It's amazing that there hasn't been more innovation in email. Part of the problem is that there isn't an easy to code, modular platform to innovate on. Here are some of the projects I've found with food for thought.
- POPFile - Automatic Email Classification. It's interesting that this is the most active project current on Sourceforge. Shows the amount of interest in improved email. It classifies mail and has to suck it in at the POP3 level. That's not a very rich API to say the least. Imagine that there was actually a rich platform for doing this kind of work. You could unleash a lot of developer talent. I had a chance to try it and it is a great idea, but the integration is really pretty bad given the current state of technology. You set up a POP3 intercepter on the local machine and then have to change every single POP3 email account entry to use it. Symantec has to do the same thing and they don't work together. I love the program BTW and have tied all my POP3 accounts to it using Outlook POP3 support. I've been using for a week and it is fantastic. This Bayesian stuff works really well.
- Squirrel Mail. This is an open source POP3 mail reader. It is amazing the number of add-ins that this has spawned. I've looked at the code and it is nice and modular with the open source ethos. Written for server only though, it's got all the disadvantages of a web/HTML only interface. There needs to be a client harness to build things like this.
- Pine. An old project from down the street in Seattle. I've used it and its more like a traditional Unix closed product.
- Chandler. This is Mitch's newest project. It is going to have a very nice underlying data syste it sounds like and has the beginnings of a platform like feel.