Reminiscent of the mooo.com screwup in the US, where Homeland Security's ICE division "accidentally" seized 84,000 sites and plastered them over with a warning graphic about how they'd been seized by the US government for child porn, the Danish police similarly "accidentally" had 8,000 legitimate sites declared as child porn sites that needed to be blocked. Among the sites listed? Google and Facebook. Visitors to those sites, from ISP Siminn were greeted with the following message (translated, of course):

The National High Tech Crime Center of the Danish National Police [NITEC], who assist in investigations into crime on the internet, has informed Siminn Denmark A/S, that the internet page which your browser has tried to get in contact with may contain material which could be regarded as child pornography...

Upon the request of The National High Tech Crime Center of the Danish National Police, Siminn Denmark A/S has blocked the access to the internet page.

And people wonder why so many people around the world were so concerned about the threat of something like SOPA -- which would make DNS blocking at the ISP level a lot more common.

So how did this "accident" happen?

According to NITEC chief Johnny Lundberg, it began when an employee at the police center decided to move from his own computer to that of a colleague.

“He sat down and was about to make an investigation, and in doing so he placed a list of legitimate sites in the wrong folder,” Lundberg explained. “Before becoming aware of the error, two ISPs retrieved the list of sites.”

It would seem that there's a problem in this process. The fact that just one employee can change the list seems wide open to abuse. And the fact that the list seems somewhat automated beyond that is even more problematic. You know what would solve this problem? A little thing called due process. What a concept.

If you believe in gender stereotypes, then you probably think that men are better drivers than women. However, auto insurance companies are inclined to believe that women are actually safer drivers. It's a hotly debated topic, but it's safe to say that there are lots of bad drivers -- both men and women -- on the road. That's why we need robot cars. Here are a few links to some driving-related studies.

• A New York City traffic study found that male drivers were responsible for 80% of car accidents in which pedestrians were severely injured or killed. Blame it on testosterone, which increases aggression and risk-taking. [url]
• A University of Michigan analysis of 6.5 million car accidents in the U.S. between 1998-2007 found that female drivers were responsible for 68.1% of all crashes. Apparently, women have trouble navigating intersections, as they were most often hit on the driver's or passenger side while trying to turn left or right. [url]
• A more recent UK study showed that women are actually better at parking than men, dispelling the myth that men have better spatial awareness than women. According to the professional driving instructor who created the study, men learn and perform better during driving lessons, but it's possible that women retain what they learn better than men. [url]
• Let the car park itself. Many automakers, including Ford, Toyota, and BMW, are offering self-parking systems in their vehicles now. While they can guide cars into parking spaces with minimal driver involvement, they aren't always reliable. [url]
• To discover more interesting car-related content, check out what's driving around StumbleUpon. [url]  

By the way, StumbleUpon can recommend some good Techdirt articles, too.

Well, well. Slashdot points us to this bizarre and slightly scary story about how everyone's favorite TV show, MythBusters, had an experiment that went really, really wrong yesterday. Apparently, it fired a home-made cannon at the Alameda County Sheriff's Department bomb disposal range. The idea was to shoot the cannonball into huge water containers.

But they missed.

Instead, the cannonball went hurtling through the suburban northern California town of Dublin, at 4:15, just as kids were getting home from school. According to the SF Chronicle report on this:

The cantaloupe-sized cannonball missed the water, tore through a cinder-block wall, skipped off a hillside and flew some 700 yards east, right into the Tassajara Creek neighborhood, where children were returning home from school at 4:15 p.m., authorities said.

There, the 6-inch projectile bounced in front of a home on quiet Cassata Place, ripped through the front door, raced up the stairs and blasted through a bedroom, where a man, woman and child slept through it all - only awakening because of plaster dust.

The ball wasn't done bouncing.

It exited the house, leaving a perfectly round hole in the stucco, crossed six-lane Tassajara Road, took out several tiles from the roof of a home on Bellevue Circle and finally slammed into the Gill family's beige Toyota Sienna minivan in a driveway on Springvale Drive.

Wow. Amazingly (and thankfully) no one got hurt in all of this. CBS has some astounding video of the carnage, including showing how the cannoball bounced around that house on Cassata Place putting holes in a bunch of places: But, perhaps equally interesting is that the Mythbusters themselves, appear to have done a little post accident tidying up. After hearing about the whole mess, Jon Laslow went and checked out the various Mythbusters' Twitter feeds and noticed that a bunch of photos of Kari, Grant and Tory posing next to cannons had been deleted. Oops. But, of course, you can't delete anything online... So, the photos & tweets have been preserved:
"Heavy Artillery"


"Canon Envy"


Tory and his Canon Like many folks around here, I'm a big fan of the show, but this seems like quite a mishap.

Here's an interesting story of a self-published Amazon ebook author, James Crawford, who discovered that Amazon "accidentally" gave away his ebook for free over 6,000 times, and then refused to pay him any royalties. The "mistake" was really a bad algorithm. Apparently when you sign up for Amazon's platform, the company makes you promise to give them the best price, and if they find you offering the same thing cheaper elsewhere, they have the right to automatically lower your price. But, here, Amazon's algorithm mistook a "sample three chapters" that was being given away free on the Nook platform, as if it were the whole book. So it automatically dropped Crawfod's ebook price from $5.99 down to $0. It took weeks to sort out, and by the time that happened, 6,116 people had downloaded the free version.

Amazon then refused to pay any royalties, saying that under its contract, it's absolved from having to pay royalties for such mistakes. While that may be true, I do wonder if there's actually a copyright claim there. Here's another situation where the murky boundary between sales and licenses gets strange. After all, Crawford appears to be technically licensing his work for Amazon to distribute with a specific condition: that it be priced at $5.99 with certain royalties going back to Crawford. Offering it free seems like a clearcut case of an unauthorized distribution, which could be open to a copyright claim.

That said, I agree with the point that many have made to Crawford, that he's probably much better off just using this experience for the publicity, and building on that, rather than suing. Suing would be expensive and uncertain. And, it actually could harm sales, as people often don't like the litigious. Readers might be more willing to support Crawford directly after just hearing his story -- and seeing that he was a good sport about it, who could recognize that perhaps this mistake could lead more readers to his work, and those readers may bring in other (paying) readers in the future. Still, it's yet another case where the difference between physical sales and digital "licensing" becomes abundantly clear.

weneedhelp points us to yet another report suggesting that red light cameras quite frequently appear to increase the number of car accidents, even as they're pitched to the public as a way to decrease accidents. In this case, police data from Philly showed that aggregate accidents apparently increased 12% at the locations where cameras have been installed for at least a year. What's odd, however, is that the local Parking Authority, who runs the red light camera program, insists that the police data is inaccurate, and that its own data shows a decrease in accidents. Of course, it's not clear why or how the Parking Authority would have crash data. The police seem much more likely to have such data, so it seems like that's a lot more trustworthy. As we've noted in the past, if municipalities really wanted to decrease accidents, they'd do two things: (1) increase the length of yellow lights and (2) add a delay after lights in one direction turn red before the lights in the other direction turn green (and, yes, many places do this already -- but lots of them do not, including most lights near where I live).

Of course, it's not difficult to see that local governments have no desire to actually stop violations at all, because things like red light cameras have never been about safety on the roads, and have always been about money:

Motorists have paid $45.3 million in fines since 2005, of which $21.1 million has gone to PennDot.

The rest of the money - $24.2 million - has gone to pay the expenses for operating the program, mainly to American Traffic Solutions Inc., the Scottsdale, Ariz., company that installs and maintains the cameras, and to the Parking Authority.

PennDot distributes its revenue from the camera program for transportation projects: half to Philadelphia and half to the rest of the state.

So far, Philadelphia has received $8.4 million; another $8.4 million has gone to 116 other municipalities in the state. Philadelphia is to get $1.5 million more, and a like amount will go to the rest of the state.

Apparently, the amount being made today absolutely dwarfs what was made in tickets before the cameras were installed. And, so, rather than actually making the roads safer, they're being made less safe in order to beef up government revenue.

There's a bunch of talk today over the news that one of Google's self-driving autonomous vehicles apparently got into a minor fender bender. Google was quick to point out that it was actually under human control at the time, so really there's not much of a story here. However, since it's leading to a variety of discussions about how "scary" autonomous vehicles are, why don't we just get an important point out of the way: there's no way that autonomous vehicles will have a perfect track record and never, ever get into an accident. They will crash. It's just a matter of time. The real question is not whether or not they will crash, but whether or not the likelihood of getting into an accident (or the likelihood of the seriousness of any such accident) is significantly higher or lower than with a human at the controls. I'm certainly not confident in the state of the art today to be safer, but I find it likely that it won't be long until such vehicles have a much higher probability of getting you to your destination safely than a human-driven vehicle.

Tom sent over a report saying that redlight cameras aren't just about making money for local police (and the companies who sell redlight cameras), but that one study in Barcelona found that the city saved a ton of money due to fewer accidents. Now, what I find interesting about this is that it seems to contradict nearly every other study we've seen on this subject. Study after study after study after study after study has shown no decrease in the number of accidents... and in many cases an increased number of accidents. So is there something that Barcelona did differently with its cameras that resulted in a decrease in accidents?

After digging in, it looks like some of this may just be bad reporting by Scientific American's Cynthia Graber. First, she focuses the article on red light cameras, but the actual study she discusses was for speed cameras. I'm surprised that Scientific American would mix up the two, since they're clearly different. Second, as some people pointed out in the comments to Graber's story, the study happened over a two year period during which unemployment in Barcelona shot up drastically, potentially leading to fewer cars on the road, especially during rush hour periods when accidents are most likely. In other words, while this might be some evidence in support of speed cameras, at best, it shows a correlation, and the causal factor may be something entirely different, such as the amount of cars on the road.

We've been suspicious of whether or not "driving while yakking" laws actually do any good. There are already laws against reckless driving, and picking out specific driving distractions doesn't seem likely to change things, since people just switch to other distractions. A study back in 2006 found that driving while yakking laws don't make the roads any safer, and a brand new study has apparently surprised researchers in showing no impact whatsoever on crash data even as studies show that fewer people are holding phones to their ears while driving (thanks Chirag). Now, there could be plenty of reasons for this -- such as that people are just switching to ear pieces which can be just as dangerous. Or it could be that common claims about driving while yakking leading to more accidents are wrong. Or it could be more complex, with other variables having an impact, but which is hidden in the data. Either way, it certainly seems worth investigating more seriously. If the goal is better road safety, then we should make sure that the laws actually lead to that result. If they don't, then it's important to understand why not.

Dark Helmet was the first of a few to send in a Chicago Tribune article looking at redlight camera accident rates and seeing (as many other reports have shown before) that in many cases the number of accidents went up. The report looked at fourteen intersections. Seven showed increases in accidents, two stayed the same, and five dropped. Even more interesting: at one of the intersections the number of t-bone "broadside" collisions (the ones that defenders of the cameras insist decrease) went up significantly. That was just one intersection, though. Others saw the more traditional decrease in broadsides, but significant increase in rear-end accidents as drivers slam on their brakes. The other interesting finding from the data: contrary to the claims of camera defenders, over time the rate of accidents did not appear to decrease "as drivers learned about the cameras."

Once again, we're left wondering why people still claim the cameras increase safety, when there appears to be no evidence to support that at all. There is a clear and proven way to increase safety though: (1) increase the length of the yellow and (2) increase the delay (or, for places like California which have no delay, put in a delay) when lights in all direction are red, before switching the new direction to green. Any municipality that puts in redlight cameras without doing those two things above, and then claims its about "safety" is lying. Bizarrely, though, the supporters in the Chicago area are still defending their system:

Even if not reflected in accident statistics, Belwood Police Chief Robert Collins Jr. said he sees drivers being more cautious as they approach stoplights. "Driver behavior has definitely changed," Collins said...

Changed for the worse, apparently. Are we to believe the police chief's anecdotal insistence or what the stats actually say?

Roger Pawlowski, a division chief at the Oak Lawn Police Department, said the benefits of red-light cameras can't always be extrapolated from crash statistics.

Ah, then what are we to extrapolate the benefits of redlight cameras from? Checks cashed by the city?

We were just talking about how current child porn laws that make you a criminal based on incidental possession alone can be quite problematic. Reader Jesse highlights an example of this. If you happen to have been in Toronto on Monday, and received a copy of Metro Toronto, a popular commuter paper, buried a few pages in was a "featured picture" of some kids celebrating at an annual parade by jumping around in a hot tub. The only problem? The male in the photo appears to be, well... having a wardrobe malfunction. Not unexpectedly, a bunch of sites were having some fun with this... until some realized that these were high school students. At that point, even Gawker -- who will post almost anything -- removed the photo realizing that under current laws, it was likely child pornography. Jesse points out that this would appear to make a bunch of people at the Metro, all the recipients of the paper on Monday, many people who visited blogs like Torontoist and Gawker while they had those photos displayed... potentially at risk for possessing child porn (and in the case of Metro employees and these blogs, for distributing it as well) -- making them all potential felons who could be required to sign up to be on sex offender lists for the rest of their lives. Isn't something wrong with the law when that happens?