Building A Better Server

We've recently been pitched by a number of teams with good next generation email clients and while at the TED conference, I saw Mitch Kapor discuss the open source PIM Chandler project. It's great to see such development going on in a space with arguably huge potential (do you know anyone who doesn't use email?) but difficult revenue prospects (do you know anyone who has paid for their email?).

It has made me wonder why no one is trying to fix email servers. A lot of the problems I have with email could arguably be better solved server-side, including filtering, offline mail sync, bounce management, address book/contacts, etc. And let's face it, Exchange is about as bad a product as Microsoft makes. I can't imagine that selling servers to enterprises is any harder than selling email clients to users, given that both are widely and freely available. The reality is selling either is very hard, I'm just surprised that no one yet seems to be focusing on the server.