Eye tracking is evil

Someone on LinkedIn says:
"I was at a public discussion with Jared Spool last week (and about 100 participants in SF at the IxDA discussion). the topic was... when usability is evil. The consensus on eye tracking... everyone hates it, no one seems to trust the results. In the end... and from what i understand as results from the consensus: eye tracking is evil. There were persons who championed eye tracking... (I think there was one), but... just some random data for your study from a room full of UX's. anyone who was there is welcome to chime in here."

You can't say usability is evil, but I would certainly agree that eye-tracking is at worst evil and at best pants. In my opinion, any user testing method that requires such long-winded and dreary analysis, in real time, afterwards, by an expert, is a bummer. If you've got a roomful of users, about the dullest thing you can do is point a camera at them. Go in! Watch! Take notes! Talk to them! *Then* you'll find out stuff. When do you ever learn about people just by watching their eyes? You'd only do it if there was no possibility of talking to them.

Perhaps some of the software available is good stuff if you use it a lot, but everything I've seen seemed to have quite a steep learning curve and be prone to disastrous errors - recording three days of video with no audio, to cite an example from one of my team, who traipsed all the way to the US to capture a lot of data that turned out to be unusable. Instead of faffing about with tracking nonsense, invest some time before your test sessions in writing really good questions and tasks, and talking to the client to make sure you're going to find out everything that they're going to want to know afterwards.