You are viewing a plain text version of this content. The canonical link for it is here.
Posted to site-dev@james.apache.org by Apache Wiki <wi...@apache.org> on 2007/01/08 14:39:10 UTC

[James Wiki] Update of "MailetIdeas" by DannyAngus

Dear Wiki user,

You have subscribed to a wiki page or wiki category on "James Wiki" for change notification.

The following page has been changed by DannyAngus:
http://wiki.apache.org/james/MailetIdeas

------------------------------------------------------------------------------
  
   *  '''Tag Message Delivery Agent''' - Java implementation of the Tagged Message Delivery Agent (TMDA).. http://tmda.net/  an excellent SPAM filter / processor
  
-  *  '''Mailing lists''' -> duh.
+  *  '''Mailing lists''' - James already has support for Mailing Lists
-    *  http://www.li18nux.org/subgroups/majavdomo/ -- Japanese (PureJava MLMS -- Majavdomo)
+    *  ''this seems to be defunct, 404 error.'' http://www.li18nux.org/subgroups/majavdomo/ -- Japanese (PureJava MLMS -- Majavdomo)
  
   *  '''HTML-to-text''' - mailet could look for HTML emails and automatically remove this markup rich copy so plain text mail clients (like pine) get a message that's easier for them to read.
  
@@ -22, +22 @@

     *  '''Message signature''' - some systems will hash the message to get a quick "signature" of a spam message, and then block further copies of these.
       * Done - see DCCIFD, this is the same principle
     *  '''Bayesian filtering''' - probably the most promising spam filtering, this learns what are frequent words that appear in spam and not spam, and then checks the frequency of these words in an incoming message to determine whether it's spam or not.  Quite effective!!
+      * Available in James 2.3
     *  '''Usage Limitations''' - Limit number of messages an authenticated user can send (daily or weekly) - this would keep users on a system from using it for mass-mailings.
     *  '''Pornography Filtering''' -  11/27/2000 CNN.com article describes how a filter checks for attachment images, opens them and counts the percentage of color pallete in the skin tone, and high amounts generally means it's a porn image.