|
Using SpamBox for E-Mail Spam Filtering
S4314 July 2006
This document provides instructions for using ITS's content filtering tool for spam, SpamBox. SpamBox keeps most spam messages out of your INBOX by filtering them into a separate folder for you, where you can do a quick check for any misidentified messages before deleting the spam.
Table of ContentsWhat SpamBox DoesSpamBox checks your incoming messages as they come in to the ITS e-mail servers. It uses Bayesian and other statistical algorithms to compare the content of the messages (broken down into text strings called "tokens") to the information in a "dictionary" that it maintains. It tags messages it suspects to be spam (by adding some text to the message header).
SpamBox is very accurate, but it will occasionally make a mistake when figuring out which messages are spam and which are innocent. For example, when spammers try a new technique, SpamBox may not recognize the messages as spam until it has been programmed to "learn" what they are. Also, if an innocent message has characteristics of mail that is typically spam, such as HTML code or embedded images, it may be misidentified as spam.
Turn SpamBox On Via the WebRequirement
WARNING! Some U-M departments provide their own e-mail services. Before turning SpamBox on, please check with your department's information technology help desk to see if you are currently using one of these services. Many departmental e-mail services already provide spam filtering. Using the SpamBox Setup Wizard
When you turn SpamBox on for your ITS IMAP e-mail mailbox, three things happen:
NOTE TO POP USERS: If, when you turn SpamBox on, you indicate that you use a POP e-mail program, only the first of these three things happens. TIP FOR PINE USERS: If you would like to be able to check for messages in your SpamBox folder by pressing the Tab key, add this to the incoming-folders line in your .pinerc file: incoming-folders={mail.umich.edu}SpamBox Your New SpamBox FolderOnce SpamBox has been enabled for your ITS IMAP e-mail account, you will notice a new folder/mailbox in your folder list called SpamBox. Any incoming messages identified as likely to be spam will be put there. You will need to regularly look in this folder to check for any non-spam messages mistakenly placed there and to delete the spam messages.If You Accidentally Delete the SpamBox FolderIf you should inadvertently delete the SpamBox folder, you can create a new one yourself. Be sure to name it SpamBox, making the name all one word and matching the capitalization and lower case letters exactly. The server-side filter will not be able to put your suspected spam messages in the folder if the folder name is different in any way.Tips for Making Your SpamBox Folder Appear Using Various E-Mail ProgramsIf you do not see your SpamBox folder after turning SpamBox on, you may need to tell your e-mail program how to find and display it. Here are some ways to do that:
Apple's Mail Application (IMAP)
Entourage for Mac OS X (IMAP)
Eudora (IMAP) for WindowsRight-click the dominant folder and select Refresh Mailbox List.
Mozilla Thunderbird
Mulberry (IMAP)If you use Mulberry on more than one computer and can see a SpamBox folder from one computer but not the other, follow these instructions. On the computer where the SpamBox folder does not appear:
Netscape 7 Mail
Outlook (IMAP)You may need to subscribe to the SpamBox folder to see it.
Outlook Express (IMAP)You may need to subscribe to the SpamBox folder to see it.
Other ProgramsAlthough ITS only provides support for the Mulberry, Pine, and web-based mail.umich.edu programs, we are happy to include tips for other programs in this document. If you have a tip for getting the SpamBox folder to display using another mail program, please send it to itcs.doc.comments@umich.edu.
It Is Important to Check Your SpamBox Folder RegularlyCheck your SpamBox folder regularly for these reasons:
Keeping Mail from a Particular Address Out of Your SpamBox FolderIf you would like to be sure that mail from a particular address (perhaps that of your parents or your spouse) will never get mistakenly placed in your SpamBox folder, you can do this by adding that address to your Accept List. Then all mail from that address will be put in your INBOX regardless of whether SpamBox "thinks" the mail looks like spam or not. See this document for details:
Turning SpamBox OffIf you later decide that you do not want to use SpamBox, you can turn it off.
Reporting SpamBox ErrorsYou do not need to report the errors that SpamBox makes. If, however, you would like to help us continually improve its ability to identify spam, here are instructions for reporting errors.IMPORTANT! To report errors, you must use the web-based web.mail.umich.edu. When you turn SpamBox on, spam-reporting links are added to web.mail.umich.edu for you. To report spam in your INBOX
To report innocent mail in your SpamBox Folder
IMPORTANT! Web.mail.umich.edu can only recognize one folder for spam mail. If you change your web.mail.umich.edu preferences to select a folder other than SpamBox for your spam folder, please be aware that the This is Not Spam link will no longer be available inside your SpamBox folder. Instead, that link will appear inside whatever folder you designate as your spam folder. Additional ResourcesVisit ITS's Information System to obtain ITS computer documentation and other resources. A list of relevant documents follows:
Creating POP Mail Filters/Rules to Work with SpamBox (S4324) We welcome your comments; please send e-mail. ITS's Online Help Desk provides a variety of computing help resources. For further help with SpamBox, send e-mail or phone (734) 764-HELP.
Appendix A: Frequently Asked QuestionsWhy can't I see my SpamBox folder?There are two possible reasons for this:
Can SpamBox be fixed to filter a specific address/keyword/domain? Why did this message get identified as spam and that similar message go into my INBOX?
If you'd like to learn more about the Bayesian approach to spam filtering, these web pages are good resources: http://paulgraham.com/spam.html and http://paulgraham.com/better.html When spammers try a new tactic, SpamBox may miss some of the new spam that is sent until information about it is added to its central spam information. That information is updated regularly.
The directory forwards my mail to my ITS mailbox and to my Comcast (or other) address. Why is all the spam still going to my Comcast address?
One copy goes off to your Comcast address, and the other goes to the machine where your U-M mail resides -- one of the ITS mail servers. SpamBox is installed on that machine. Therefore, it can only work on mail that goes to your ITS mailbox.
Can I make Apple's Mail Application tell me when there are messages in my SpamBox folder?
I see my SpamBox folder in mail.umich.edu (or some other program), but not in Pine. Why not?
Appendix B: The Software Behind SpamBox -- DSPAM, Ingo, and SieveThe spam-identifying software behind SpamBox is DSPAM. DSPAM is open source software available to the University at no charge. ITS staff members participate with others in the open source community to continually improve DSPAM.Technical details about DSPAM are on the Nuclear Elephant web site. DSPAM is set to tag mail as either spam or innocent. It works in conjunction with server-side filters to have the tagged mail sorted into your INBOX and SpamBox folder. The SpamBox setup wizard uses Ingo software (which is actually an application inside IMP, which is the software behind the web-based mail.umich.edu) as an interface to the Sieve mail filtering language to create a mail filter on the ITS mail servers. Technical information about IMP and Ingo is available in the Projects section of the Horde web site. The Horde Project is responsible for this open source software.
|