I've written a response to someone else, explaining some issues you mention here.
A little note on UI: we don't need web UI. It's a good addition but unnecessary for the beginning. There are many free-software desktop mail clients. Some are big and complicated, but some are very simple and very easy to use, just like Gmail is. So UI is not a critical issue right now, we just need to be able to easily configure a mail client, e.g. Evolution, to work with the server. And it's great to hear people like the idea and want to help! With hard work and cooperation, everything is possible! - Anatoly Krasner On ד', 2013-02-27 at 18:39 -0500, Bennett Todd wrote: > The operational cost is non-zero. Besides hardware, which must include > backups, and enough physical diversity to offer availability, an email > server is an attractive nuisance; spammers and other criminals > constantly attempt sabotage and burglary, and it takes ongoing > manpower to attempt to hold them temporarily at bay. > > And unless you put hard caps on message sizes, people will use their > mailboxes as backup drives, or just email their vacation movies to > family, and you'll be buying drives, and hence replacing them, often. > > I love the idea, I'm fond of running mailservers myself. But I've gone > Google. > > As for software, I won't pitch my favorite components to this wide > list, but I know how to find all the pieces I'd need except the > webmail front-end for the utterly non-technical. > > If you limited the scope to IMAP and SMTP, both SSL authenticated, it > wouldn't be too hard to spec out. > > Host on AWS EC3 or the like, then find an affordable solution to spam, > and you can sell to anyone who doesn't expect their email to be > private from governments. > > Anybody know of a well-engineered and maintained SSL library? >