I took a big step today… I’m in the process of uploading about 500MB of mail to gmail. I will use it as my primary email client/storage repository.

I’m changing two things at the same time - my email client (mutt) and the server. Up to this point, all my mail has been stored exclusively on my own servers. My reluctance to put it elsewhere has been a matter of control - loss of privacy, loss of flexibility. At the moment, my email address will forward both to Google and store a copy locally. Longer term, I will have a short-term backup queue combined with a complete synchronization of Google’s store.

I still remain nervous about my potential loss of privacy.

Update (11:57 PM): Well, after “uploading” my ~500 MB worth of email, I ended up with only 353 on gmail. gmail also didn’t do so well with mail entries w/o domain names (or some other form of a mangled From header.) I’ve clearly got more work to do before I can switch over. I’ll likely need to write a tool to do the upload myself (I had used mutt to “save” the mail into the gmail folders) so I can verify that nothing is missed.

I’m going to take a step back and assess what I’m really looking for. I want ubiquitous access via IMAP. I can do that with my current mail server. I also want a decent webmail client which is based on a “search” model versus an “organize” model - that will take some additional research (I don’t think Squirrelmail cuts it.) I also need to get a new Crackberry which supports additional IMAP servers.

If I had all of that with my own server, the long-term questions remain:

  1. Am I willing to give up my “privacy” by storing my mail elsewhere? (Presuming I have perfect backups.)
  2. Can I provide the compute resources necessary to match my workflow (search)?
  3. Is my workflow more unique such that it will require further customization?

I think the long-term answer is still gmail (or some other hosted service), even though it is hard for me to give up that control.