Interesting experiment, but not a win.<p>For a new UI to be a success, it needs to be faster for the user to USE or faster for the user to LEARN. The former is more important for tools with repeat users (like a mail app), the latter is more important for sites with transitory users. You could argue that it could perform okay on the former point IF you ignore the meaningful population of people who use trackpads or mobile devices... But how can you possible argue that ignoring that population is a good idea?