I retrieve the email and open it in an editor. I don't render the html ... if someone sends me an email that's html only then I queue it for another day. Then

I retrieve the email and open it in an editor. I don't render the html ... if someone sends me an email that's html only then I queue it for another day. Then I batch process the html-only emails, using lynx locally to display the text. At no point do any of my processes download anything except the email itself, and it does that via a raw SMTP connection. People call me paranoid ... they're probably right ... but I just do all this as a matter of course, and it barely affects my flow.