John Goerzen blogs about his experience of using vim for about 8 months. Having been an emacs user a few years back, I understand both editors on a basic level, and can relate to some of the things he says. I have to disagree with one particular bit, though:
It starts faster. I'm not sure if that really was true even when I switched, but it certainly isn't true on any of my machines today. Both Vim and Emacs have had major version upgrades (v7 and v22, respectively) since I started using Vim. People seem to say that Emacs 22 feels faster, though I don't know if that's true. The startup times of the two, if they're different, are imperceptible.
Sorry, but that just isn't true. Using my very non-scientific method of measuring startup times ("click on the icon, watch the clock, and start counting until the editor is ready to use"), I can tell that gvim on my hardware takes less than a second to start, whereas emacs requires approximately 5 seconds to do the same. For reference, this is on my laptop, a 1.3Ghz PowerPC G4.
In any case, I'm not a fan of either editor. There are things that vi does well, and there are things that emacs does well. On emacs' end, I like M-x gdb, I slightly prefer its M-x compile over vim's :make, and I just love psgml. I've also occasionally used gnus for news reading (never for mail, though). On vim's end, I find regex search & replace easier (as you mention in your blog post), and I prefer its syntax highlighting (I've seen emacs get confused because the first line in the output window started in the middle of a string. Such things will never happen to vim: if it gets confused once on a given file, it will always be confused on the same file. While annoying, at least you know it's there).
All in all, I don't think it matters much. I've seen people who use an editor as an operating system, both in the vim and in the emacs camp, by having a ~/.vimrc or ~/.emacs that's larger than my collection of pictures[1]. I'm not one of them; to me, an editor that doesn't properly work in its defaults shouldn't exist.
"Nick Dilmore has published the second edition in his Snarky Halloween History series, featured on Slashdot last year. This time around, he concentrates on movie monsters: vampires, werewolves and zombies. From the article: '[D]id you know the movie monsters we've all to come to know and love (in a platonic way, of course) have colorful histories stretching back to the earliest civilizations? What, you didn't think some Hollywood hack actually had enough imagination to invent vampires, werewolves, and zombies, did you? Silly, silly non-monster-trivia knowing person.'"
"Project WAND, for Wireless Adaptive Network Development, will exploit commercial radio components, rather than custom ones, and use a variety of software techniques and algorithms, many of them only just now emerging in mature form. These $500 walkie-talkie-size radios will form large-scale, peer-to-peer ad hoc nets, which can shift frequencies, sidestep interference, and handle a range of events that today completely disrupt wireless communications ... [right now] 'The average soldier on the ground doesn't have a radio,' says Jason Redi, principle scientist for BBN's network technologies group, and the man overseeing the software work. Radios are reserved for platoon and company commanders, in part because of their cost: typically $15,000 to $20,000 each, with vehicle-mounted radios reaching $80,000."