The first comment about this - on the blog Sound Politics, where it was posted - said, "Who the hell cares. Why is this being posted on a politics forum?"
But it's all political, in its way. And something basic like this should be: The duration of yellow lights at traffic intersections in Seattle, where drivers are watched by auto-cameras for running red lights.
Poster Carter Mackley measured the amount of time, at three intersections where red-light cameras are installed, that the lights stayed yellow. He came up with 3.5 seconds each, and concluded, "The 3.5 seconds of yellow is too short in my view for 35 MPH streets, and it is shorter than the 4 seconds recommended by others. If borrowing a second or so from the green light would screw up the city's carefully optimized system, I would support moving .5 seconds from the all red period to the yellow light, at least on an experimental basis to gather data. That would give drivers more time to make the stop-or-go decision, and still leave a half second of all red to allow the intersection to clear." He is calling on others to help with additional time measurement, to build a base of information.
One of the aspects of the red-light cameras - which have a clearly useful goal- is that if they're not structured just right, they can have unexpected side effects. People jamming on their brakes at the sight of a yellow, for instance, for fear of being nailed by the camera on what they would otherwise consider to be a safe crossing.
Or maybe that's not a problem. But the discussion of it is clearly useful. The sort of thing political debate is supposed to be for . . .