Take my advice - throw away every email older than 2 years. Not only will you never need them, but they're what the black hats will hang you by should they get the urge. You will feel a huge sense of relief - a weight lifted from your soul as you relinquish the crushing burden of carting about all those dead emails.
(Unless, of course, you work in an industry where longer retention times are mandated.)
In the UK, business documents must be retained for six years.
Im not a lawyer so there may be subtle details I don't understand, but broadly in England and Wales, you can be sued within either 6 years or 12 years from the date of a breach of contract depending on type of contract, in Scotland 15 years. I don't know about the US or rest of EU, but think carefully about deleting all your old emails. You might have something in there which would be very useful to defend yourself if this ever happened. You would regret this if you had built the next Facebook and some old acquaintance decided to stake a claim based on some gmail conversation from 10 years ago. Better to archive to a redundant backup system I would say.
The MailPile page (https://www.mailpile.is/) mentions GMVault (http://gmvault.org/) as their method for backing up Gmail. I haven't heard of GMVault (or MailPile) until just now. Both look promising. I'm glad there is a push in this direction.
I've been using GMVault for the past month and I have really enjoyed it. I wrote a cron job to periodically pull down my latest emails and back them up (encrypted).
Does GMVault store your mail on your machine in a text-readable (I include XML in this) format? Because if it's a binary blob, it's not as useful as it could be.
All emails are stored individually as EML files which is just text. It can also export all your emails into the standard maildir or mbox format so you can easily import them elsewhere.