I've hooked a Python Bayesian classifier into procmail and deliver things into folders. In essence it follows the GTD input process flow and delivers emails into project, answer required (AR), archive and delete folders. The AR emails I scan and reply to every day, refiling those that have been misclassified. My weekly GTD pass checks the other folders, allowing me to reclassify when they're wrong, and deal with appropriately otherwise.
For archiving I just put it all in either a project specific repository or a single folder called "other". I can then search for keywords, but I usually pop out to a command line and grep for things if I need them.
I'm sure there's more available in existing systems, but the Bayesian classifiers also work as spam filters. I wrote them years ago and they are a great tool to have to hand.
If you have the inclination, I'm sure many people (myself included) would be interested to see your code, even if it's raw and unsupported. Cool solution.
Hmm. I'll have to think about that. There's a lot I really wouldn't release, but I'll see what I can do. Too busy just now, but I'll put it on my list.
Thanks for the interest. It never really occured to me that someone would be interested.
For archiving I just put it all in either a project specific repository or a single folder called "other". I can then search for keywords, but I usually pop out to a command line and grep for things if I need them.
I'm sure there's more available in existing systems, but the Bayesian classifiers also work as spam filters. I wrote them years ago and they are a great tool to have to hand.