper 'Reply-To-List' function (TBird doesn't, although it has an extension that tries to do it, it isn't very good at it) discussion lists should definitely (imnsho) be configured to reply to list, and I always change the ones I manage to do so.
--
Best regards,
Charles
![](https://secure.gravatar.com/avatar/efd26422479c7fa3f0f080b07b6986bb.jpg?s=120&d=mm&r=g)
Charles Marcus wrote:
The point is that there are lots of MUAs out there that are broken, and if you screw with the Reply-To: header, they are completely and totally unable to change who the reply is sent to. This is how private information gets exposed on public lists, with consequences ranging from just being personally embarassing, to getting you fired, to actually being life-threatening in some cases.
Do you really want to be responsible for something that could get someone killed?
Then you must not have read the FAQ I referenced.
-- Brad Knowles <brad@python.org> Member of the Python.org Postmaster Team & Co-Moderator of the mailman-users and mailman-developers mailing lists
![](https://secure.gravatar.com/avatar/920e6c5fc32991f18ad3ebaeff6ed8ad.jpg?s=120&d=mm&r=g)
For users who have their mail go through procmail the problem can be
fixed by a rule like the following. But in light of Brads letter, I
feel it is necessary to add a disclaimer saying that if this setting
gets you killed or fired, I cannot be held responsible! :p
:0 fw
* ^TO_mailman-users@python.org|^List-id: .*<mailman-
users.python.org> | formail -i "Reply-To: mailman-users@python.org"
A more fancy solution could extract the list address from the list- post header (and target all list letters w/o a reply-to).
I use an anonymous remailer so personally wouldn’t work for me, plus
when people reply to me and cc the list, there are no list headers in
that message, even though it is a list message (one of the reasons
“reply to poster” sucks).
Btw: does anyone know any of these many MUAs that Brad speak of? I
can’t help but wonder how users of these can participate in list
conversation with Mailman’s default setting, given their inability to
control where their letters are sent to.
On 19 Jun 2008, at 23:23, Brad Knowles wrote:
![](https://secure.gravatar.com/avatar/374ba076bead84905266162abc728fb8.jpg?s=120&d=mm&r=g)
You replied on list, so I will too... but I will not respond to further arguments about it, since neither of us are likely to change our mind.
On 6/19/2008, Brad Knowles (brad@python.org) wrote:
The point is that there are lots of MUAs out there that are broken,
I didn't intend to open up this debate, because I *have* read - and *understand* the arguments on both sides. *Ideally*, I agree with you and the Mailman devs... *realistically*, though, most discussion lists - for most people, using the most popular mail clients - operate much more smoothly when Reply-to munging is implemented.
*My* point was simply pointing out that there *is a preference setting* in the Mailman GUI for changing this, so if someone *wants* to change it, they obviously *can* - so what difference does the *default* make?
and if you screw with the Reply-To: header, they are completely and totally unable to change who the reply is sent to.
Sorry, but this isn't true for any mail client I've ever used... ever heard of copy/cut/paste? Yeah, it requires some manual labor, rather than clicking a button, but it can still be done.
In fact, I must do this a little of this when participating *on this list*. First, I have to hit 'Reply All' to get the list address in the CC field, then I change the 'CC' to 'To', then *delete* the *posters* email address so that they don't get a duplicate - which, by the way, I notice you don't have the courtesy to do.
I wouldn't have to do any of that, if the Reply-to was set to the list.
But without being specific - like, what specific MUA's have this problem
- its kind of hard to argue.
If someone is dumb enough to send information of such a nature without actually *looking* at where it is going, then yeah, they might actually be required to pay the consequences...
Lol! Thanks, I needed that...
Then you must not have read the FAQ I referenced.
I did... I just believe that it is *ideally* correct, but *realistically* incorrect, due to the reality of limitations in both the most popular mail clients *and* the *behavior* of most people on discussion lists (they don't know the difference, and more importantly *don't care*).
--
Best regards,
Charles
![](https://secure.gravatar.com/avatar/56f108518d7ee2544412cc80978e3182.jpg?s=120&d=mm&r=g)
Charles Marcus wrote:
There is also an mm_cfg.py setting to override the default for new lists, but the fact remains that many people, even competent people, are busy and don't have time to research every setting and option, so they tend to go with defaults until there is a problem. That's why the choice of default is important to many people.
<snip>
It is not a matter of courtesy. There are many arguments for not editing out the addiditional recipients including: that person in the loop.
- one or more may be a reply-to that is not a list member address but an address the poster wants the mail to go to, and similarly for Ccs.
- the poster may be a digest member and Ccing back to the poster keeps
Note that Mailman will remove addresses from Cc: if they are list members with 'nodupes' set.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
![](https://secure.gravatar.com/avatar/efd26422479c7fa3f0f080b07b6986bb.jpg?s=120&d=mm&r=g)
Charles Marcus wrote:
You replied on list, so I will too...
Not really a very wise idea. When I use my @python.org e-mail address (which I very rarely do), I speak from a position of a certain amount of authority on the subject.
If you want to try to have a civilized private conversation with us on the subject, we can probably find a way to do that.
But if you continue to aggressively and publicly challenge us on our own mailing list over a topics that were considered dead and buried years ago, the only likely outcomes are ones that you probably will not like.
In your experience. And how many decades have you been specializing in Internet e-mail systems administration?
How many of the screwed-up MUAs that we talk about have you actually encountered? None?
So, it is out of ignorance that you say everyone should completely and totally ignore our advice?
Do you really want to use ignorance as your only defense?
The default makes a huge difference. Ask Microsoft.
In my experience, over 90% never change the defaults on the software they install, whether you're talking about our mailing list management software or anything else.
If those defaults are not secure, then they are not secure -- again, go talk to Microsoft.
There are lots of admins out there who are capable of doing an "apt-get mailman" (or whatever), and not much beyond that. Therefore, we have to be extra careful in terms of what is enabled or disabled by default.
There are battered womens shelters who use our software, and some of those battered women literally do have stalkers coming after them. There are dissident groups in authoritarian countries that use our software, and some of those dissident groups really do have the secret police coming after them.
These kinds of things are always in the backs of our minds as we develop and maintain our software, and while we won't necessarily leave out certain features of our software just because it could be dangerous if mis-configured, we certainly do keep in mind the fact that we should probably ship those features disabled by default.
But as strongly as we hold our views that Reply-To: munging should not be done, we do acknowledge that there are certain limited circumstances where it might potentially be acceptable to do this kind of munging -- like when a company is running an internal discussion list and they want to force all their employees to keep all replies on the list. And if those employees screw up and post sensitive private information on the list, then the only thing at risk is their jobs, and the jobs of any others who might have also been accidentally exposed.
So, we allow people like you to choose to configure your software differently. Why can you not accept that we choose to configure the default for this option to be disabled?
Do you really want to take that fascist approach with us, where we hold an opinion but we allow you to speak, but you do not in return allow us the same courtesy?
Just because you have not encountered something does not mean it does not exist. There are those of us who've actually been around for a while who have seen all sorts of seriously weird crap.
And your claim that you've never run into this weird crap is not proof that this weird crap does not and cannot exist.
Then you be the one to try to explain that to their surviving family members. Go back to the top of this response and re-read the part where I talked about who some of our customers are for this software.
And no, as much as I might like to, I'm not going to provide those family members with firearms, so that they can demonstrate to you the danger that their dear departed one faced.
I did... I just believe that it is *ideally* correct, but *realistically* incorrect,
I'm glad you feel privileged that you are the only person on the planet who should be allowed to define what reality is.
-- Brad Knowles <brad@python.org> Member of the Python.org Postmaster Team & Co-Moderator of the mailman-users and mailman-developers mailing lists
![](https://secure.gravatar.com/avatar/374ba076bead84905266162abc728fb8.jpg?s=120&d=mm&r=g)
On 6/20/2008 1:27 PM, Brad Knowles wrote:
Charles Marcus wrote:
You replied on list, so I will too...
Not really a very wise idea.
Really. Well, I had started to reply point by point until it became obvious that you are confused about something...
Read the following carefully...
I am *not* arguing for changing the default - whatever gave you that idea?????
In fact, I was pointing out to the person that *was* arguing for changing it that, since the devs had graciously provided the ability to *change* this setting, that the default *did not really matter*.
Are you seriously suggesting that the mere fact that you posted from an @python.org address is supposed to mean something to the average list participant here? I rarely even look at the original posters address, and even if I had noticed it, it wouldn't have meant anything to me other than you have an email address @python.org.
That said, if this really was a subtle way for you to intend to be speaking from a position of authority, then maybe you should read things a little more carefully before you threaten someone on a public list from such a position.
Sheesh.
--
Best regards,
Charles
![](https://secure.gravatar.com/avatar/efd26422479c7fa3f0f080b07b6986bb.jpg?s=120&d=mm&r=g)
Charles Marcus wrote:
I am *not* arguing for changing the default - whatever gave you that idea?????
You may say that, but then virtually everything else you wrote was in strong, even violent, support of changing the default.
So, what am I supposed to believe -- the tiny little disclaimer, or the vast bulk of what you write?
And you don't ever look at the signature block, either?
Do you drive on the roads like you drive your computer? By not looking or paying any attention whatsoever to any of the signs or signals? Does your insurance company know about this?
It wasn't a threat. It was a statement of fact -- continued arguing with us on a subject considered long since dead is likely to have results that you're probably not going to like.
And that's the last I'm going to say on this subject. Anything further on this subject will be action, and no words involved.
-- Brad Knowles <brad@python.org> Member of the Python.org Postmaster Team & Co-Moderator of the mailman-users and mailman-developers mailing lists
![](https://secure.gravatar.com/avatar/374ba076bead84905266162abc728fb8.jpg?s=120&d=mm&r=g)
On 6/20/2008, Brad Knowles (brad@python.org) wrote:
I was talking about your cute little remark about giving someone a FIREARM so that they could 'come demonstrate to me the danger their deear departed one faced'.
If thats not a threat of VIOLENCE then I don't know what is.
That chip on your shoulder is getting pretty heavy, isn't it?
--
Best regards,
Charles
![](https://secure.gravatar.com/avatar/efd26422479c7fa3f0f080b07b6986bb.jpg?s=120&d=mm&r=g)
Charles Marcus wrote:
I said I wouldn't do it. So there's no threat there.
That chip on your shoulder is getting pretty heavy, isn't it?
Ad hominem attacks are not going to get you anywhere, at least not anywhere you like.
-- Brad Knowles <brad@python.org> Member of the Python.org Postmaster Team & Co-Moderator of the mailman-users and mailman-developers mailing lists
![](https://secure.gravatar.com/avatar/b476c22cb2e32de59459a302eea18a45.jpg?s=120&d=mm&r=g)
Brad Knowles wrote:
For a real-world example: We run several thousand lists here at our university (not all in Mailman) and we had a class list set up with "replies go to list" (the professor/owner had inadvertantly changed that setting from the default). One night, the professor sent out a message to the class list, giving them some instructions or something. Well one student on the list didn't realize this and thought the prof had sent the message only to her, basically targetting her (I guess she didn't look at the headers). She replied with a pretty personal message back to the prof, and it went to the entire class (180 students). She was so embarrassed and humiliated the next day to know that entire class had read her personal message and were talking about her, that she dropped the class.
This is basically why you want people to specifically address the list, if they mean to send to it. Sure it was her fault, but it's wiser to use the "strongly recommended" defaults than to try to teach a revolving group of 15,000 students an email lesson.
- jim -
![](https://secure.gravatar.com/avatar/36c620309a75f21b0a535d27e25d5c73.jpg?s=120&d=mm&r=g)
adsarebad@gmail.com wrote:
I just wanted to say I did not mean to start a header war here. What I meant was I should have looked more closely where my reply was going and fixed it. I had not really even thought about the question of changing headers until now as my only lists are announcement only, so the reply should always go back to the sender. Also a quick question to the side the believes we should not do header mangling. After reading the FAQ and the links it pointed to, I noticed the counter reply talked about RFC 822. If that document defines email standards then isn't that the correct way to do it?
Anyway sorry about starting this whole thing.
participants (7)
-
adsarebad@gmail.com
-
Allan Odgaard
-
Brad Knowles
-
Charles Marcus
-
Mark Sapiro
-
Savoy, Jim
-
Stephen J. Turnbull