Related to Alan's post, this was in the NY Times yesterday

<http://www.nytimes.com/2009/03/08/nyregion/thecity/08bike.html?
pagewanted=1&_r=1>

Regarding observing rider behaviour, some of what Alan is suggesting is
very much down to interpretation: what one person might regard as good
riding technique others would be appalled at, so care would be need in
listing such behaviour.

Likewise with what was referred to as red light jumping.
Marilyn Johnston from Monash has been studying red light breaches and
they come down to a number of variations.

What she refers to as red light jumping is, in fact, what we were all
taught to do years ago, move off just before the green, when everyone
has a red, because that gets the rider up and rolling before the cars
start off, and the rider can usually be straight and up to speed and
much of the way across the intersection, and thus out of the cars' way,
before the cars get the green.
I will admit that that was in a time when the only red turn arrow in
Melbourne was at Kew Junction, there were no green turn assist arrows
anywhere, and it was not a normal training ride unless you had something
thrown at you from a passing car.

There are variations on actual riding through red lights, such as
blasting through without stopping, and stop and look first. The latter
works when traffic signals do not respond to bikes, which unfortunately,
is still typical in Melbourne, at least in my experience.

So by all means collect data, but don't be too hard and fast, and
consider in whose hands the data may end up. A rabidly anti-cycling
politician or shock jock doesn't need any help.

Graham B