[randallsquared]

While Google Instant is somewhat annoying to me, and therefore turned off, on my Mac (because of the interaction between it and Safari), I think that the process of throwing any old idea against the wall and seeing if it sticks is a sound one in the usual case. The argument the article is making seems to be that theory is better than experiment; I think this would only be true in the case where experiment has shown over and over to be in line with theory, but UI design is such a young field that I'm strongly skeptical of that claim. Surely there have been enough recent accidental or surprising improvements that experiment with unusual ideas is still worthwhile?