Hi All
I've just been cleaning up a list that I administer and suddenly it came to me that we could have a report Scam/Spam button for such items being sent to Mailman lists.
Would this be feasible? I know gmail has such a feature - I don't quite know the process it follows once it is reported.
Where I see the biggest advantage of this is that individual mail users are not going to / seldom do go to the trouble of reporting because it is too much trouble. Do be able to do so at the click of a button, would exponentially increase the number of senders being reported.
Regards,
Rui
Rui Correia Advocacy, Media and Language Consultant 38 Finch St, Ontdekkers Park, Roodepoort, Johannesburg, South Africa Tel/ Fax (+27-11) 766-4336 Cell (+27) (0) 83-368-1214
I've just been cleaning up a list that I administer and suddenly it came to me that we could have a report Scam/Spam button for such items being sent to Mailman lists.
Where would you have this button? In the admindb interface? In the post itself somewhere?
Would this be feasible? I know gmail has such a feature - I don't quite know the process it follows once it is reported.
Ay. There's the rub. To whom would it be reported, and what might they do with it once reported?
Where I see the biggest advantage of this is that individual mail users are not going to / seldom do go to the trouble of reporting because it is too much trouble. Do be able to do so at the click of a button, would exponentially increase the number of senders being reported.
Yes, but as with all distributed effort spam control measures, who is going to police the mob-ocracy?
A good idea, to be sure. But there remain many sticky details to iron out.
-jag
-- Joshua Ginsberg <jag@fsf.org> Free Software Foundation - Senior Systems Administrator
On Tue, Dec 13, 2005 at 12:55:30PM -0500, Joshua Ginsberg wrote:
Would this be feasible? I know gmail has such a feature - I don't quite know the process it follows once it is reported.
Ay. There's the rub. To whom would it be reported, and what might they do with it once reported?
Perhaps spamcop? This can be done simply by forwarding the offending spam as an attachment to a spamcop email address that you are given for submissions.
Regards, Msquared...
At 2:33 PM +0200 2005-12-13, Rui Correia wrote:
I've just been cleaning up a list that I administer and suddenly it came to me that we could have a report Scam/Spam button for such items being sent to Mailman lists.
Where would this button exist? Who would use it, and how?Where I see the biggest advantage of this is that individual mail users are not going to / seldom do go to the trouble of reporting because it is too much trouble. Do be able to do so at the click of a button, would exponentially increase the number of senders being reported.
Mailman does not provide the end-user MUA functionality. It is not possible for Mailman to pop up a button in Exchange saying "Report this message as spam?", nor is it possible to support all MUAs in this manner.
You could add a new "List-" header which would allow compliant MUAs to give users a single button to click on, but there are currently no MUAs that would support such a feature. Moreover, to whom would they report the message as spam -- the mail servers operated by their own providers, or the mail server(s) operated by the mailing list host? How do you prevent abuse of such a button by an attacker?
There are a lot of questions you'd have to answer before you could come up with a reasonable scheme for creating a new "List-" header that would perform the desired function while avoiding the creation of any new major weaknesses.
-- Brad Knowles, <brad@stop.mail-abuse.org>
"Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety."
-- Benjamin Franklin (1706-1790), reply of the Pennsylvania
Assembly to the Governor, November 11, 1755LOPSA member since December 2005. See <http://www.lopsa.org/>.
At 7:38 PM -0600 2005-12-13, Brad Knowles wrote:
At 2:33 PM +0200 2005-12-13, Rui Correia wrote:
I've just been cleaning up a list that I administer and suddenly it came to me that we could have a report Scam/Spam button for such items being sent to Mailman lists.
Where would this button exist? Who would use it, and how?
BTW, this thread is operational in nature, and doesn't really belong on this list. If you wanted to discuss Python code that you've written to help implement such features, then this would be the appropriate place to continue this discussion.
Otherwise, this thread should be re-started on mailman-users.-- Brad Knowles, <brad@stop.mail-abuse.org>
"Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety."
-- Benjamin Franklin (1706-1790), reply of the Pennsylvania
Assembly to the Governor, November 11, 1755LOPSA member since December 2005. See <http://www.lopsa.org/>.
On Tue, 13 Dec 2005 19:38:46 -0600 "Brad Knowles" <brad@stop.mail-abuse.org> wrote:
Where would this button exist? Who would use it, and how?
That is the clearest part of the suggestion. This would be yet another option in moderation. Now "discard" would have a sibling "discard and report as spam." Whether that option shows up and what "report" means would have to be configured by either a the site admin or maybe a list admin. There are a number of reasonable options corresponding with the various spam tools out there, but they all pretty much boil down to "forward it to a pre-specified email address." Not only is this suggestion appropriate for this list, but I also think I recall a previous discussion where the idea was hashed out. I would be willing to bet there's a patch to implement it, because I think the result of the previous discussion was: "Good idea, but too big a new feature for a patch-level release. Provide a patch and people that want it can use it."
-Dale Newfield Dale@Newfield.org
At 11:11 PM -0500 2005-12-13, Dale Newfield wrote:
Not only is this suggestionappropriate for this list, but I also think I recall a previous discussion where the idea was hashed out.
Unless you're talking about Python code you've developed to implement this feature, or commenting on Python code that someone else has developed to implement this feature, I'm pretty certain that this discussion belongs on mailman-users and not here.
When we get down to the point where we're talking about going to the archives of the list to see the previous discussion on this topic, and to see the patch that was produced, etc... then we are definitely into mailman-users territory and not mailman-developers.
If Barry or JC disagree, then I'm willing to bow to their greater knowledge on this topic, since they've been moderators of this list longer than I have. If Tokio or Mark disagree, I'm willing to bow to their greater knowledge because I know they've been hacking on the code longer than I've been associated with the project.
Anyone else is going to have to work pretty hard to convince me.-- Brad Knowles <brad@python.org> Python.org Postmaster Team
Brad Knowles wrote:
Unless you're talking about Python code you've developed to implement this feature, or commenting on Python code that someone else has developed to implement this feature, I'm pretty certain that this discussion belongs on mailman-users and not here.
When we get down to the point where we're talking about going to the archives of the list to see the previous discussion on this topic, and to see the patch that was produced, etc... then we are definitely into mailman-users territory and not mailman-developers.
If Barry or JC disagree, then I'm willing to bow to their greater knowledge on this topic, since they've been moderators of this list longer than I have. If Tokio or Mark disagree, I'm willing to bow to their greater knowledge because I know they've been hacking on the code longer than I've been associated with the project.
IMHO, -users is for discussions about using mailman (installing it, using the features it has, integrating it with other software, etc.). The posts we are hoping to keep OFF of -dev and ON -users are the posts by users who are having what they believe are "complicated" use problems and they want "advanced support" and thus try to post a -users appropriate question to -dev to get an answer faster or from someone with more "knowledge" than those who answer posts on -users. (This is almost always an inappropriate use of -dev and an intrusion on developer time.)
As long as the topic brought here is about *how* to develop a new feature, I believe it's on-topic for -dev. It can often be a good idea to get feedback on the idea and the proposed implementation before spending time on writing the code. This way one can avoid writing one's way into a rat-hole because of a lack of knowledge about why it is the way it is now, and learn the best way to incorporate the proposed fix into the existing design. Then go code it!
I believe the discussion about a "this is spam" button is appropriate for -dev. I agree that there are a lot of technical issues that need to be addressed about if or how to implement it.[1] I believe that -dev IS the right list for that discussion.
When all is said and done though, I'm not a developer. My role here is more of "product manager" - I help with input on priorities, user interface, and of course with the boring details of administering -users and -dev so that the developers can spend their time on actual code development. So Barry (and Tokio and Mark, et al) really have the final say on what they want -dev to include or exclude. Until one of the key developers opines that "this is off-topic" I encourage using -dev for further discussion on the "this is spam" button topic.
jc
[1] Does it belong in the headers, the footer, the body? Should it be a configuration option to place it in one or more of these locations? What happens when the button (or link) is clicked? What if non-spam is reported? What about RFCs? Etc.
"JC" == JC Dill <lists05@equinephotoart.com> writes:
JC> The posts we are hoping to keep OFF of -dev and ON -users are
JC> the posts by users who are having what they believe are
JC> "complicated" use problems and they want "advanced support"[...]
JC> As long as the topic brought here is about *how* to develop a
JC> new feature, I believe it's on-topic for -dev.+1. And good job of presenting, at that. FAQ material?
JC> [1] Does it belong in the headers, the footer, the body?
JC> Should it be a configuration option to place it in one or more
JC> of these locations? What happens when the button (or link) is
JC> clicked? What if non-spam is reported? What about RFCs?
JC> Etc.Niggle: these questions are appropriate for -developers, but should also be polled on -users.
-- School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.
Hi,
Dale Newfield wrote:
On Tue, 13 Dec 2005 19:38:46 -0600 "Brad Knowles" <brad@stop.mail-abuse.org> wrote:
Where would this button exist? Who would use it, and how?
"forward it to a pre-specified email address." Not only is this suggestion appropriate for this list, but I also think I recall a previous discussion where the idea was hashed out. I would be willing to bet there's a patch to implement it, because I think the result of the previous discussion was: "Good idea, but too big a new feature for a patch-level release. Provide a patch and people that want it can use it."
I think you are talking about: http://sourceforge.net/tracker/index.php?func=detail&aid=1008865&group_id=103&atid=300103
Hopefully, we will be back to the MAIN CVS tree for the development of mailman-2.2 after the release of 2.1.7. This should be the chance that the patch is integrated in the 2.2 release.
Regards,
-- Tokio Kikuchi, tkikuchi@ is.kochi-u.ac.jp http://weather.is.kochi-u.ac.jp/
On Wed, 2005-12-14 at 16:38 +0900, Tokio Kikuchi wrote:
I think you are talking about: http://sourceforge.net/tracker/index.php?func=detail&aid=1008865&group_id=103&atid=300103
Hopefully, we will be back to the MAIN CVS tree for the development of mailman-2.2 after the release of 2.1.7. This should be the chance that the patch is integrated in the 2.2 release.
There's also
http://sourceforge.net/tracker/index.php?func=detail&aid=713522&group_id=103&atid=300103
which is my Spambayes integration patch. I do think mailman-developers is a reasonable place to discuss this. We can talk about whether it's even reasonable to have anti-spam defenses in Mailman, and if so whether we want to pick one such product to support, or have a pluggable architecture (possibly shipping one by default).
I also think there are interesting usability/user interface questions to ask, as well as whether spam filters should be per-list, site-wide, or domain-wide.
-Barry
On 12/14/05 3:32 PM, "Barry Warsaw" <barry@python.org> wrote:
I do think mailman-developers is a reasonable place to discuss this. We can talk about whether it's even reasonable to have anti-spam defenses in Mailman, and if so whether we want to pick one such product to support, or have a pluggable architecture (possibly shipping one by default).
I also think there are interesting usability/user interface questions to ask, as well as whether spam filters should be per-list, site-wide, or domain-wide.
We're happy with spam and virus defenses in the MTA which receives from the world and does the spam and virus work before handing off the messages to the MTA on the Mailman machine. The Mailman machine can't be reached on port 25 (or 587 or 465) from the world.
If anything is shipped with Mailman (beyond tuneups to what is there now, which we mostly don't have to use), we'd likely want it able to be turned off.
--John
--On 13 December 2005 19:38:46 -0600 Brad Knowles <brad@stop.mail-abuse.org> wrote:
You could add a new "List-" header which would allow compliant MUAs to give users a single button to click on, but there are currently no MUAs that would support such a feature.
There is an extension to Thunderbird which supports List- header URLs. Turns them into clickable links. I'm not sure whether it would work for a new List- header - it might look for only the RFC 2369 defined List- headers.
Of course, there's also the issue that such a new List- header has no justification in the RFCs. That is to say, we'd have no reason to believe that anyone would add support for it to any MUA.
-- Ian Eiloart Servers Team Sussex University ITS
On Wed, 2005-12-14 at 10:21 +0000, Ian Eiloart wrote:
There is an extension to Thunderbird which supports List- header URLs. Turns them into clickable links. I'm not sure whether it would work for a new List- header - it might look for only the RFC 2369 defined List- headers.
Of course, there's also the issue that such a new List- header has no justification in the RFCs. That is to say, we'd have no reason to believe that anyone would add support for it to any MUA.
And technically, we'd have to add them as X-List-* headers, so it's very doubtful MUAs would support them.
-Barry
On Tue, 2005-12-13 at 14:33 +0200, Rui Correia wrote:
I've just been cleaning up a list that I administer and suddenly it came to me that we could have a report Scam/Spam button for such items being sent to Mailman lists.
A long while ago, I did an experimental integration of Spambayes with Mailman that had a similar concept. Essentially it let you train ham and spam for its bayesian filter. I'm sure the patch is still on SF, but off-hand I don't know whether it still applies to CVS.
While in general I think the mailing list is a lousy place to do spam removal (it is better done upstream of the MLM), I do think we could add some useful controls to help here.
-Barry
At 6:26 PM -0500 2005-12-14, Barry Warsaw wrote:
While in general I think the mailing list is a lousy place to do spam removal (it is better done upstream of the MLM), I do think we could add some useful controls to help here.
I can certainly see the advantage in allowing Mailman to take advantage of additional information placed within the headers of the message by the anti-spam processing system (improved in 2.1.6 over previous versions). And I can see the usefulness of having a "discard and report as spam" option for list owners and moderators, as mentioned by Dale -- basically, just forward the message to a pre-configured e-mail address.
However, I would seriously question the usefulness of trying to integrate a full-blown anti-spam system into Mailman (e.g., SpamBayes). IMO, that kind of thing needs to be integrated into the MTA, not Mailman.
My bigger concern here is that people understand who sees this "discard as spam" button, and how it works. This is not something that would be exposed or otherwise available to the regular mailing list recipients, and I believe that it will be important to manage user expectations in this respect.
-- Brad Knowles, <brad@stop.mail-abuse.org>
"Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety."
-- Benjamin Franklin (1706-1790), reply of the Pennsylvania
Assembly to the Governor, November 11, 1755LOPSA member since December 2005. See <http://www.lopsa.org/>.
I personally agree that MM is a poor place to do spam management. What we're considering doing is using some Courier-IMAP-fu to make it work all pretty like. If SA/dspam/whatever tags a post to a list as spam, it is delivered to an IMAP mailbox folder that the list administrators can access. To approve messages mismarked as spam, they simply move the message to another IMAP folder that is just a pipe back into Mailman.
-jag
On Thu, 2005-12-15 at 03:06 -0600, Brad Knowles wrote:
At 6:26 PM -0500 2005-12-14, Barry Warsaw wrote:
While in general I think the mailing list is a lousy place to do spam removal (it is better done upstream of the MLM), I do think we could add some useful controls to help here.
I can certainly see the advantage in allowing Mailman to take advantage of additional information placed within the headers of the message by the anti-spam processing system (improved in 2.1.6 over previous versions). And I can see the usefulness of having a "discard and report as spam" option for list owners and moderators, as mentioned by Dale -- basically, just forward the message to a pre-configured e-mail address.
However, I would seriously question the usefulness of trying to integrate a full-blown anti-spam system into Mailman (e.g., SpamBayes). IMO, that kind of thing needs to be integrated into the MTA, not Mailman.
My bigger concern here is that people understand who sees this "discard as spam" button, and how it works. This is not something that would be exposed or otherwise available to the regular mailing list recipients, and I believe that it will be important to manage user expectations in this respect.
-- Joshua Ginsberg <jag@fsf.org> Free Software Foundation - Senior Systems Administrator
While agreeing that MM is not really a good spam control place, my life would be made easier if:- * Marking a message as spam killed all other messages sent by the same sender address to that and all other lists currently awaiting moderation. * Marking a message as spam killed all other messages sent with the same subject to that and all other lists currently awaiting moderation.
Those 2 together would reduce the messages I see in moderation substantially - in general I get multiple lists hit with copies of the same stuff within a few seconds....
Additionally automatically discarding messages where identical subject combos hit *different* lists on the box and both going to moderation within a short timeframe would also be a great helper.
Although frankly I'm tempted to just say sod it and make everything that would be moderated get trashed.
Nigel.-- [ Nigel Metheringham Nigel.Metheringham@InTechnology.co.uk ] [ - Comments in this message are my own and not ITO opinion/policy - ]
Nigel Metheringham wrote:
While agreeing that MM is not really a good spam control place, my life would be made easier if:- * Marking a message as spam killed all other messages sent by the same sender address to that and all other lists currently awaiting moderation. * Marking a message as spam killed all other messages sent with the same subject to that and all other lists currently awaiting moderation.
If this marking is taking place within the mailman admin webpage structure, this could be accomplished by providing checkboxes for each of these actions, and the default setting would be to have those boxes pre-checked so that when you click the button, those actions happen. The setting should be changeable within the mailman configuration so that one or both of these boxes are not pre-checked if desired by that mailman admin.
jc
participants (13)
-
Barry Warsaw -
Brad Knowles -
Brad Knowles -
Dale Newfield -
Ian Eiloart -
JC Dill -
John W. Baxter -
Joshua Ginsberg -
Msquared -
Nigel Metheringham -
Rui Correia -
Stephen J. Turnbull -
Tokio Kikuchi