From rhubbell at ihubbell.com Fri May 2 15:38:40 2008 From: rhubbell at ihubbell.com (R.H.) Date: Fri, 02 May 2008 12:38:40 -0700 Subject: [Mulberry-discuss] plugin for bayesian spam control Message-ID: <7396C254BC6EEFB5B636128E@[192.168.1.121]> Hello, Just began using Mulberry and having issues figuring out what to do with spam. What are others doing now to combat spam? I was looking for a bayesian plugin or something like that to manage spam in Mulberry. It worked pretty good in other mail apps. From samuel at translate.org.za Sun May 4 18:40:54 2008 From: samuel at translate.org.za (Samuel Murray (Groenkloof)) Date: Mon, 05 May 2008 00:40:54 +0200 Subject: [Mulberry-discuss] Off-topic: Translators wanted for opensource Message-ID: <481E3B76.9040302@translate.org.za> G'day everyone Please allow me to tell you about our translation project. I'm hoping that some of you might be able to help us with translations or ideas. One of the types of programs that I'm sure will be useful to translate, is e-mail programs. Since you guys already work on or translate an e-mail program, would you mind telling me what you think one should look for in an e-mail program from the point of view of translating it? What things do you think one needs to keep in mind when selecting a single e-mail program for a translation project with volunteers? We call our project the Decathlon because we want to encourage people who feel passionate about their language to translate up to ten or more opensource programs into their languages in 2008. This year, we limit our selection of translated programs to applications aimed at end-users, and preferably programs that run on multiple platforms. All translations are done in our web-based translations system, Pootle. The value of Pootle is that a team of translators can work together on a single file. Pootle also has quality checking features, to ensure that translations don't break the software they are used in. Pootle requires the Gettext PO format, but we can convert certain other formats to and from PO. You can read more about the Decathlon project at http://translate.sourceforge.net/wiki/decathlon/mainpage. Or take a look at the online translation system: http://pootle.locamotion.org/ Or join the low-volume newsletter mailing list: https://lists.sourceforge.net/lists/listinfo/translate-decathlon Or contact the project leader, Samuel Murray, at samuel at translate.org.za. I look forward to hearing from you. Sincerely Samuel Murray Decathlon project leader From mulberry at lists.blurk.net Tue May 6 21:10:33 2008 From: mulberry at lists.blurk.net (Bernie Maier) Date: Wed, 07 May 2008 11:10:33 +1000 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <7396C254BC6EEFB5B636128E@[192.168.1.121]> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> Message-ID: <9623ED89AC7108E2D69FCA91@knockando> R.H.: > Just began using Mulberry and having issues figuring out what to do with > spam. What are others doing now to combat spam? I was looking for a > bayesian plugin or something like that to manage spam in Mulberry. It > worked pretty good in other mail apps. Either I'm missing some list messages or this didn't draw much of a response. For my part, I've never really used serious client-side spam filtering. These days, I rely on server-side spam filtering provided by the various email/domain hosting services I use. Maybe many others do this too, which is why there hasn't been much response? Cheers, Bernie From rhubbell at ihubbell.com Tue May 6 23:59:10 2008 From: rhubbell at ihubbell.com (rh) Date: Tue, 06 May 2008 20:59:10 -0700 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <9623ED89AC7108E2D69FCA91@knockando> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> Message-ID: <4821290E.804@ihubbell.com> I like client-side for the case of a false-positive. Some people send email rarely and I don't like missing their emails. From sombrero at mm.st Wed May 7 00:03:04 2008 From: sombrero at mm.st (Gleason Pace) Date: Wed, 07 May 2008 00:03:04 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <4821290E.804@ihubbell.com> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <4821290E.804@ihubbell.com> Message-ID: <482129F8.6030400@mm.st> rh, > I like client-side for the case of a false-positive. Some people send > email rarely and I don't like missing their > emails. Fastmail has a pretty good configurable/trainable server side spam filter. -- Gleason From david.lang at digitalinsight.com Thu May 8 16:43:42 2008 From: david.lang at digitalinsight.com (David Lang) Date: Thu, 8 May 2008 13:43:42 -0700 (PDT) Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <4821290E.804@ihubbell.com> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]><9623ED89AC7108E2D69FC A91@knockando> <4821290E.804@ihubbell.com> Message-ID: On Tue, 6 May 2008, rh wrote: > I like client-side for the case of a false-positive. Some people send email > rarely and I don't like missing their > emails. how does client-side vs server-side affect this? David Lang P.S. I use popfile which runs on a client machine and makes IMAP connections to the server to move mail between folders. From rhubbell at ihubbell.com Thu May 8 18:05:21 2008 From: rhubbell at ihubbell.com (rh) Date: Thu, 08 May 2008 15:05:21 -0700 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: References: <7396C254BC6EEFB5B636128E@[192.168.1.121]><9623ED89AC7108E2D69FC A91@knockando> <4821290E.804@ihubbell.com> Message-ID: <48237921.8000101@ihubbell.com> David Lang wrote: > On Tue, 6 May 2008, rh wrote: > >> I like client-side for the case of a false-positive. Some people >> send email rarely and I don't like missing their >> emails. > > how does client-side vs server-side affect this? > > David Lang > > P.S. I use popfile which runs on a client machine and makes IMAP > connections to the server to move mail between folders. > Sorry, don't understand the question. From david.lang at digitalinsight.com Thu May 8 17:25:52 2008 From: david.lang at digitalinsight.com (David Lang) Date: Thu, 8 May 2008 14:25:52 -0700 (PDT) Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <48237921.8000101@ihubbell.com> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]><9623ED89AC7108E2D69FC A91@knockando> <4821290E.804@ihubbell.com> <48237921.8000101@ihubbell.com> Message-ID: On Thu, 8 May 2008, rh wrote: > David Lang wrote: >> On Tue, 6 May 2008, rh wrote: >> >>> I like client-side for the case of a false-positive. Some people send >>> email rarely and I don't like missing their >>> emails. >> >> how does client-side vs server-side affect this? >> >> David Lang >> >> P.S. I use popfile which runs on a client machine and makes IMAP >> connections to the server to move mail between folders. >> > Sorry, don't understand the question. you said that you like client-side for the case of false-positives. I'm asking what difference client-side or server-side makes in this case. David Lang From rhubbell at ihubbell.com Thu May 8 18:12:16 2008 From: rhubbell at ihubbell.com (rh) Date: Thu, 08 May 2008 15:12:16 -0700 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: References: <7396C254BC6EEFB5B636128E@[192.168.1.121]><9623ED89AC7108E2D69FC A91@knockando> <4821290E.804@ihubbell.com> <48237921.8000101@ihubbell.com> Message-ID: <48237AC0.40701@ihubbell.com> David Lang wrote: > On Thu, 8 May 2008, rh wrote: > >> David Lang wrote: >>> On Tue, 6 May 2008, rh wrote: >>> >>>> I like client-side for the case of a false-positive. Some people >>>> send email rarely and I don't like missing their >>>> emails. >>> >>> how does client-side vs server-side affect this? >>> >>> David Lang >>> >>> P.S. I use popfile which runs on a client machine and makes IMAP >>> connections to the server to move mail between folders. >>> >> Sorry, don't understand the question. > > you said that you like client-side for the case of false-positives. > > I'm asking what difference client-side or server-side makes in this case. > > David Lang That's what you said already. So I'll ask you. What difference doesn't it make? From Hagedorn at uni-koeln.de Thu May 8 18:54:45 2008 From: Hagedorn at uni-koeln.de (Sebastian Hagedorn) Date: Fri, 09 May 2008 00:54:45 +0200 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <48237AC0.40701@ihubbell.com> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FC A91@knockando> <4821290E.804@ihubbell.com> <48237921.8000101@ihubbell.com> <48237AC0.40701@ihubbell.com> Message-ID: <3B629F68EB99D7DE1203947B@G5> -- rh is rumored to have mumbled on 8. Mai 2008 15:12:16 -0700 regarding Re: [Mulberry-discuss] plugin for bayesian spam control: >> I'm asking what difference client-side or server-side makes in this case. >> >> David Lang > That's what you said already. So I'll ask you. What difference doesn't > it make? It doesn't make any difference! Both on the client and on the server you can ever discard messages (in which case false positives are a bad problem) or you file them in a separate mailbox (in which case false positives aren't that bad, as longs as their number isn't too high). -- Sebastian Hagedorn - RZKR-R1 (Flachbau), Zi. 18, Robert-Koch-Str. 10 Zentrum f?r angewandte Informatik - Universit?tsweiter Service RRZK Universit?t zu K?ln / Cologne University - Tel. +49-221-478-5587 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 194 bytes Desc: not available Url : http://lists.mulberrymail.com/pipermail/mulberry-discuss/attachments/20080509/c982fe07/attachment.bin From Hagedorn at uni-koeln.de Thu May 8 18:56:48 2008 From: Hagedorn at uni-koeln.de (Sebastian Hagedorn) Date: Fri, 09 May 2008 00:56:48 +0200 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <3B629F68EB99D7DE1203947B@G5> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FC A91@knockando> <4821290E.804@ihubbell.com> <48237921.8000101@ihubbell.com> <48237AC0.40701@ihubbell.com> <3B629F68EB99D7DE1203947B@G5> Message-ID: -- Sebastian Hagedorn is rumored to have mumbled on 9. Mai 2008 00:54:45 +0200 regarding Re: [Mulberry-discuss] plugin for bayesian spam control: > It doesn't make any difference! Both on the client and on the server you > can ever That was supposed to read "either" instead of "ever" ... > discard messages (in which case false positives are a bad > problem) or you file them in a separate mailbox (in which case false > positives aren't that bad, as longs as their number isn't too high). -- Sebastian Hagedorn - RZKR-R1 (Flachbau), Zi. 18, Robert-Koch-Str. 10 Zentrum f?r angewandte Informatik - Universit?tsweiter Service RRZK Universit?t zu K?ln / Cologne University - Tel. +49-221-478-5587 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 194 bytes Desc: not available Url : http://lists.mulberrymail.com/pipermail/mulberry-discuss/attachments/20080509/1d0142dd/attachment.bin From rhubbell at ihubbell.com Thu May 8 19:16:42 2008 From: rhubbell at ihubbell.com (rh) Date: Thu, 08 May 2008 16:16:42 -0700 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <3B629F68EB99D7DE1203947B@G5> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FC A91@knockando> <4821290E.804@ihubbell.com> <48237921.8000101@ihubbell.com> <48237AC0.40701@ihubbell.com> <3B629F68EB99D7DE1203947B@G5> Message-ID: <482389DA.1080508@ihubbell.com> Sebastian Hagedorn wrote: > -- rh is rumored to have mumbled on 8. Mai > 2008 15:12:16 -0700 regarding Re: [Mulberry-discuss] plugin for > bayesian spam control: > >>> I'm asking what difference client-side or server-side makes in this >>> case. >>> >>> David Lang >> That's what you said already. So I'll ask you. What difference doesn't >> it make? > > It doesn't make any difference! Both on the client and on the server > you can ever discard messages (in which case false positives are a bad > problem) or you file them in a separate mailbox (in which case false > positives aren't that bad, as longs as their number isn't too high). Ok, now I got it. I like to get stuff off the server and manage it in the mailer. I've just always done it that way. I don't abhor change but I resist when it seems arbitrary or maybe not important. This seems like a six-of-one-and-half-dozen-of-the-other kind of situation. Maybe in Germany there's a metric version of that like "10mm-of-one-and-1cm-of-the-other". I'm pretty sure dozen isn't metric. And FWIW the metric system is a good example of a change that I'm down with. I'll check into filtering on the server and see if I can live with it. Also I'm glad to see that someone's on this list. Thanks everyone for the replies. > -- > Sebastian Hagedorn - RZKR-R1 (Flachbau), Zi. 18, Robert-Koch-Str. 10 > Zentrum f?r angewandte Informatik - Universit?tsweiter Service RRZK > Universit?t zu K?ln / Cologne University - Tel. +49-221-478-5587 From srb at umich.edu Thu May 8 20:23:48 2008 From: srb at umich.edu (Steve Burling) Date: Thu, 08 May 2008 20:23:48 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <482389DA.1080508@ihubbell.com> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FC A91@knockando> <4821290E.804@ihubbell.com> <48237921.8000101@ihubbell.com> <48237AC0.40701@ihubbell.com> <3B629F68EB99D7DE1203947B@G5> <482389DA.1080508@ihubbell.com> Message-ID: <68E5EE77D3D214A105139831@srb.icpsr.umich.edu> --On May 8, 2008 4:16:42 PM -0700 rh wrote: > I like to get stuff off the server and manage it in the mailer. I've > just always done it that way. To which I reply: You might investigate using server-side mail storage; there are some major benefits unless you always, and only, read mail from a single machine. All of my mail has lived on an IMAP server for more than 10 years now, and I can get at it from wherever I happen to find a computer (or an iPhone or ...). If it was all sitting on my laptop somewhere, well, that'd suck if I wasn't near that laptop and happened to want to look up something. If course, it helps that I have access to an IMAP server with lots of disk, and am not constrained by the somewhat meager mailbox quotas imposed by our local Exchange dudes. Ick, Exchange. -- Steve Burling University of Michigan, ICPSR Voice: +1 734 615.3779 330 Packard Street FAX: +1 734 647.8700 Ann Arbor, MI 48104-2910 From iane at sussex.ac.uk Fri May 9 05:38:28 2008 From: iane at sussex.ac.uk (Ian Eiloart) Date: Fri, 09 May 2008 10:38:28 +0100 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <9623ED89AC7108E2D69FCA91@knockando> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> Message-ID: --On 7 May 2008 11:10:33 +1000 Bernie Maier wrote: > R.H.: > >> Just began using Mulberry and having issues figuring out what to do with >> spam. What are others doing now to combat spam? I was looking for a >> bayesian plugin or something like that to manage spam in Mulberry. It >> worked pretty good in other mail apps. > > Either I'm missing some list messages or this didn't draw much of a > response. For my part, I've never really used serious client-side spam > filtering. These days, I rely on server-side spam filtering provided by > the various email/domain hosting services I use. Maybe many others do > this too, which is why there hasn't been much response? > Perhaps. Especially as Mulberry is more popular in institutions that have local control of server side filtering. Server side filtering is MUCH MUCH more beneficial than client side, especially if you reject unwanted email - then there's no chance of false positives going unseen in spam mailboxes. -- Ian Eiloart IT Services, University of Sussex x3148 From iane at sussex.ac.uk Fri May 9 05:40:57 2008 From: iane at sussex.ac.uk (Ian Eiloart) Date: Fri, 09 May 2008 10:40:57 +0100 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <48237AC0.40701@ihubbell.com> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FC A91@knockando> <4821290E.804@ihubbell.com> <48237921.8000101@ihubbell.com> <48237AC0.40701@ihubbell.com> Message-ID: --On 8 May 2008 15:12:16 -0700 rh wrote: > David Lang wrote: >> On Thu, 8 May 2008, rh wrote: >> >>> David Lang wrote: >>>> On Tue, 6 May 2008, rh wrote: >>>> >>>>> I like client-side for the case of a false-positive. Some people >>>>> send email rarely and I don't like missing their >>>>> emails. >>>> >>>> how does client-side vs server-side affect this? >>>> >>>> David Lang >>>> >>>> P.S. I use popfile which runs on a client machine and makes IMAP >>>> connections to the server to move mail between folders. >>>> >>> Sorry, don't understand the question. >> >> you said that you like client-side for the case of false-positives. >> >> I'm asking what difference client-side or server-side makes in this case. >> >> David Lang > That's what you said already. So I'll ask you. What difference doesn't > it make? This is silly. I'll answer the question: If you filter server side, you have more options. You can do everything that the mail client can do. Also, you can reject the message (rather than generating a bounce message), and that should make the sender aware. You can't reject client side. -- Ian Eiloart IT Services, University of Sussex x3148 From iane at sussex.ac.uk Fri May 9 05:42:37 2008 From: iane at sussex.ac.uk (Ian Eiloart) Date: Fri, 09 May 2008 10:42:37 +0100 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <482389DA.1080508@ihubbell.com> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FC A91@knockando> <4821290E.804@ihubbell.com> <48237921.8000101@ihubbell.com> <48237AC0.40701@ihubbell.com> <3B629F68EB99D7DE1203947B@G5> <482389DA.1080508@ihubbell.com> Message-ID: <0A83B8820FB23570CA679752@lewes.staff.uscs.susx.ac.uk> --On 8 May 2008 16:16:42 -0700 rh wrote: > This seems like a six-of-one-and-half-dozen-of-the-other > kind of situation. Maybe in Germany there's a metric version of that > like "10mm-of-one-and-1cm-of-the-other". In the North of England, they use the rather more efficient "six and two threes"! -- Ian Eiloart IT Services, University of Sussex x3148 From sombrero at mm.st Fri May 9 10:05:48 2008 From: sombrero at mm.st (Gleason Pace) Date: Fri, 09 May 2008 10:05:48 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> Message-ID: <48245A3C.9070406@mm.st> Ian Eiloart, > > > --On 7 May 2008 11:10:33 +1000 Bernie Maier > wrote: > >> R.H.: >> >>> Just began using Mulberry and having issues figuring out what to do >>> with >>> spam. What are others doing now to combat spam? I was looking for a >>> bayesian plugin or something like that to manage spam in Mulberry. It >>> worked pretty good in other mail apps. >> >> Either I'm missing some list messages or this didn't draw much of a >> response. For my part, I've never really used serious client-side spam >> filtering. These days, I rely on server-side spam filtering provided by >> the various email/domain hosting services I use. Maybe many others do >> this too, which is why there hasn't been much response? >> > > Perhaps. Especially as Mulberry is more popular in institutions that > have local control of server side filtering. > > Server side filtering is MUCH MUCH more beneficial than client side, > especially if you reject unwanted email - then there's no chance of > false positives going unseen in spam mailboxes. What am I missing here? If unwanted email is rejected, don't false positives dissappear without being seen? Specifically, Mulberry is popular with administrators of institutional IT. When other users make requests for things like client side spam filters or full display of html in email, the refusal of IT admistrators to see the virtue of such can be deafening. -- Gleason From rhubbell at ihubbell.com Fri May 9 15:50:19 2008 From: rhubbell at ihubbell.com (rh) Date: Fri, 09 May 2008 12:50:19 -0700 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <0A83B8820FB23570CA679752@lewes.staff.uscs.susx.ac.uk> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FC A91@knockando> <4821290E.804@ihubbell.com> <48237921.8000101@ihubbell.com> <48237AC0.40701@ihubbell.com> <3B629F68EB99D7DE1203947B@G5> <482389DA.1080508@ihubbell.com> <0A83B8820FB23570CA679752@lewes.staff.uscs.susx.ac.uk> Message-ID: <4824AAFB.5090502@ihubbell.com> Ian Eiloart wrote: > > > --On 8 May 2008 16:16:42 -0700 rh wrote: > >> This seems like a six-of-one-and-half-dozen-of-the-other >> kind of situation. Maybe in Germany there's a metric version of that >> like "10mm-of-one-and-1cm-of-the-other". > > In the North of England, they use the rather more efficient "six and > two threes"! > Ha, good one! Very efficient.. From lispamateur at yahoo.de Fri May 9 12:42:06 2008 From: lispamateur at yahoo.de (Oliver Cromm) Date: Fri, 9 May 2008 12:42:06 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> Message-ID: <1upl9g56eo1kz.dlg@mid.crommatograph.info> * "R.H." wrote: > Just began using Mulberry and having issues figuring out what to do with > spam. What are others doing now to combat spam? I was looking for a > bayesian plugin or something like that to manage spam in Mulberry. It > worked pretty good in other mail apps. In the past, I have been using Spambayes as a Proxy, and it worked quite well. As it's an external app, you'll have to see how it fits into your workflow (POP? IMAP? Local server? etc.) -- Performance: A statement of the speed at which a computer system works. Or rather, might work under certain circumstances. Or was rumored to be working over in Jersey about a month ago. From dmz-lists at tffenterprises.com Fri May 9 13:34:54 2008 From: dmz-lists at tffenterprises.com (Daniel M. Zimmerman) Date: Fri, 09 May 2008 10:34:54 -0700 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <48245A3C.9070406@mm.st> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> Message-ID: <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> --On 9 May 2008 10:05:48 -0400 Gleason Pace wrote: > What am I missing here? If unwanted email is rejected, don't false > positives dissappear without being seen? No, they don't "disappear without being seen"; they _bounce_, letting the sender know that something went wrong and to either fix _their_ problem or pursue an alternative contact method. > Specifically, Mulberry is popular with administrators of institutional > IT. When other users make requests for things like client side spam > filters or full display of html in email, the refusal of IT admistrators > to see the virtue of such can be deafening. I'd think that client-side spam filtering would be an absolute nightmare for institutional IT, as they'd have basically no way to figure out what happened to a message (such as a false positive) once it left the server (unless the client keeps detailed logs, which is unlikely). HTML in email is a different story, and we've already had the arguments about that ad nauseum on this list... -Dan ------------------------------------------------------------------ Daniel M. Zimmerman TFF Enterprises 1900 Commerce St. Box 358426 http://www.tffenterprises.com/ Tacoma, Washington 98402 USA dmz at tffenterprises.com From sombrero at mm.st Fri May 9 17:09:04 2008 From: sombrero at mm.st (Gleason Pace) Date: Fri, 09 May 2008 17:09:04 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> Message-ID: <4824BD70.2070208@mm.st> Daniel M. Zimmerman, >> What am I missing here? If unwanted email is rejected, don't false >> positives dissappear without being seen? > > No, they don't "disappear without being seen"; they _bounce_, letting > the sender know that something went wrong and to either fix _their_ > problem or pursue an alternative contact method. Good in an academic situation perhaps, where the person receiving the message has less interest in successful delivery than the sender. If the recipient doesn't want to make the sender think he _has_ a *problem*, such treatment is inappropriate. In other words, most false positives I see are inoccous enough. I would have a hard time figuring out what to change myself. It is no good forcing my correspondents to struggle with the ineptness of my spam filter, not in a business situation, anyway. And I am just a simple programmer wanting to keep in effective touch with users and buyers of my software. >> Specifically, Mulberry is popular with administrators of institutional >> IT. When other users make requests for things like client side spam >> filters or full display of html in email, the refusal of IT admistrators >> to see the virtue of such can be deafening. > > I'd think that client-side spam filtering would be an absolute > nightmare for institutional IT, as they'd have basically no way to > figure out what happened to a message (such as a false positive) once > it left the server (unless the client keeps detailed logs, which is > unlikely). Yes, I am the one who has interest in making sure I don't loose a false positive. I don't expect my email server to be involved or care about that in any way. > HTML in email is a different story, and we've already had the > arguments about that ad nauseum on this list... Oh, yes, indeed. The instructive thing is to notice is who is nauseated by the discussion. -- Gleason From dmz-lists at tffenterprises.com Fri May 9 17:17:11 2008 From: dmz-lists at tffenterprises.com (Daniel M. Zimmerman) Date: Fri, 09 May 2008 14:17:11 -0700 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <8E08E132673B2FCD41A77AD3@dhcp017107106066.corp.apple.com> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <8E08E132673B2FCD41A77AD3@dhcp017107106066.corp.apple.com> Message-ID: <4B42D74F208B1437D582871E@D-128-208-244-210.dhcp4.washington.edu> --On 9 May 2008 13:58:14 -0700 Capitan Holy Hippie wrote: > > --On May 9, 2008 10:34:54 AM -0700 "Daniel M. Zimmerman" > wrote: > >> --On 9 May 2008 10:05:48 -0400 Gleason Pace wrote: >> >>> What am I missing here? If unwanted email is rejected, don't false >>> positives dissappear without being seen? >> >> No, they don't "disappear without being seen"; they _bounce_, letting >> the sender know that something went wrong and to either fix _their_ >> problem or pursue an alternative contact method. > > This is not a wise policy. In fact, it can get other sites quite annoyed > at you. Nevertheless, it is how the RFC is written. If an SMTP server _rejects_ a message (as opposed to accepting it and then doing something else with it), it _must_ generate a bounce. If, on the other hand, the SMTP server _accepts_ the message and then filters it in some way, that's another story. Since the original question used the word "rejected", I assumed the word was being used properly. Accepting a message and filing it in a spam mailbox (which is how most server-side filtering that I've seen and/or written - http://www.tffenterprises.com/cgpsa/ - works) is not the same as rejecting the message. -Dan ------------------------------------------------------------------ Daniel M. Zimmerman TFF Enterprises 1900 Commerce St. Box 358426 http://www.tffenterprises.com/~dmz/ Tacoma, WA 98402 USA dmz at tffenterprises.com From dmz-lists at tffenterprises.com Fri May 9 17:29:01 2008 From: dmz-lists at tffenterprises.com (Daniel M. Zimmerman) Date: Fri, 09 May 2008 14:29:01 -0700 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <4824BD70.2070208@mm.st> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> Message-ID: <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> --On 9 May 2008 17:09:04 -0400 Gleason Pace wrote: > > > Daniel M. Zimmerman, >>> What am I missing here? If unwanted email is rejected, don't false >>> positives dissappear without being seen? >> >> No, they don't "disappear without being seen"; they _bounce_, letting >> the sender know that something went wrong and to either fix _their_ >> problem or pursue an alternative contact method. > > Good in an academic situation perhaps, where the person receiving the > message has less interest in successful delivery than the sender. If the > recipient doesn't want to make the sender think he _has_ a *problem*, > such treatment is inappropriate. In other words, most false positives I > see are inoccous enough. I would have a hard time figuring out what to > change myself. It is no good forcing my correspondents to struggle with > the ineptness of my spam filter, not in a business situation, anyway. > And I am just a simple programmer wanting to keep in effective touch with > users and buyers of my software. If the sender actually does have a problem (like using a blacklisted SMTP server), then covering it up is counterproductive. But there is (as I just said in a previous message) a substantial difference between messages that are _rejected_ (the word you originally used) and messages that are _accepted and filtered_. The latter can be, and generally are, filed in a spam mailbox for review by the user with no message loss - or even just tagged with a spam score so that client-side filters can sort it into appropriate mailboxes/priorities/etc. >>> Specifically, Mulberry is popular with administrators of institutional >>> IT. When other users make requests for things like client side spam >>> filters or full display of html in email, the refusal of IT admistrators >>> to see the virtue of such can be deafening. >> >> I'd think that client-side spam filtering would be an absolute >> nightmare for institutional IT, as they'd have basically no way to >> figure out what happened to a message (such as a false positive) once >> it left the server (unless the client keeps detailed logs, which is >> unlikely). > > Yes, I am the one who has interest in making sure I don't loose a false > positive. I don't expect my email server to be involved or care about > that in any way. I'd say that most people _serviced by institutional IT_ (the context you were talking about) who did not receive a message they were expecting (or who otherwise suspected that some message loss was occurring) would demand that their mail administrator find out what happened, which they could of course not do if the filtering was client-side and un-logged. That's the only point I was trying to make here. It is _easier for institutional IT_ for these things to be server-side, in terms of maintenance, accountability, disaster recovery, and a host of other considerations. >> HTML in email is a different story, and we've already had the >> arguments about that ad nauseum on this list... > > Oh, yes, indeed. The instructive thing is to notice is who is nauseated > by the discussion. I'd say that's probably just about everybody at this point, since the discussion has been occurring on and off for almost 10 years (the earliest message I can find on the list archives asking for HTML images and tables is dated 10 June 1998). -Dan ------------------------------------------------------------------ Daniel M. Zimmerman TFF Enterprises 1900 Commerce St. Box 358426 http://www.tffenterprises.com/~dmz/ Tacoma, WA 98402 USA dmz at tffenterprises.com From sombrero at mm.st Fri May 9 17:42:53 2008 From: sombrero at mm.st (Gleason Pace) Date: Fri, 09 May 2008 17:42:53 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> Message-ID: <4824C55D.9060903@mm.st> Daniel M. Zimmerman, >>>> Specifically, Mulberry is popular with administrators of institutional >>>> IT. When other users make requests for things like client side spam >>>> filters or full display of html in email, the refusal of IT >>>> admistrators >>>> to see the virtue of such can be deafening. >>> >>> I'd think that client-side spam filtering would be an absolute >>> nightmare for institutional IT, as they'd have basically no way to >>> figure out what happened to a message (such as a false positive) once >>> it left the server (unless the client keeps detailed logs, which is >>> unlikely). >> >> Yes, I am the one who has interest in making sure I don't loose a false >> positive. I don't expect my email server to be involved or care about >> that in any way. > > I'd say that most people _serviced by institutional IT_ (the context > you were talking about) who did not receive a message they were > expecting (or who otherwise suspected that some message loss was > occurring) would demand that their mail administrator find out what > happened, which they could of course not do if the filtering was > client-side and un-logged. That's the only point I was trying to make > here. It is _easier for institutional IT_ for these things to be > server-side, in terms of maintenance, accountability, disaster > recovery, and a host of other considerations. Preferable is that all mail received by the server is delivered. With spam codes is fine. That way, there is no need to ask the administrator what happened to a message. There are other ways of dealing with disaster recovery and maintenance issues than taking spam control out of the hands of users. >>> HTML in email is a different story, and we've already had the >>> arguments about that ad nauseum on this list... >> >> Oh, yes, indeed. The instructive thing is to notice is who is nauseated >> by the discussion. > > I'd say that's probably just about everybody at this point, since the > discussion has been occurring on and off for almost 10 years (the > earliest message I can find on the list archives asking for HTML > images and tables is dated 10 June 1998). No, I think you'll find that those of us on the advocating side will be interested in continuing the discussion until our needs are recognized and fulfilled. I can think of several here. -- Gleason From srb at umich.edu Fri May 9 19:03:58 2008 From: srb at umich.edu (Steve Burling) Date: Fri, 09 May 2008 19:03:58 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <4824C55D.9060903@mm.st> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> Message-ID: --On May 9, 2008 5:42:53 PM -0400 Gleason Pace wrote: > There are other ways of dealing with disaster recovery and maintenance > issues than taking spam control out of the hands of users. Who has said that server-side filtering has to take spam control out of the hands of users? I use server-side filtering, but it is *entirely* under my control what happens with those filters. -- And on the subject of HTML support, you wrote: > No, I think you'll find that those of us on the advocating side will be > interested in continuing the discussion until our needs are recognized > and fulfilled. > I can think of several here. You can continue the discussion (or at least try to continue it) as long as you have the taste for it. But until you find someone who agrees with you, and who either: i) has the skills to do the work, or ii) has the money to pay someone with the skills to do the work your needs are going to go unfulfilled. I'm sorry if this comes across as snide; it's really not meant to be that way. But the simple truth seems to be that the people with the interest and skills to do further Mulberry development have little or no interest in improving Mulberry's HTML support beyond its current rudimentary state. -- Steve Burling University of Michigan, ICPSR Voice: +1 734 615.3779 330 Packard Street FAX: +1 734 647.8700 Ann Arbor, MI 48104-2910 From sombrero at mm.st Fri May 9 23:18:58 2008 From: sombrero at mm.st (Gleason Pace) Date: Fri, 09 May 2008 23:18:58 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> Message-ID: <48251422.3000604@mm.st> Steve Burling, > You can continue the discussion (or at least try to continue it) as > long as you have the taste for it. But until you find someone who > agrees with you, and who either: > > i) has the skills to do the work, or > ii) has the money to pay someone with the skills to do the work > > your needs are going to go unfulfilled. > > I'm sorry if this comes across as snide; it's really not meant to be > that way. But the simple truth seems to be that the people with the > interest and skills to do further Mulberry development have little or > no interest in improving Mulberry's HTML support beyond its current > rudimentary state. I noticed. -- Gleason From davidbo at kth.se Sat May 10 04:31:09 2008 From: davidbo at kth.se (david) Date: Sat, 10 May 2008 10:31:09 +0200 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> Message-ID: <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> --On 2008-05-09 19.03 -0400 Steve Burling wrote: >> There are other ways of dealing with disaster recovery and maintenance >> issues than taking spam control out of the hands of users. > > Who has said that server-side filtering has to take spam control out of > the hands of users? I use server-side filtering, but it is *entirely* > under my control what happens with those filters. This is a very narrow minded view on it. I think you two talk about (very) different situations, e.g., mail handling on a university department and at a big ISP. On a university department, server side is fine because (as I did) you can simply knock on the postmaster's door and ask him to make an exception for one mail account if the spam filter for some reason works badly for you. Try that with e.g., Gmail - which, btw is a good example of the problems with server side filtering: when I get bounces on my Gmail-accounts because I have used an erroneous address (typically misspelled) most of the time they end among the spam. Very bad! And I can't do anything about it. -- Visit the Mulberry IRC-channel irc://irc.freenode.net/mulberry From dmz-lists at tffenterprises.com Sat May 10 04:58:47 2008 From: dmz-lists at tffenterprises.com (Daniel M. Zimmerman) Date: Sat, 10 May 2008 01:58:47 -0700 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> Message-ID: <4610357BC5CB05E3B4E51411@whitestar.local> --On 10 May 2008 10:31:09 +0200 david wrote: > --On 2008-05-09 19.03 -0400 Steve Burling wrote: > >>> There are other ways of dealing with disaster recovery and maintenance >>> issues than taking spam control out of the hands of users. >> >> Who has said that server-side filtering has to take spam control out of >> the hands of users? I use server-side filtering, but it is *entirely* >> under my control what happens with those filters. > > This is a very narrow minded view on it. I think you two talk about > (very) different situations, e.g., mail handling on a university > department and at a big ISP. On a university department, server side is > fine because (as I did) you can simply knock on the postmaster's door and > ask him to make an exception for one mail account if the spam filter for > some reason works badly for you. Try that with e.g., Gmail - which, btw > is a good example of the problems with server side filtering: when I get > bounces on my Gmail-accounts because I have used an erroneous address > (typically misspelled) most of the time they end among the spam. Very > bad! And I can't do anything about it. Actually, it seems as though you are thinking only about specific types of server-side filtering, as the distinction has nothing whatsoever to do with "academic" vs. "large ISP". There are several different types of server-side filtering, some of which are: - realtime blackhole lists, which reject potential spam before it's ever sent to the receiving server in the first place - server-side spam scoring (which may factor in data from blackhole lists) where spam above a certain threshold is discarded - server-side spam scoring where spam above a certain threshold is sequestered in a separate mailbox - server-side spam scoring where spam is tagged and its disposition is left entirely to the client side app The first two of these clearly take control away from the end user, though with the second the actual threshold at which spam is discarded may be tunable by the end user (and filtering might even be able to be turned off entirely); the third is debatable (it depends on how the end user feels about having a dedicated spam mailbox, and whether the end user can customize the spam threshold); and the fourth clearly takes _no_ control away from the end user, leaving it entirely up to her to decide how to treat spam that scores above a certain threshold based on the headers inserted by the filter, or even to ignore the filter's "advice" completely. As for the bounces on your Gmail accounts that end up among the spam, you _can_ do something about them by using the "Not Spam" button, just like you presumably use the "Report Spam" button on spam messages that _don't_ go into the spam mailbox. If you do that enough times, Gmail - in a manner very similar to that of a client-side Bayesian filter - should learn that you like seeing bounces of messages you sent. All of this having been said, I suspect this topic is now straying somewhat outside the purview of this mailing list, and might better be discussed offline or elsewhere. The original question, about a plugin for Bayesian spam control, has (I think) pretty much been answered - there currently isn't one for Mulberry, and as far as I can see nobody has expressed interest in writing one... -Dan ------------------------------------------------------------------ Daniel M. Zimmerman TFF Enterprises 1900 Commerce St. Box 358426 http://www.tffenterprises.com/~dmz/ Tacoma, WA 98402 USA dmz at tffenterprises.com From sombrero at mm.st Sat May 10 08:40:30 2008 From: sombrero at mm.st (Gleason Pace) Date: Sat, 10 May 2008 08:40:30 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <4610357BC5CB05E3B4E51411@whitestar.local> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> <4610357BC5CB05E3B4E51411@whitestar.local> Message-ID: <482597BE.80907@mm.st> Daniel M. Zimmerman, > All of this having been said, I suspect this topic is now straying > somewhat outside the purview of this mailing list, and might better be > discussed offline or elsewhere. The original question, about a plugin > for Bayesian spam control, has (I think) pretty much been answered - > there currently isn't one for Mulberry, and as far as I can see nobody > has expressed interest in writing one... I have been a shoeless wanderer among various online groups (including usenet) for some years. One of the rarer creatures that you meet in the bush is the discussionShutDown Bear. The one who announces when a discussion must be finished. This is distinct from Moderator Rabbits which tend to more generous in the matter. -- Gleason From srb at umich.edu Sat May 10 15:57:16 2008 From: srb at umich.edu (Steve Burling) Date: Sat, 10 May 2008 15:57:16 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> Message-ID: <1F7D34683D20C21D829A721F@srb.local> --On May 10, 2008 10:31:09 AM +0200 david wrote: > > Who has said that server-side filtering has to take spam control out of > > the hands of users? I use server-side filtering, but it is *entirely* > > under my control what happens with those filters. > > This is a very narrow minded view on it. I think you two talk about > (very) different situations, e.g., mail handling on a university > department and at a big ISP. On a university department, server side is > fine because (as I did) you can simply knock on the postmaster's door and > ask him to make an exception for one mail account if the spam filter for > some reason works badly for you. Try that with e.g., Gmail - which, btw > is a good example of the problems with server side filtering: when I get > bounces on my Gmail-accounts because I have used an erroneous address > (typically misspelled) most of the time they end among the spam. Very > bad! And I can't do anything about it. To which I reply: What part of my original statement above was unclear? I'll repeat it: I use server-side filtering, but it is *entirely* under my control what happens with those filters. I don't have to knock on the postmaster's door and ask him (or her, in umich's case) to make an exception. She has *nothing* to do with how the filtering is done. *I* set up the filters. *I* control what they do. If they're broken in some way, *I* fix them. And this is on the university's central IMAP service, which serves tens of thousands of users, each and every one of whom has the same ability to control their destiny that I do. Now, if I was in an unenlightened environment where the SPAM filtering was actually blocking or deleting mail, rather than just tagging it, you might have an argument. But I don't, and (for me, at least) you don't. -- Steve Burling University of Michigan, ICPSR Voice: +1 734 615.3779 330 Packard Street FAX: +1 734 647.8700 Ann Arbor, MI 48104-2910 From davidbo at kth.se Sun May 11 05:44:32 2008 From: davidbo at kth.se (david) Date: Sun, 11 May 2008 11:44:32 +0200 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <4610357BC5CB05E3B4E51411@whitestar.local> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> <4610357BC5CB05E3B4E51411@whitestar.local> Message-ID: --On 2008-05-10 01.58 -0700 "Daniel M. Zimmerman" wrote: > - realtime blackhole lists, which reject potential spam before it's ever > sent to the receiving server in the first place > - server-side spam scoring (which may factor in data from blackhole > lists) where spam above a certain threshold is discarded > - server-side spam scoring where spam above a certain threshold is > sequestered in a separate mailbox > - server-side spam scoring where spam is tagged and its disposition is > left entirely to the client side app I agree with you. My point is that with client side filtering this is a non-issue and if we want Mulberry to become more popular it can not be the only client without client side filtering. Regarding Gmail, I have "non-spam" marked bounces for years, to no avail. Since I even worked at Google I reported this problem in Google's internal issue system which should give it higher priority than external reports. -- Visit the Mulberry IRC-channel irc://irc.freenode.net/mulberry From davidbo at kth.se Sun May 11 05:56:20 2008 From: davidbo at kth.se (david) Date: Sun, 11 May 2008 11:56:20 +0200 Subject: [Mulberry-discuss] Custom addresses Message-ID: I like to tag the e-mail address I use in web forms and similar with some keywords to indicate where I entered it as well as what type of matter the message was about, for example: davidbo+domain.com+bug at ... However, when I get replies on these mails and reply again there seem to be no way to use this custom e-mail address less than creating a new identity. Any ideas? Or must I file a feature request? -- Visit the Mulberry IRC-channel irc://irc.freenode.net/mulberry From davidbo at kth.se Sun May 11 05:59:51 2008 From: davidbo at kth.se (david) Date: Sun, 11 May 2008 11:59:51 +0200 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <1F7D34683D20C21D829A721F@srb.local> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> <1F7D34683D20C21D829A721F@srb.local> Message-ID: <68FC75038D1D92C6FA7A5F7E@c80-216-42-109.bredband.comhem.se> --On 2008-05-10 15.57 -0400 Steve Burling wrote: > I don't have to knock on the postmaster's door and ask him (or her, in > umich's case) to make an exception. She has *nothing* to do with how the > filtering is done. *I* set up the filters. *I* control what they do. > If they're broken in some way, *I* fix them. And this is on the > university's central IMAP service, which serves tens of thousands of > users, each and every one of whom has the same ability to control their > destiny that I do. So you, as a regular user, can install programs (e.g., different spam filters) on your mail server? In that case, you have access to a very impressive mail setup. -- Visit the Mulberry IRC-channel irc://irc.freenode.net/mulberry From srb at umich.edu Sun May 11 06:52:01 2008 From: srb at umich.edu (Steve Burling) Date: Sun, 11 May 2008 06:52:01 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <68FC75038D1D92C6FA7A5F7E@c80-216-42-109.bredband.comhem.se> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> <1F7D34683D20C21D829A721F@srb.local> <68FC75038D1D92C6FA7A5F7E@c80-216-42-109.bredband.comhem.se> Message-ID: <9C66B29D904870B2F02EAEA2@srb.local> --On May 11, 2008 11:59:51 AM +0200 david wrote: > So you, as a regular user, can install programs (e.g., different spam > filters) on your mail server? In that case, you have access to a very > impressive mail setup. To which I reply: No, I can't install filtering software -- the mail server admins do that. I can install *rules* that control what that filtering software does with *my* mail (and only my mail). The important part about this is that the mail server admins are sufficiently advanced in their thinking that the filtering software that they install does one thing, and one thing only -- it tags incoming mail. It's up to my rules to decide how to act on those tags. The rest of this is not really relevant to Mulberry, so feel free to skip it... This is not to say that the umich mail folks don't take other actions. They do, for example, have some very active processes that happen at our mail gateways, including some very clever locally-developed gray listing. But once again, the goal is to err on the side of caution, and let possible badness through rather than take the chance of mis-handling non SPAM. Even then, it's possible for badly-configured sender mail systems to fail to deliver legitimate mail, but there's nothing on earth that can protect you from a badly-run mail system. I have a huge amount of respect for people who run large institutional mail systems well. One of my goals in life is to *not* have to run mail systems -- it's bloody hard work to do it well. And the evil-doers are out there in huge numbers, which makes the job even harder. -- Steve Burling University of Michigan, ICPSR Voice: +1 734 615.3779 330 Packard Street FAX: +1 734 647.8700 Ann Arbor, MI 48104-2910 From sombrero at mm.st Sun May 11 08:43:49 2008 From: sombrero at mm.st (Gleason Pace) Date: Sun, 11 May 2008 08:43:49 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> <4610357BC5CB05E3B4E51411@whitestar.local> Message-ID: <4826EA05.7020004@mm.st> david, > --On 2008-05-10 01.58 -0700 "Daniel M. Zimmerman" > wrote: > >> - realtime blackhole lists, which reject potential spam before it's ever >> sent to the receiving server in the first place >> - server-side spam scoring (which may factor in data from blackhole >> lists) where spam above a certain threshold is discarded >> - server-side spam scoring where spam above a certain threshold is >> sequestered in a separate mailbox >> - server-side spam scoring where spam is tagged and its disposition is >> left entirely to the client side app > > I agree with you. My point is that with client side filtering this is > a non-issue and if we want Mulberry to become more popular it can not > be the only client without client side filtering. It seems that many here are not concerned about Mulberry's popularity. Something else is more important to them. > Regarding Gmail, I have "non-spam" marked bounces for years, to no > avail. Since I even worked at Google I reported this problem in > Google's internal issue system which should give it higher priority > than external reports. -- Gleason From sombrero at mm.st Sun May 11 08:59:32 2008 From: sombrero at mm.st (Gleason Pace) Date: Sun, 11 May 2008 08:59:32 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <9C66B29D904870B2F02EAEA2@srb.local> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> <1F7D34683D20C21D829A721F@srb.local> <68FC75038D1D92C6FA7A5F7E@c80-216-42-109.bredband.comhem.se> <9C66B29D904870B2F02EAEA2@srb.local> Message-ID: <4826EDB4.4080301@mm.st> Steve Burling, > No, I can't install filtering software -- the mail server admins do > that. I can install *rules* that control what that filtering software > does with *my* mail (and only my mail). > > The important part about this is that the mail server admins are > sufficiently advanced in their thinking that the filtering software > that they install does one thing, and one thing only -- it tags > incoming mail. It's up to my rules to decide how to act on those tags. > > The rest of this is not really relevant to Mulberry, so feel free to > skip it... > > This is not to say that the umich mail folks don't take other actions. > They do, for example, have some very active processes that happen at > our mail gateways, including some very clever locally-developed gray > listing. But once again, the goal is to err on the side of caution, > and let possible badness through rather than take the chance of > mis-handling non SPAM. Even then, it's possible for badly-configured > sender mail systems to fail to deliver legitimate mail, but there's > nothing on earth that can protect you from a badly-run mail system. > > I have a huge amount of respect for people who run large institutional > mail systems well. One of my goals in life is to *not* have to run > mail systems -- it's bloody hard work to do it well. And the > evil-doers are out there in huge numbers, which makes the job even > harder. I'm glad you are happy with your well run mail server. Mulberry should not be limited to doing only the spam actions you would use in your situation. For me in my situation, I am not adverse to maybe accidentally reading a spam once in a while, or encountering them in my in box occasionally. Evil as they might be. But I understand that other poeple might be more sensitive, get rashes, etc. from such exposure. So yes, I think it would be ok for Mulberry to make them email happy too. Even if they don't have a well run mail server. -- Gleason From alex at alex.org.uk Sun May 11 09:38:08 2008 From: alex at alex.org.uk (Alex Bligh) Date: Sun, 11 May 2008 14:38:08 +0100 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <4826EDB4.4080301@mm.st> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> <1F7D34683D20C21D829A721F@srb.local> <68FC75038D1D92C6FA7A5F7E@c80-216-42-109.bredband.comhem.se> <9C66B29D904870B2F02EAEA2@srb.local> <4826EDB4.4080301@mm.st> Message-ID: --On 11 May 2008 08:59:32 -0400 Gleason Pace wrote: > But I understand that other poeple might be more sensitive, get rashes, > etc. from such exposure. So yes, I think it would be ok for Mulberry to > make them email happy too. Even if they don't have a well run mail > server. Send code. Alex From IYESEYTVMXTN at spammotel.com Sun May 11 09:49:39 2008 From: IYESEYTVMXTN at spammotel.com (IYESEYTVMXTN at spammotel.com) Date: Sun, 11 May 2008 09:49:39 -0400 Subject: [Mulberry-discuss] Returned mail: Transaction failed Message-ID: <1210514617.IYESEYTVMXTN@spammotel.com> From rhubbell at ihubbell.com Sun May 11 12:36:35 2008 From: rhubbell at ihubbell.com (rh) Date: Sun, 11 May 2008 09:36:35 -0700 Subject: [Mulberry-discuss] Bounces of messages to some recipients to this list Message-ID: <48272093.7050308@ihubbell.com> Hello all, (apologies if I should have sent this to an admin address, seems like there are a few here that deal with these sorts of things on this list) I get bounces from a couple of addresses when mail is sent to this list. I'm not sure if it's because of my domain or if it's due to mulberrymail.com. One address is on comcast and the other is a umich.edu address. The Sender of the bounce is mulberry-discuss-bounces at lists.mulberrymail.com I think I read somewhere that comcast uses block lists that are a bit zealous. From iane at sussex.ac.uk Mon May 12 10:47:41 2008 From: iane at sussex.ac.uk (Ian Eiloart) Date: Mon, 12 May 2008 15:47:41 +0100 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <4824BD70.2070208@mm.st> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> Message-ID: --On 9 May 2008 17:09:04 -0400 Gleason Pace wrote: > > Daniel M. Zimmerman, >>> What am I missing here? If unwanted email is rejected, don't false >>> positives dissappear without being seen? >> >> No, they don't "disappear without being seen"; they _bounce_, letting >> the sender know that something went wrong and to either fix _their_ >> problem or pursue an alternative contact method. > Good in an academic situation perhaps, where the person receiving the > message has less interest in successful delivery than the sender. If the > recipient doesn't want to make the sender think he _has_ a *problem*, > such treatment is inappropriate. In other words, most false positives I > see are inoccous enough. I would have a hard time figuring out what to > change myself. It is no good forcing my correspondents to struggle with > the ineptness of my spam filter, not in a business situation, anyway. > And I am just a simple programmer wanting to keep in effective touch with > users and buyers of my software. Yes, but the problem is how do you dig the false positives out of your spam mailbox? Studies have shown that spam filters can be more effective than humans at identifying spam. If you have to examine a spam mailbox, you're in trouble. With SMTP time rejection, the sender gets notified, and spam gets thrown away by the sending server. Use a generous filter, and you'll still get some spam in your INBOX, but little enough that it's manageable. Your false positive rate will be low, but your sender will be alerted. -- Ian Eiloart IT Services, University of Sussex x3148 From iane at sussex.ac.uk Mon May 12 10:50:06 2008 From: iane at sussex.ac.uk (Ian Eiloart) Date: Mon, 12 May 2008 15:50:06 +0100 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <5CDEE005561F7FEDFA76352B@D-128-208-244-210.dhcp4.washington.edu> <4824C55D.9060903@mm.st> <9D6783F42A454037752F9747@c80-216-42-109.bredband.comhem.se> Message-ID: --On 10 May 2008 10:31:09 +0200 david wrote: > Try that with e.g., Gmail - which, btw > is a good example of the problems with server side filtering: when I get > bounces on my Gmail-accounts because I have used an erroneous address > (typically misspelled) most of the time they end among the spam. Very > bad! And I can't do anything about it. They even hide emails that they've sent! I sent a message to a Google Group from an account in my domain hosted by Google. Google bounced the message, because I accidentally used the wrong account. Then they filtered their own bounce into my spam mailbox! -- Ian Eiloart IT Services, University of Sussex x3148 From sombrero at mm.st Mon May 12 11:32:03 2008 From: sombrero at mm.st (Gleason Pace) Date: Mon, 12 May 2008 11:32:03 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> Message-ID: <1905626949.20080512113203@mm.st> Ian, >> Good in an academic situation perhaps, where the person receiving the >> message has less interest in successful delivery than the sender. If the >> recipient doesn't want to make the sender think he _has_ a *problem*, >> such treatment is inappropriate. In other words, most false positives I >> see are inoccous enough. I would have a hard time figuring out what to >> change myself. It is no good forcing my correspondents to struggle with >> the ineptness of my spam filter, not in a business situation, anyway. >> And I am just a simple programmer wanting to keep in effective touch with >> users and buyers of my software. > Yes, but the problem is how do you dig the false positives out of your spam > mailbox? Studies have shown that spam filters can be more effective than > humans at identifying spam. Maybe they can, but I have tried a lot of them, and never found that to be the case. > If you have to examine a spam mailbox, you're in trouble. With SMTP time > rejection, the sender gets notified, Right, the whole thing should be entirely transparent to the sender. My spam filter is not their problem. Outside of an academic environment, that is. > and spam gets thrown away by the > sending server. Bad, bad. > Use a generous filter, and you'll still get some spam in > your INBOX, but little enough that it's manageable. Your false positive > rate will be low, but your sender will be alerted. No good. -- Gleason From iane at sussex.ac.uk Mon May 12 11:48:36 2008 From: iane at sussex.ac.uk (Ian Eiloart) Date: Mon, 12 May 2008 16:48:36 +0100 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <1905626949.20080512113203@mm.st> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <1905626949.20080512113203@mm.st> Message-ID: <17CCC7827B31B77958DB44B9@lewes.staff.uscs.susx.ac.uk> --On 12 May 2008 11:32:03 -0400 Gleason Pace wrote: >> and spam gets thrown away by the >> sending server. > > Bad, bad. > Really, you'd rather a spambot did something else with messages that it can't deliver? >> Use a generous filter, and you'll still get some spam in >> your INBOX, but little enough that it's manageable. Your false positive >> rate will be low, but your sender will be alerted. > > No good. > > Then you need to live without a spam filter. -- Ian Eiloart IT Services, University of Sussex x3148 From klensin+mulberry at jck.com Mon May 12 12:03:34 2008 From: klensin+mulberry at jck.com (John C Klensin) Date: Mon, 12 May 2008 12:03:34 -0400 Subject: [Mulberry-discuss] Custom addresses In-Reply-To: References: Message-ID: <4118B0DB2FB2F4CA325223A2@p3.JCK.COM> --On Sunday, 11 May, 2008 11:56:20 +0200, david wrote... > I like to tag the e-mail address I use in web forms and > similar with some keywords to indicate where I entered it as > well as what type of matter the message was about, for > example: > > davidbo+domain.com+bug at ... > > However, when I get replies on these mails and reply again > there seem to be no way to use this custom e-mail address > less than creating a new identity. Any ideas? Or must I file > a feature request? Well, I can tell you what does work but, IMO, the handling of subaddresses (as these things are broadly known) has never been satisfactory in Mulberry and an (another) feature request would be in order. You can use the "custom" address by going through the following sequence after initiating your reply: (1) Click on the From: line (this may require setting a flag somewhere; I don't remember). This will pull down a menu containing all of the identities you have defined. (2) Click on "Custom Identity". (3) A new icon will now appear to the right of the "From:" line that looks like a person with a pencil stuck into it. Click that icon. (4) You will now see a popup that looks suspiciously like the new identity creation one. Ignore the "Identity" line. Click the checkbox next to "Active" (5) Type in your name, if you care about personal names. (6) Type in your complete "custom" address. (7) Click "ok" and compose your message normally, with great confidence that the new "From:" address will be used in spite of the fact that you can't see it (the "From: line in the window continues to show "Custom Identity") except by going back through step (3) of the above process. That process can be made a bit more complicated by poor choices of options and defaults. Note that, if you do create new identities for each of your specialized addresses, what you probably want to happen (i.e., replying to an incoming message that reaches you via a custom address will put that address in the "From:" field of the reply) will happen most of the time, although there are some odd cases I've never understood. But you probably don't want to do that. I don't know about your situation, but I've got hundreds of subaddresses in active use. I don't know what would happen to Mulberry if I set up many hundreds of identities, but I suspect we wouldn't like it. Now, if that is too complicated or too error-prone for you, or if you feel that seeing "Custom Identity" in the "From:" line doesn't permit you to adequately confirm what you are doing, you need a feature request. I've always thought (and suggested in some detail several years ago) that Mulberry needs three options/features to support subaddresses at least reasonably well: (i) A "support subaddress in From" option that would change the "From:" lines of composed windows from From: [Named-identity v] (where "v" is the pull-down indicator) to From: [Named-identity][ ][v] (where that empty box was one in which the user could alternately just type "+foobar" or whatever). Note that this is real subaddress support, since subaddresses are attached to registered identities, rather than being treated as completely separate one-time-use identities. It would also permit users who are not permitted to change their own configurations (or identity list) to use subaddresses. One could make it a bit more robust by having the subaddress separator set in a configuration option so that one would type "foobar", rather than "+foobar" and have the separator supplied automatically. (ii) A "use incoming address in reply" option, such that, if one of the addresses in the To: or cc: headers of an incoming message matched a string in the "Outgoing addresses to count as mine" list (in the Identity Tab of Preferences), then that address would be copied into the "From:" field of a reply (using the subaddress form of the From field as above, so that the user could see, and override, the subaddress if needed). Note that this would take precedence over any "link identity" options -- those, and folder-default identities, would be effective on replies only if none of the incoming addresses matched. (iii) Finally, while this is separate enough from the above to possibly just be a bug fix, I believe that, if there is a custom identity and the material in the "From" tab of the identity box is active, then the personal name and address given should show up on the "From:" line of the composition box (either instead of or in addition to "Custom Identity..."). john From sombrero at mm.st Mon May 12 12:14:49 2008 From: sombrero at mm.st (Gleason Pace) Date: Mon, 12 May 2008 12:14:49 -0400 Subject: [Mulberry-discuss] plugin for bayesian spam control In-Reply-To: <17CCC7827B31B77958DB44B9@lewes.staff.uscs.susx.ac.uk> References: <7396C254BC6EEFB5B636128E@[192.168.1.121]> <9623ED89AC7108E2D69FCA91@knockando> <48245A3C.9070406@mm.st> <5EABDC838ABB566256258EF8@babylon3.tffenterprises.com> <4824BD70.2070208@mm.st> <1905626949.20080512113203@mm.st> <17CCC7827B31B77958DB44B9@lewes.staff.uscs.susx.ac.uk> Message-ID: <319578514.20080512121449@mm.st> Ian, >>> and spam gets thrown away by the >>> sending server. >> >> Bad, bad. >> > Really, you'd rather a spambot did something else with messages that it > can't deliver? Give me the spambot that can deliver that message. >>> Use a generous filter, and you'll still get some spam in >>> your INBOX, but little enough that it's manageable. Your false positive >>> rate will be low, but your sender will be alerted. >> >> No good. >> >> > Then you need to live without a spam filter. I can do fine with the client side spam filter that throws nothing away. -- Gleason Using 4.0.24.8 on Windows XP, 5.1, Build 2600. IMAP email provider is Fastmail, which uses Cyrus server software. From klensin+mulberry at jck.com Tue May 13 16:36:38 2008 From: klensin+mulberry at jck.com (John C Klensin) Date: Tue, 13 May 2008 16:36:38 -0400 Subject: [Mulberry-discuss] Setting up a search condition or rule on an additional header Message-ID: Hi. I'm trying to set up search (or rule selection) criteria based on the contents of the Return-path header field. That field isn't in the list of headers in the usual pull-down (the canonical IMAP set). I should be able to access this via "Header Field", but that gives me two blank boxes. I can find nothing about the Header Field option in the manual or on the wiki. It appears from the tool tips that show up in Search (but not in Rules) that the first of these is the field name and the second is text to be searched for. After a bit of experimentation, it appears that the field name should be specified without the colon. But I'd think that either there should be a third box with a "Contains"/"Does not contain" option (like other text fields) or that the second box should be a full regular expression in which I can, e.g., get "Does not contain" with "^[string]" or equivalent. In case it is relevant and not obvious, I'm trying to isolate delivery notifications and the like with a "Return-path Equals '<>'" rule. Am I correctly understanding this? Do we need a feature request, or just more documentation? john From Hagedorn at uni-koeln.de Wed May 14 08:10:08 2008 From: Hagedorn at uni-koeln.de (Sebastian Hagedorn) Date: Wed, 14 May 2008 14:10:08 +0200 Subject: [Mulberry-discuss] Setting up a search condition or rule on an additional header In-Reply-To: References: Message-ID: Hi, --On 13. Mai 2008 16:36:38 -0400 John C Klensin wrote: > I'm trying to set up search (or rule selection) criteria based > on the contents of the Return-path header field. that isn't really a header field at all. It's not part of the actual message. Actually it's the sender from the message's envelope. Some mail systems save the return-path on final delivery, but I don't think you can count on that. According to RFC 2822, it's an optional "trace field": > That field > isn't in the list of headers in the usual pull-down (the > canonical IMAP set). Which should be no suprise, given that it's no proper header. > I should be able to access this via > "Header Field", but that gives me two blank boxes. I can find > nothing about the Header Field option in the manual or on the > wiki. > > It appears from the tool tips that show up in Search (but not in > Rules) that the first of these is the field name and the second > is text to be searched for. Exactly. > After a bit of experimentation, it > appears that the field name should be specified without the > colon. That's correct. > But I'd think that either there should be a third box > with a "Contains"/"Does not contain" option (like other text > fields) or that the second box should be a full regular > expression in which I can, e.g., get "Does not contain" with > "^[string]" or equivalent. The IMAP Specification does not support that, According to RFC 3501: "HEADER Messages that have a header with the specified field-name (as defined in [RFC-2822]) and that contains the specified string in the text of the header (what comes after the colon). If the string to search is zero-length, this matches all messages that have a header line with the specified field-name regardless of the contents." > In case it is relevant and not obvious, I'm trying to isolate > delivery notifications and the like with a > "Return-path Equals '<>'" > rule. *If* Return-Path can be searched, why don't you just put <> in the second field?? Update: it *is* possible. I just tested it: . search header "return-path" "<>" * SEARCH 33 . OK Completed (1 msgs in 0.010 secs) > Am I correctly understanding this? Do we need a feature > request, or just more documentation? Perhaps the latter. The former doesn't seem possible. -- .:.Sebastian Hagedorn - RZKR-R1 (Geb?ude 52), Zimmer 18.:. Zentrum f?r angewandte Informatik - Universit?tsweiter Service RRZK .:.Universit?t zu K?ln / Cologne University - ? +49-221-478-5587.:. .:.:.:.Skype: shagedorn.:.:.:. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 186 bytes Desc: not available Url : http://lists.mulberrymail.com/pipermail/mulberry-discuss/attachments/20080514/08c9af78/attachment.bin From klensin+mulberry at jck.com Wed May 14 09:17:29 2008 From: klensin+mulberry at jck.com (John C Klensin) Date: Wed, 14 May 2008 09:17:29 -0400 Subject: [Mulberry-discuss] Setting up a search condition or rule on an additional header In-Reply-To: References: Message-ID: <40B99E54ACF70CB224EB728C@p3.JCK.COM> --On Wednesday, 14 May, 2008 14:10 +0200 Sebastian Hagedorn wrote: > Hi, > > --On 13. Mai 2008 16:36:38 -0400 John C Klensin > wrote: > >> I'm trying to set up search (or rule selection) criteria based >> on the contents of the Return-path header field. > > that isn't really a header field at all. It's not part of the > actual message. Actually it's the sender from the message's > envelope. Some mail systems save the return-path on final > delivery, but I don't think you can count on that. According > to RFC 2822, it's an optional "trace field": > > That is an interesting reading of 2822. Given the requirement in 2821 that the Return-path be inserted (a MUST), there is no sense in which Return-path is optional at the time of delivery. Certainly something between the message delivery process, the message store, and the IMAP client could delete it, but they could also delete the other trace information (and that is very much part of the canonical IMAP form issue). I've copied you on a fairly lengthy note about this issue and 2822upd; I won't bore the rest of the list with it. >> That field >> isn't in the list of headers in the usual pull-down (the >> canonical IMAP set). > > Which should be no suprise, given that it's no proper header. It should be no surprise given that the canonical IMAP set was defined the way it was defined. I believe it is entirely appropriate to scan the entire header set to examine/ test "Return-path", just as I would have to scan the entire header set to examine, e.g., X"-Bogus-Stuff". But that is precisely what the "Header Field" option is supposed to be for. >> I should be able to access this via >> "Header Field", but that gives me two blank boxes. I can >> find nothing about the Header Field option in the manual or >> on the wiki. >> >> It appears from the tool tips that show up in Search (but not >> in Rules) that the first of these is the field name and the >> second is text to be searched for. > > Exactly. > >> After a bit of experimentation, it >> appears that the field name should be specified without the >> colon. > > That's correct. > >> But I'd think that either there should be a third box >> with a "Contains"/"Does not contain" option (like other text >> fields) or that the second box should be a full regular >> expression in which I can, e.g., get "Does not contain" with >> "^[string]" or equivalent. > > The IMAP Specification does not support that, According to RFC > 3501: > > "HEADER > Messages that have a header with the specified > field-name (as > defined in [RFC-2822]) and that contains the > specified string > in the text of the header (what comes after the > colon). If the > string to search is zero-length, this matches all > messages that > have a header line with the specified field-name > regardless of > the contents." That is relevant if this test is performed on the server. But it appears clear from watching what is going on that Mulberry is, at least with the server I'm using, downloading the entire header set and making the test locally. If it is, it can, in principle, make any test it likes. >> In case it is relevant and not obvious, I'm trying to isolate >> delivery notifications and the like with a >> "Return-path Equals '<>'" >> rule. > > *If* Return-Path can be searched, why don't you just put <> in > the second field?? I did. > Update: it *is* possible. I just tested it: > > . search header "return-path" "<>" > * SEARCH 33 > . OK Completed (1 msgs in 0.010 secs) Yes, this works fine. If I put the same construction in a filtering Rule, it also works if I select a range and then explicitly apply the rule (although I see a lot of "message downloading" action). However, if I bind the rule to an incoming message trigger on a particular folder (a trigger that works fine with other rules), nothing happens. So something peculiar is going on that, at least, needs better description. >> Am I correctly understanding this? Do we need a feature >> request, or just more documentation? > > Perhaps the latter. The former doesn't seem possible. See above. john From daboo at mulberrymail.com Wed May 14 09:37:39 2008 From: daboo at mulberrymail.com (Cyrus Daboo) Date: Wed, 14 May 2008 09:37:39 -0400 Subject: [Mulberry-discuss] Setting up a search condition or rule on an additional header In-Reply-To: <40B99E54ACF70CB224EB728C@p3.JCK.COM> References: <40B99E54ACF70CB224EB728C@p3.JCK.COM> Message-ID: <307CBD0A5E7CC01BF46A161E@caldav.corp.apple.com> Hi John, --On May 14, 2008 9:17:29 AM -0400 John C Klensin wrote: >> The IMAP Specification does not support that, According to RFC >> 3501: >> >> "HEADER >> Messages that have a header with the specified >> field-name (as >> defined in [RFC-2822]) and that contains the >> specified string >> in the text of the header (what comes after the >> colon). If the >> string to search is zero-length, this matches all >> messages that >> have a header line with the specified field-name >> regardless of >> the contents." > > That is relevant if this test is performed on the server. But > it appears clear from watching what is going on that Mulberry > is, at least with the server I'm using, downloading the entire > header set and making the test locally. If it is, it can, in > principle, make any test it likes. The search criteria in rules should be executing on the server. As regards a "does not contain" operator, the IMAP spec defines a NOT operator that can be applied before the HEADER test to negate its behavior. Mulberry's UI for the generic header test does not have a 'contains/does not contain' popup like the other tests for the simple reason that there was not enough room to put one in given that the header name had to be an editable field. So originally Mulberry always defaulted to a "contains" for that. Several people complained about that so a "hidden" (undocumented) option was added whereby if you prefix the header name with a single "!" character, a 'does not contain" operation is done instead. That should at the very least have an FAQ. >>> In case it is relevant and not obvious, I'm trying to isolate >>> delivery notifications and the like with a >>> "Return-path Equals '<>'" >>> rule. >> >> *If* Return-Path can be searched, why don't you just put <> in >> the second field?? > > I did. > >> Update: it *is* possible. I just tested it: >> >> . search header "return-path" "<>" >> * SEARCH 33 >> . OK Completed (1 msgs in 0.010 secs) > > Yes, this works fine. If I put the same construction in a > filtering Rule, it also works if I select a range and then > explicitly apply the rule (although I see a lot of "message > downloading" action). However, if I bind the rule to an > incoming message trigger on a particular folder (a trigger that > works fine with other rules), nothing happens. So something > peculiar is going on that, at least, needs better description. > Can you run with IMAP logging on and check to see whether the proper SEARCH command is being executed when the rule triggers? -- Cyrus Daboo From Hagedorn at uni-koeln.de Wed May 14 09:41:55 2008 From: Hagedorn at uni-koeln.de (Sebastian Hagedorn) Date: Wed, 14 May 2008 15:41:55 +0200 Subject: [Mulberry-discuss] Setting up a search condition or rule on an additional header In-Reply-To: <40B99E54ACF70CB224EB728C@p3.JCK.COM> References: <40B99E54ACF70CB224EB728C@p3.JCK.COM> Message-ID: --On 14. Mai 2008 09:17:29 -0400 John C Klensin wrote: >> that isn't really a header field at all. It's not part of the >> actual message. Actually it's the sender from the message's >> envelope. Some mail systems save the return-path on final >> delivery, but I don't think you can count on that. According >> to RFC 2822, it's an optional "trace field": >> >> > > That is an interesting reading of 2822. Given the requirement > in 2821 that the Return-path be inserted (a MUST), Thanks, I wasn't aware of that. -- .:.Sebastian Hagedorn - RZKR-R1 (Geb?ude 52), Zimmer 18.:. Zentrum f?r angewandte Informatik - Universit?tsweiter Service RRZK .:.Universit?t zu K?ln / Cologne University - ? +49-221-478-5587.:. .:.:.:.Skype: shagedorn.:.:.:. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 186 bytes Desc: not available Url : http://lists.mulberrymail.com/pipermail/mulberry-discuss/attachments/20080514/179897f2/attachment.bin From klensin+mulberry at jck.com Wed May 14 09:49:20 2008 From: klensin+mulberry at jck.com (John C Klensin) Date: Wed, 14 May 2008 09:49:20 -0400 Subject: [Mulberry-discuss] Setting up a search condition or rule on an additional header In-Reply-To: <307CBD0A5E7CC01BF46A161E@caldav.corp.apple.com> References: <40B99E54ACF70CB224EB728C@p3.JCK.COM> <307CBD0A5E7CC01BF46A161E@caldav.corp.apple.com> Message-ID: <80516A97C8B4ABEA8353E4EF@p3.JCK.COM> --On Wednesday, 14 May, 2008 09:37 -0400 Cyrus Daboo wrote: > Hi John, > > --On May 14, 2008 9:17:29 AM -0400 John C Klensin > wrote: > >... > The search criteria in rules should be executing on the server. > > As regards a "does not contain" operator, the IMAP spec > defines a NOT operator that can be applied before the HEADER > test to negate its behavior. Mulberry's UI for the generic > header test does not have a 'contains/does not contain' popup > like the other tests for the simple reason that there was not > enough room to put one in given that the header name had to be > an editable field. So originally Mulberry always defaulted to > a "contains" for that. Several people complained about that so > a "hidden" (undocumented) option was added whereby if you > prefix the header name with a single "!" character, a 'does > not contain" operation is done instead. That should at the > very least have an FAQ. I will try writing it if you or Sebastian will check and edit. > Can you run with IMAP logging on and check to see whether the > proper SEARCH command is being executed when the rule triggers? Yep. Later today. thanks, john From quanah at fast-mail.org Thu May 15 14:03:26 2008 From: quanah at fast-mail.org (Quanah Gibson-Mount) Date: Thu, 15 May 2008 11:03:26 -0700 Subject: [Mulberry-discuss] Local cache for IMAP? Message-ID: <52048EA116B78C984D624163@[192.168.1.198]> Does Mulberry have a local cached set of data for IMAP folders? If so, does this cache persist across restarts? And if so, what can be done to delete it so it gets reset? Thanks, Quanah From Hagedorn at uni-koeln.de Thu May 15 14:11:15 2008 From: Hagedorn at uni-koeln.de (Sebastian Hagedorn) Date: Thu, 15 May 2008 20:11:15 +0200 Subject: [Mulberry-discuss] Local cache for IMAP? In-Reply-To: <52048EA116B78C984D624163@[192.168.1.198]> References: <52048EA116B78C984D624163@[192.168.1.198]> Message-ID: <302B0B52F00C6DBCD8B7FA3B@G5> -- Quanah Gibson-Mount is rumored to have mumbled on 15. Mai 2008 11:03:26 -0700 regarding [Mulberry-discuss] Local cache for IMAP?: > Does Mulberry have a local cached set of data for IMAP folders? Only if you use disconnected mode. > If so, does this cache persist across restarts? If you use disconnected mode: yes. > And if so, what can be > done to delete it so it gets reset? I guess you could delete the mailboxes in "Disconnected", but I've never heard of issues with them. AFAIK Mulberry never uses the cached data when you're connected, only when you're disconnected. -- Sebastian Hagedorn - RZKR-R1 (Flachbau), Zi. 18, Robert-Koch-Str. 10 Zentrum f?r angewandte Informatik - Universit?tsweiter Service RRZK Universit?t zu K?ln / Cologne University - Tel. +49-221-478-5587 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 194 bytes Desc: not available Url : http://lists.mulberrymail.com/pipermail/mulberry-discuss/attachments/20080515/5bbb9ecd/attachment-0001.bin From quanah at fast-mail.org Thu May 15 14:26:08 2008 From: quanah at fast-mail.org (Quanah Gibson-Mount) Date: Thu, 15 May 2008 11:26:08 -0700 Subject: [Mulberry-discuss] Local cache for IMAP? In-Reply-To: <302B0B52F00C6DBCD8B7FA3B@G5> References: <52048EA116B78C984D624163@[192.168.1.198]> <302B0B52F00C6DBCD8B7FA3B@G5> Message-ID: --On Thursday, May 15, 2008 8:11 PM +0200 Sebastian Hagedorn wrote: > -- Quanah Gibson-Mount is rumored to have mumbled > on 15. Mai 2008 11:03:26 -0700 regarding [Mulberry-discuss] Local cache > for IMAP?: > >> Does Mulberry have a local cached set of data for IMAP folders? > > Only if you use disconnected mode. > >> If so, does this cache persist across restarts? > > If you use disconnected mode: yes. > >> And if so, what can be >> done to delete it so it gets reset? > > I guess you could delete the mailboxes in "Disconnected", but I've never > heard of issues with them. AFAIK Mulberry never uses the cached data when > you're connected, only when you're disconnected. Great, thanks. I never use disconnected mode. :) --Quanah From hone+mulberrydiscuss at oak.cats.ohiou.edu Thu May 22 14:05:54 2008 From: hone+mulberrydiscuss at oak.cats.ohiou.edu (Don Hone) Date: Thu, 22 May 2008 14:05:54 -0400 Subject: [Mulberry-discuss] Mac: Opening .doc files w/ Word 08 Message-ID: I have a customer who cannot open Word 04 attachments (.doc) from within Mulberry on a Mac. This person has Word 08 installed. The only way to open is to save the attachment, open Word 08, and file -> open. The same computer can open the same document from within Mac Mail. Can this be fixed from a system setting or from within Mulberry? Don Hone Ohio University From dave at scocca.org Thu May 22 14:12:01 2008 From: dave at scocca.org (Dave Scocca) Date: Thu, 22 May 2008 14:12:01 -0400 Subject: [Mulberry-discuss] Mac: Opening .doc files w/ Word 08 In-Reply-To: References: Message-ID: <4E389131759BFA5864F71585@[10.8.2.178]> --On 5/22/2008 2:05 PM -0400 Don Hone wrote: > > I have a customer who cannot open Word 04 attachments (.doc) from within > Mulberry on a Mac. This person has Word 08 installed. The only way to > open is to save the attachment, open Word 08, and file -> open. The same > computer can open the same document from within Mac Mail. I think this may be a known problem with Word 08/service pack 1, which prevents it from opening certain kinds of files in the finder: http://tinyurl.com/6qfyrt See the tail end of the thread at: http://emperor.tidbits.com/webx?50@@.3cc2ccd6 for a few more links on the issue--particularly Diane Ross's message #11 in the thread. Dave Scocca From pwilson at apnic.net Thu May 22 18:57:27 2008 From: pwilson at apnic.net (Paul Wilson) Date: Fri, 23 May 2008 08:57:27 +1000 Subject: [Mulberry-discuss] Mac: Opening .doc files w/ Word 08 In-Reply-To: References: Message-ID: <9EEF498AEC069D511E79F288@as-paul-l-1097.local> I have a propblem which may be related. Often I receive Word attachements, which are shown in Mulberry as MIMEtype "application/applefile". If I open or save them, I get a 0-length result. This has happened only since moving to Mac OS from Windoze, and I have not been able to investigate it further so far. Paul --On 22 May 2008 2:05:54 PM -0400 Don Hone wrote: > > I have a customer who cannot open Word 04 attachments (.doc) from within > Mulberry on a Mac. This person has Word 08 installed. The only way to > open is to save the attachment, open Word 08, and file -> open. The same > computer can open the same document from within Mac Mail. > > Can this be fixed from a system setting or from within Mulberry? > > Don Hone > > Ohio University From pschmehl_lists at tx.rr.com Tue May 27 12:54:28 2008 From: pschmehl_lists at tx.rr.com (Paul Schmehl) Date: Tue, 27 May 2008 11:54:28 -0500 Subject: [Mulberry-discuss] Future of mulberry Message-ID: I'm at a crossroads. I need to decide whether or not I should stick with Mulberry or start looking for a different email client. Three issues are driving me; handling of S/MIME certificates, handling of POP mail and stability. I run Mulberry for LInux under emulation of FreeBSD, so I don't know if my problems are unique to that setup or common for all Linux users. 1) S/MIME - this problem isn't going away until openssl updates its code to handle the new types of certs that Verisign is issuing, and it's not unique to Mulberry. So it's not really a "deal-killer". 2) POP mail - for some reason, even though I've selected the option of remove mail from the server on local expunge, it fails on the Linux client running under emulation. It works fine on Mulberry on my Mac at home as well as on Windows, so maybe it's just a unique problem associated with emulation. This isn't a deal-killer either, but it is an irritation. 3) Stability - I'm having huge problems running Mulberry as I would like to. E.g. if I click on a link in an email (either a mailto or an href) Mulberry core dumps. If I scroll through a list of messages rapidly, and I go past a message with a digital signature attached, Mulberry core dumps. This is really making life difficult for me. Combined with the other issues, it's getting frustrating, to put it mildly. OTOH, I love Mulberry. It has features that don't exist in any other client that I'm aware of. So, my question is, is anybody working on a "native" unix client that would compile on several platforms? I would much prefer a FreeBSD client to the emulated Linux client I have to use now. I'd offer to help, but I am not a programmer. The best I can do is muddle through compiling and making and report problems when I encounter them. It seems like not much of anything is happening on the development end of things. Is that a false impression? I don't want to give up Mulberry, but I'm rapidly approaching the point where I may have to, on unix. -- Paul Schmehl As if it wasn't already obvious, my opinions are my own and not those of my employer. From pwilson at apnic.net Tue May 27 13:34:42 2008 From: pwilson at apnic.net (Paul Wilson) Date: Wed, 28 May 2008 03:34:42 +1000 Subject: [Mulberry-discuss] Future of mulberry In-Reply-To: References: Message-ID: <3DC66B4E6F298542F8861AF7@as-paul-l-1097.local> I have the same dilemma as Paul, running Mulberry on Mac OSX. Frequent crashes and various bugs are frustrating; but my main issue is that offline mode does not work. Offline updates are not properly synchronised - often being totally abandoned. As the fantastic offline functionality was my main reason for using Mulberry in the first place (under Windows initially) I really can't justify the frustrations without some light on the horizon. Paul. --On 27 May 2008 11:54:28 AM -0500 Paul Schmehl wrote: > I'm at a crossroads. I need to decide whether or not I should stick with > Mulberry or start looking for a different email client. Three issues are > driving me; handling of S/MIME certificates, handling of POP mail and > stability. > > I run Mulberry for LInux under emulation of FreeBSD, so I don't know if > my problems are unique to that setup or common for all Linux users. > > 1) S/MIME - this problem isn't going away until openssl updates its code > to handle the new types of certs that Verisign is issuing, and it's not > unique to Mulberry. So it's not really a "deal-killer". > > 2) POP mail - for some reason, even though I've selected the option of > remove mail from the server on local expunge, it fails on the Linux > client running under emulation. It works fine on Mulberry on my Mac at > home as well as on Windows, so maybe it's just a unique problem > associated with emulation. This isn't a deal-killer either, but it is an > irritation. > > 3) Stability - I'm having huge problems running Mulberry as I would like > to. E.g. if I click on a link in an email (either a mailto or an href) > Mulberry core dumps. If I scroll through a list of messages rapidly, and > I go past a message with a digital signature attached, Mulberry core > dumps. This is really making life difficult for me. Combined with the > other issues, it's getting frustrating, to put it mildly. > > OTOH, I love Mulberry. It has features that don't exist in any other > client that I'm aware of. > > So, my question is, is anybody working on a "native" unix client that > would compile on several platforms? I would much prefer a FreeBSD client > to the emulated Linux client I have to use now. I'd offer to help, but I > am not a programmer. The best I can do is muddle through compiling and > making and report problems when I encounter them. > > It seems like not much of anything is happening on the development end of > things. Is that a false impression? I don't want to give up Mulberry, > but I'm rapidly approaching the point where I may have to, on unix. > > > -- > Paul Schmehl > As if it wasn't already obvious, > my opinions are my own and not > those of my employer. > From pschmehl_lists at tx.rr.com Wed May 28 14:15:30 2008 From: pschmehl_lists at tx.rr.com (Paul Schmehl) Date: Wed, 28 May 2008 13:15:30 -0500 Subject: [Mulberry-discuss] S/MIME Message-ID: <09CBBA57B4F6A1B183063D55@utd65257.utdallas.edu> I finally resolved the problem of importing certs into mulberry. Turns out the certs I was trying to import had been encrypted by Safeboot somehow. A fresh export/import solved the problem. However, I had to downgrade mulberry to version 4.0.6 to make use of the certs. Neither 4.0.7 nor 4.0.8 loads the S/MIME plugin. I never know if I should log these as bugs or not, because I'm running the Linux binary under emulation on FreeBSD. So, what I'm seeing might not be an accurate representation of the app's behavior in the environment under which it's built to run. -- Paul Schmehl As if it wasn't already obvious, my opinions are my own and not those of my employer. From davidbo at kth.se Sat May 31 16:12:05 2008 From: davidbo at kth.se (david) Date: Sat, 31 May 2008 22:12:05 +0200 Subject: [Mulberry-discuss] Future of mulberry In-Reply-To: <3DC66B4E6F298542F8861AF7@as-paul-l-1097.local> References: <3DC66B4E6F298542F8861AF7@as-paul-l-1097.local> Message-ID: <94D31699F893FC3958AD0B1F@c80-216-44-178.bredband.comhem.se> --On 2008-05-28 03.34 +1000 Paul Wilson wrote: > I have the same dilemma as Paul, running Mulberry on Mac OSX. Frequent > crashes I also experience these crashes on my Mac. I created a report on Trac with many crash logs attached. Hopefully someone will be able to work something out of these data. > and various bugs are frustrating; but my main issue is that > offline mode does not work. Offline updates are not properly > synchronised - often being totally abandoned. As the fantastic offline > functionality was my main reason for using Mulberry in the first place > (under Windows initially) I really can't justify the frustrations without > some light on the horizon. I have no problems with offline mode. In the beginning it was very counterintuitive but when I started to understand the design offline mode has worked fine. What is your exact problem? -- Visit the Mulberry IRC-channel irc://irc.freenode.net/mulberry