So, we got one of the $100 "obstructing snow plow" tickets for parking over at twinknj last night. When, of course, there was no snow. (My fault, I knew there was a snow emergency at 4; I had gotten confused and thought it was 4am, when there would actually be snow). Bah.
The trouble is, Somerville has a declared policy of "ticket and/or tow way before there's snow" (including two years ago when they ticketed 3000 cars and towed 150 when it snowed an inch). By being particularly draconian about this, they can, in fact, keep cars out of the way of plows when there *is* snow. And I can't argue with that, because it works.
But it's yet another example of the Way Things Work that bugs me. In order to get reasonable behavior, we have to outlaw not only unreasonable behavior but a whole area around the edge, because otherwise you can't be sure you get the edge cases. If the speed limit is 60 MPH, then the prevailing flow of traffic goes 70[*]. Schools write harassment policies that end up suspending 6-year-olds for sexual harassment when they grab another 6-year-old by the waistband, so they can be sure that they're on firm ground for things that actually resemble harassment as most of us know it. "Traffic calming" solutions, like those raised intersection areas, slow me down to about 15mph, when I have to assume that what they really care about is that people not drive 50 through the neighborhood.
Essentially, they're systems which are skewed towards allowing a lot of false positives, rather than running the risk of allowing false negatives. I used to think that in an ideal world, we could make the rules actually between right and wrong, but now I wonder if it just always has to be a choice between which failure case you care more about.
*: I make the assumption here that the (possibly evolved) desire is to stop people from going 80, rather than really stop them from going 60. If it's the latter, they're just failing utterly.
The trouble is, Somerville has a declared policy of "ticket and/or tow way before there's snow" (including two years ago when they ticketed 3000 cars and towed 150 when it snowed an inch). By being particularly draconian about this, they can, in fact, keep cars out of the way of plows when there *is* snow. And I can't argue with that, because it works.
But it's yet another example of the Way Things Work that bugs me. In order to get reasonable behavior, we have to outlaw not only unreasonable behavior but a whole area around the edge, because otherwise you can't be sure you get the edge cases. If the speed limit is 60 MPH, then the prevailing flow of traffic goes 70[*]. Schools write harassment policies that end up suspending 6-year-olds for sexual harassment when they grab another 6-year-old by the waistband, so they can be sure that they're on firm ground for things that actually resemble harassment as most of us know it. "Traffic calming" solutions, like those raised intersection areas, slow me down to about 15mph, when I have to assume that what they really care about is that people not drive 50 through the neighborhood.
Essentially, they're systems which are skewed towards allowing a lot of false positives, rather than running the risk of allowing false negatives. I used to think that in an ideal world, we could make the rules actually between right and wrong, but now I wonder if it just always has to be a choice between which failure case you care more about.
*: I make the assumption here that the (possibly evolved) desire is to stop people from going 80, rather than really stop them from going 60. If it's the latter, they're just failing utterly.
no subject
Date: 2006-02-15 02:06 pm (UTC)What bugs me the most, though, is lack of clarity. If everyone knows at any given moment whether they're breaking the rules as enforced, any given incentive structure works better. What irks me is that setting the speed limit at 60 mph but only ticketing people for going some unknown amount over that leads to uncertainty about whether you're actually breaking the rules or not.
Now, if the probability of getting caught when breaking the rules were very high, this uncertainty would lead to extra caution, which would be the point of the skewed system. But the probability of getting caught for speeding or parking wrong is also fairly low, so the uncertainty is compounded to where it's easier to just model parking tickets or speeding tickets as completely random events that don't depend on my behavior at all.
(The sexual harassment thing seems like a different issue. I feel like you could write an overly conservative policy that didn't result in complete absurdities like banning hugging or suspending six year olds.)