I recently came across Joel Spolsky’s description of a usability bug in the elevators at 7WTC. The elevator tried to perform a form of “passenger clustering” to reduce the number of stops made by each elevator, but it failed because the new system doesn’t match up with the mental models of most elevator users.
The essay reminded me of why I don’t try to do lots of user-interface design. Although it’s a discipline that involves lots of code, there is much more to it than writing decent code. I can tell you if a UI is good, but I will sometimes struggle to tell you why, and what should be changed to make it better. (I find the whole task of assessing user interfaces to be uncomfortably close to Justice Potter Stewart’s definition of pornography — I may not know how to build a good UI, but I can tell when I use one.)
Joel’s article also reminded me of a much older salon.com article about personal rapid transit (PRT) in the wake of seeing The Incredibles in 2004. The idea behind the new-fangled PRT system is that you enter a station, punch in a destination, and get in a car that whisks you to your destination. It has the potential to suffer from the same sort of UI problem as the elevators, though the company’s description of the ticketing process makes it seem less likely. A PRT system is more like a taxi, which makes it less likely that users would trip over a pre-conceived notion of how things are “supposed to work.”
Interestingly, I might be able to ride the Taxi 2000 system in the distant future. I work in the Hacienda Business Park in (un-)Pleasanton, California, which is apparently considering a PRT system to link up with the BART station. On its face, Pleasanton seems like an idea trial zone for trying a PRT feeder system instead of a bus feeder system. Pleasanton is a rich, car-dependent suburb; as far as I can tell, the bus exists to get lower-paid service workers to their jobs. The shuttle bus that runs from BART to my office building is available only for limited hours, and the commutes end to early for a typical high-tech engineer whose day starts late and ends late. Mostly, though, I’m interested to see if the PRT concept works. Proponents of these systems claim very high capacity, but most rail systems that run tight headways don’t need to throw track switches anywhere near as frequently as the proposed PRTs.