Slashdot Mirror
DUI Defendant Wins Source Code to Breathalyzer
MyrddinBach writes "CNet's Police Blotter column looks into a Minnesota drunk driving defendant case with a twist. The defendant says he needs the source code to the Intoxilyzer 5000EN to fight the charges in court. Apparently the company has agreed to turn over the code to the defense. 'A judge granted the defendant's request, but Michael Campion, Minnesota's commissioner in charge of public safety, opposed it. Minnesota quickly asked an appeals court to intervene, which it declined to do. Then the state appealed a second time. What became central to the dispute was whether the source code was owned by the state or CMI, the maker of the Intoxilyzer.'"
Actually, no, in many jurisdictions. Often the way it works is you fail the field sobriety test, whatever it is. They take you back to the station where there's a large breathalyzer (which is what the one in question is, the thing weight 30 pounds, it's not a field unit). They test you with that, if you fail, you are charged with DUI.
In most cases you have the right to demand an alternate test and/or to demand that a same is taken so that your defense team can have it independently tested. However as notification of that fact isn't required, people usually aren't told. So quite often the evidence is solely that taken from one of these units.
Also no, they don't tend to set up devices to be lax. Prosecuting drunk drivers is big business, so to speak. There are a lot of groups that want law enforcement that is tough on drunk driving thus it helps DAs get reelected.
The reasons why they don't are several:
1) Most companies don't like having their source revealed, they feel it would allow competitors to make a competing device more cheaply.
2) This isn't something lawyers/judges are so likely to think of. If the RIAA cases tell you anything, it should be that legal professionals usually know little about technology.
3) These devices are not as accurate as they'd like to pretend. For one there is the fact that the way they measure isn't extremely accurate but the bigger issue is one of measuring something indirectly. The only way to really determine the amount of alcohol in blood is to in fact test the blood in question. However that is expensive and due to the invasive nature, courts might rule that either a court order or consent (and not just implied consent) is needed. As such an indirect method is used. This is something akin to measuring the lumen output of a light by measuring it's energy input. In theory, that works fine, so long as you know the kind of light and the efficiency involved. In practise, you information could be wrong and the only way to really know is to, in fact, directly measure the lumen output.
4) If it is open to review, that means that flaws could be found. Unfortunately having the code open doesn't mean you find all the problems right away. If this were the case, why are there any bugs of any kind in any OSS that's been out for a while? Things can lurk unnoticed by everyone for years. For example have a look at the BIND exploit back in 2000. It affected essentially all version of BIND ever released. Well BIND has been around a long time, and the source is open, so how did this happen? It happened because nobody had noticed it before, despite all the people who'd looked. Finally someone stumbled across it. Same deal here. Maybe it gets vetted and everyone says it is good. Then 2 years down the road some guy notices something everyone missed and bring it up. Now you have a shitload of cases to review, which the courts don't want.
We encountered this on campus when we tried to buy campus wide firewalls. Finding a firewall that on paper did what we needed was easy. After all there were firewalls rated to a GIGABIT of traffic, when at the time we only had like 200mbps of total bandwidth to the Internet. Ok no problem right? Wrong. We'd put them in the network (or rather a mirror of the network they got to play with) and they'd fall flat on their face. We are talking lagging out at best, and just crashing at worst. But how could that be? They'd been tested in labs and shown to work! Well, the kind of traffic you generate on a network simulator isn't nearly as tough as the traffic of a bunch of undergraduates screwing around on the Internet. So while these worked fine for many companies, they couldn't take a situation like ours (Cisco blade firewalls for their 6500s worked in the end, if you are wondering).
Testing it great and all but the real proof is when the shit is actually in the field. Doesn't always work like you planned.
For more on this phenomena, see Mythbusters and the number of times they've had something that should work perfectly that didn't.
Do we ban all other impairments, and to what point? For example driving while tired is a severe problem that has a lot of the same effects as driving drunk: Your reaction time is slower, you don't pay attention as well, you have trouble focusing, etc. So do we make it illegal to drive while tired? If so do we take the same hardline attitude and say "There's no compelling reason why anyone should have to drive with any fatigue, if you haven't gotten a minimum of 8 hours of sleep not less than 10 hours before hitting the road, you've got no business driving."
I mean hell, all you have to do is play twitch video games to know the significance of sleep on accuracy and reaction time. Back when I played Team Fortress and then Action Quake 2 competitively online I'd actually make sure to nap before important matches, be well fed, etc. Reason is I showed a measurable increase in performance. When I was really exhausted I didn't do so well.
Significant isn't a weasel word, it's an important term. Your driving WILL be impaired by a number of different things at any given time. It is rare that you are in peek condition to drive. So the question becomes what amount of impairment is acceptable? You can't say "none" or you essentially ban almost all driving all of the time. You have to be reasonable about what amount of what kinds of impairment really make it more dangerous to drive.
1) It's more expensive (and slower). With a blood test they have to send the sample to a qualified lab (could be their own) and have a technician do the work on it. With a breathalyzer you just blow in it, it gives results. Not hard to see which is cheaper.
2) You'd have problems making it mandatory. A blood test is highly invasive, and in fact is against some religions. Thus generally to get a blood test the state requires either the consent of the person giving it, or a court order. The whole "implied consent" probably wouldn't cut it for a blood test. You'd get people who would sue for a violation of civil rights of various kinds and probably win.