I like the idea very much. It’s the execution I have doubts about. I mean, the human brain is hands down the best computer ever to see the light of day. The one thing that mechanical brains have on us is speed. But in terms of managing disparate, highly variable (and varied) inputs, people win every time. A human driver can get out of most problems that he gets into without resorting to calling police or actually damaging the car, himself, or his passengers. But historially robots are not great at that particular kind of problem solving. For it to work propoerly, a robot needs to exist in a very specific environment: a clean, clear hallway with an evenly tiled floor, as opposed to a typical city street. I can see a robot car following a highly distinctive pattern of starts and stops, for example, a tram that goes in straight (or mildly curving) lines down a number of streets in sequence and never does anything else, stopping at predetermined points along the way for a set time to pick up and drop off passengers. But such a system requires that the track never break, the passengers never dawdle or crowd each other, and the position of the stops never change. But the world, outside of very highly selctive environments (say, the monorail at Disneyworld or a internal track at an airport passenger terminal), does not generally work that way.
So, yes, I like the idea. I have questions about the feasability and expense. And the necessity, for that matter. Why not just reinstate the public transportation systems many cities in the U.S. abandoned in the 1930s and 40s? It’s the same basic idea.
On to library stuff: the latest edition of the NYAM Grey Literature Report (vol. 4, no. 8, July 2006) is now available here.