Tue 11 Apr 2006
Today I found out that the City of Berkeley does not change the timing of traffic lights with daylight savings time. How? Well, I'll tell ya:
When I moved to a new house last June, I started taking a new bike route to campus (of course), straight down Telegraph Ave. Stopped at one light at a major intersection, I began to challenge myself to get to the light at the next block before the little 'walk' signal started counting down to 0 for a yellow light. I almost always got there.
But then, sometime last Fall, I began to notice that it was much, much harder for me to do. In fact, I was barely getting to that next light before it changed. From a dead stop at the previous light, I'd have to really push it, just so I wouldn't have to stop at the next light too. (Unlike most cyclists in Berkeley, I follow the traffic laws.) Then today, I happened to be riding and noticed that I had no trouble beating the light – in fact it was still on the 'walk' sign as I passed through.
And all of a sudden it dawned on me – daylight savings time. Last week it switched, and today was the first time that I noticed the difference. Apparently Berkeley doesn't change the timing of their lights to accommodate the 1 hour shift. That seems awfully silly and unsophisticated to me. Not to mention probably detrimental for rush-hour traffic. But it made for an interesting bike ride today…
That's pretty interesting, but I'm curious as to just how you expect them to change the timing? It's easy to forget in this internet age that most traffic lights are not networked, but instead planned. Unless the entire system has recently been redone, the timing for each light probably sits there in the base. In order for the timing to be changed a crew would need to go out and do each one manually!
Interesting. I admit I know nothing about how traffic lights work. But it seems like they need not be networked to make this possible. All they'd have to do is know the time and date (which I'd wager they already do), since daylight savings is predictable each year…
I think a lot of cities are beginning to network their traffic lights. There was a Sun news item a few months ago that announced Baltimore's plans spend a whole lot of dough to install a new networked system. Right now some of them still run on clockwork, and the resulting traffic is maddening. (Also, Baltimor's traffic czar proudly proclaims that he does not use theory in his "planning" of the traffic system, and it shows.)
"All they’d have to do is know the time and date (which I’d wager they already do)"
I'll take that bet big time. Its an interesting assumption and I think it actually tells us a lot about our own culture and how much we've taken technological change for granted. Certainly Dave G is right that cities are starting to network their lights. But infrastructure stuff changes WAY slower than digital technology. I'm no expert on traffic lights, but I have some experience programming date and time info into microchips and I'd say the odds of a system that hasn't been redone in the past 5 years knowing the date are pretty low, and approach zero as you get back into the 80's. And of course many of these systems don't even have microchips at all, but are all mechanical.
A great example of this sort of infrastructure lock in is the NY subway system, which for the most part only knows where trains our based on mechanical switches. That means they know a train is in about a 1/4 mile region and that's it. They are slowly upgrading this stuff of course, but it's taken them years to do the first line and it's still not done…
Interesting. I guess I assumed that some mechanical mechanism for reliably tracking time and date has been around for a long time. But I stand corrected!
The NYC subway is an interesting example. I remember hearing about their antiquated (but apparently durable) mechanical system last year when they had a fire in a control room that wiped out a good chunk of the system for a while. At first they thought it would take months and months for them to fix it because the mechanical switches in the room wern't manufactured anymore, and hundreds were destroyed…
actually, I would assume that the timing is a function of available light (light sensors being easy compared to accurate absolute times) and that a by-product of the time change is that the available light is different when you're at the light. Anyway…