Evolution SpamBayes Plugin is a filter system that integrates to Evolution to provide better email filtering experience based on the bayesian algorithm offered by SpamBayes.
The Evolution SpamBayes Plugin is a fantastic addition to Evolution that allows for the use of the SpamBayes Bayesian filtering system. This plugin is based on the Bogofilter EPlugin by Mikhail Zabaluev but also incorporates some "pass message to process" code from the built-in SpamAssassin filtering plugin by Vivek Jain.
This plugin is easy to implement using two python support scripts. One sets up a DBus Server that holds the SpamBayes database. The other queries the client and trains or untrains on messages received on stdin. The Evolution plugin itself is simple and implemented as a C EPlugin for Evolution. However, it uses some junk-mail hooks in the evolution EPlugin architecture, which is not yet published on the web for unknown reasons.
One downside to the plugin is that messages that are not marked as spam have no way of being marked as ham. One possible solution would be to train all messages marked as ham in the checking function, then call untrain ham and then train spam for messages when they are marked as spam.
The same problem arises when messages arrive that are marked as spam but are ham. Should they automatically be trained as spam, then untrained as spam and re-trained as ham? It could also be that SpamBayes is only intended to be trained when it makes a mistake, rather than when it's right. Unsure results from the SpamBayes engine, specifically, could be further facilitated through the plugin.
Considering a potential fix, Evolution should ideally support the unsure option and make it possible to mark the message as spam or ham – to activate both buttons/menu-options – while still delivering the message in the normal fashion.
Conflicting with existing spam filtering plugins, Evolution supports only one junk filtering plugin at a time. Therefore, users need to disable the built-in SpamAssassin plugin before the SpamBayes one will work. Overall, the Evolution SpamBayes Plugin is a valuable and productive addition to Evolution which deserves attention for anyone looking to improve their workflow.
This plugin is easy to implement using two python support scripts. One sets up a DBus Server that holds the SpamBayes database. The other queries the client and trains or untrains on messages received on stdin. The Evolution plugin itself is simple and implemented as a C EPlugin for Evolution. However, it uses some junk-mail hooks in the evolution EPlugin architecture, which is not yet published on the web for unknown reasons.
One downside to the plugin is that messages that are not marked as spam have no way of being marked as ham. One possible solution would be to train all messages marked as ham in the checking function, then call untrain ham and then train spam for messages when they are marked as spam.
The same problem arises when messages arrive that are marked as spam but are ham. Should they automatically be trained as spam, then untrained as spam and re-trained as ham? It could also be that SpamBayes is only intended to be trained when it makes a mistake, rather than when it's right. Unsure results from the SpamBayes engine, specifically, could be further facilitated through the plugin.
Considering a potential fix, Evolution should ideally support the unsure option and make it possible to mark the message as spam or ham – to activate both buttons/menu-options – while still delivering the message in the normal fashion.
Conflicting with existing spam filtering plugins, Evolution supports only one junk filtering plugin at a time. Therefore, users need to disable the built-in SpamAssassin plugin before the SpamBayes one will work. Overall, the Evolution SpamBayes Plugin is a valuable and productive addition to Evolution which deserves attention for anyone looking to improve their workflow.
What's New
Version 0.1.7: N/A