(Dynamic) threading of email conversations. Given gigabytes of email messages, how do you efficiently group them into threads and efficiently update them when new messages arrive? This presents my solution (building on jwz's classic algorithm).
If MIMEs could talk: Email structures in the wild. Email messages can be structured all different ways, which makes it pretty tricky to do things like convert them to plain text. I wrote a tool and did some statistical analyses on the only useful corpus of emails I could find -- my own.