MN Supreme Court Backs Reasoned Requests For Breathalyzer Source Code 199
viralMeme writes with news that the Minnesota Supreme Court has upheld the right of drunk-driving defendants to request the source code for the breathalyzer machines used as evidence against them, but only when the defendant provides sufficient arguments to suggest that a review of the code may have an impact on the case. In short: no fishing expeditions. The ruling involves two such requests (PDF), one of which we've been covering for some time. In that case, the defendant, Dale Underdahl simply argued that to challenge the validity of the charges, he had to "go after the testing method itself." The Supreme Court says this was not sufficient. Meanwhile, the other defendant, Timothy Brunner, "submitted a memorandum and nine exhibits to support his request for the source code," which included testimony from a computer science professor about the usefulness of source code in finding voting machine defects, and a report about a similar case in New Jersey where defects were found in the breathalyzer's source code. This was enough for the Supreme Court to acknowledge that an examination of the code could "relate to Brunner's guilt or innocence."
Hm. (Score:3, Interesting)
Fishing expeditions (Score:5, Interesting)
Re:Fishing expeditions (Score:4, Interesting)
So essentially challenging evidence gathering methods is insufficient, but making colorful posters and waving around a PhD is fine?
Yes. Like a slashdot legal opinion is worthless, and someone who has passed the bar has value. Who is questioning the procedure is relevant.
Re:Fishing expeditions (Score:5, Interesting)
I've never found it very difficult to understand the lawyers and judge's case findings. Yes, they use a little Latin, but you can learn the 100 most used words and look up anything more that comes up on the web. And they cite cases, which you can look up too.
Are you talking about contracts?
Re:Hm. (Score:4, Interesting)
That is mentioned as a consequence of the ruling. Now it remains to be seen whether the manufacturer will release the source code. If they won't, presenting a copy of Bruner's exhibits will be a 'get out of jail free' card for drunk driving in Minnesota. Which will mean the state will have to go with a manufacturer that WILL provide the source. Nice.
Re:Hm. (Score:3, Interesting)
Re:Hm. (Score:3, Interesting)
But that's contrary to the prohibitionist agenda that defines "drunk" using arbitrarily low readings from a breathalyzer, so it's unlikely to occur.
Re:That's how science works (Score:3, Interesting)
Re:Fishing expeditions (Score:3, Interesting)
Re:That's how science works (Score:2, Interesting)
it can only measure the absolute count of methyl-group molecules in the sample chamber
Well, it's alcohol molecules, not methyl groups (I presume you're not a chemist), but yes, that's true.
not the amount of alcohol in your blood, or even the sort of things you'd need to make a reasonable estimate given the amount of alcohol in your exhaled breath
The exhaled alcohol comes from the blood vessels in your lungs. The amount of alcohol in your exhaled breath strongly correlates to the alcohol concentration in your blood, with only very small variations with body size & lung capacity.
(subject weight would be a good place to start).
Subject weight is irrelevant to BAC. You might argue BAC isn't a good measure of impairment, but that's a different story.
Some breathalyzer models may have glitches & errors (which is the subject of this article), but the idea of using exhaled alcohol content to calculate BAC has been proven to be very accurate and has been tested to death.
Re:Hm. (Score:2, Interesting)
... I think we need to notify the authorities.
I dunno, maybe we should vote on that.
My father wrote the first Felony DWI laws in MN (Score:4, Interesting)
Thought I would share this, and before a bunch of you start posting BS about the claim of who wrote what, that is not the point. The point is the evolution of MN DWI law and technology.
My father as a prosecutor in Minneapolis in the 60's and 70's started prosecuting drunk drivers for things like felony manslaughter and such. At the time it was just misdemeanor, and often the police would just give someone a ride home. The State legislators and several lobby groups caught wind of it and asked him to write the laws. Those became the first felony laws for DWI in MN, and later where used as a model for other States around the country. Obviously they have been super modified since then, but the fundamental principle that DWI is something serious is still there.
My father went on in private practice as defense attorney in the 80's. Almost all of his acquittals on DWI came down to discrediting the probable cause (i.e. the officer) for the arrest in the first place. Typically the officer's judgment was always front and center (e.g. did he really see him cross the center line on an ice covered road). It got progressively harder as they started adding video cameras and other technology to get someone off on a DWI charge, as the officer's judgment became less important.
I suspect since my father's time, the only thing left to really attack is the validity of the technology itself that measures the crime.
Re:Hm. (Score:2, Interesting)
Well I vote [NO]. What? I meant to type [NO]. What's going on here?
Re:Fishing expeditions (Score:4, Interesting)
I remember a discussion about this on /. awhile ago, and somebody brought up a point I couldn't quite remember. I hopped over to the wikipedia page to look for something about found this:
"Some breath analysis machines assume a hematocrit (cell volume of blood) of 47%. However, hematocrit values range from 42 to 52% in men and from 37 to 47% in women. A person with a lower hematocrit will have a falsely high BAC reading."
and this:
"Breathalyzers assume that the subject being tested has a 2100-to-1 'partition ratio' in converting alcohol measured in the breath to estimates of alcohol in the blood [. . .] However, this assumed 'partition ratio' varies from 1300:1 to 3100:1 or wider among individuals and within a given individual over time."
I'm not sure what, exactly, I was remembering from the previous discussion; these may or may not be it. What I do remember is that it was essentially that, somewhere in this code, there are assumptions made and that the validity of the assumptions is going to directly affect the validity of the code.
Without knowing what, exactly, this machine is measuring and what it is assuming about the individual taking the test, it's impossible to know whether or not there's any reason to believe the test was inaccurate. Since both of these people argued this case to the Minnesota Supreme Court, I hope they both feel they're innocent.
I suppose this guy's lawyer should have made that argument. On the other hand, I don't think it's unreasonable for judges who are going to oversea DUI cases to understand that a breathalyzer is not, by any stretch, conclusive evidence. Use it to haul somebody in, by all means -- then get yourself a blood test. Bill the person being charged for the test for all I care. The breathalyzer itself should not be admissible in court. (I'm ignoring, by the way, the fact that something like having taken cough medicine could also affect the results.)
Re:That's how science works (Score:3, Interesting)
Daubert superseded Frye and does specifically require that the technique be subject to peer review AND PUBLICATION. That means that the super secret proprietary code in the brethalyser disqualifies it (or should disqualify it) as scientific evidence.
As far as 'generally accepted' goes, since this is essentially a medical test, I would have to presume the medical community's opinion would be relevant. As far as I know, when a doctor wants to know a patient's BAC, they do a blood test every time. Evidently the medical community doesn't consider a breath test to be adequate for determining BAC.
Re:That's how science works (Score:3, Interesting)
Of course, the SD-2 can be used to convict you of being a minor in consumption of alcohol, which makes sense because, whereas the DUI law punishes 0.08% or higher and an inaccurate measurement by the SD-2 can make or break the case, an MIC punishes anything greater than 0%, so an inaccurate measurement only matters if it finds alcohol where there is none, which is vastly less probable than inaccurately measuring the amount of alcohol where there is some.
Actually, it can be a problem. Many mouth washes contain alcohol (in that case, the breathalyzer is technically correct, but the assumption that alcohol on your breath means you consumed an alcoholic beverage is faulty). Many breathalyzers cannot distinguish ketones from alcohol. Even is the breathalyzer functions perfectly every time, it's not valid to presume guilt based on the slightest trace of a reading.
Re:These guys are no heroes (Score:1, Interesting)
So you're saying that it should be possible for anyone to be hauled in for drunk driving if the cops don't like the smell of their breath?
I sure hope the police in your area don't have any reason to dislike you...
Re:If it looks like a drunk, it probably is. (Score:3, Interesting)
>
The bottom line is, if someone is getting pulled over for DUI, its because they were already obviously driving smashed. You can see them when you drive - weaving, going slow, forgot to put the headlights on, maybe stopping too soon or too often... and for what? Really, for what do people do this?
I hope you are never the cop who pulls someone over in diabetic keto acidosis... Looks like a drunk, smells like a drunk, acts like a drunk... And if you just dump them in the drunk tank, they will die and you will be on trial.
Re:That's how science works (Score:3, Interesting)
We bar evidence all the time if it's not reliable. That is particularly true when it might be prejudicial, that is when the judge knows it will LOOK damning to the jury but is in actuality either irrelevant or unreliable. A judge may also instruct the jury to disregard testimony if it fails to meet the relevant standards.A judge may even penalize the prosecution if it knowingly brings a case based only on such inadmissible evidence (though it is more common to simply admonish them and throw the case out). Essentially, refusing to admit evidence is a simple and effective way to avoid adding months to what should be a simple trial and avoid confusing and potentially prejudicing the jury with what will clearly be ruled irrelevant. If that leaves nothing to prosecute, so be it! It's better to let 100 guilty people go than to wrongly convict one innocent.
It could be argued that breathalyzer evidence is hearsay. That is, the breathalyzer is acting as a proxy for it's programmers and designers who claim that it's in-built methodology is valid. That is, the officer is really testifying that "Intoxilizer 5000EN told me that CMI told it that the defendant was intoxicated".
The adversarial process is supposed to be the next line of defense when that is not done. "I think I saw him at..." will be torn apart in court. If we are really appropriately observing the standard of proof and placing the burden of proof on the prosecution, all the defense should need to do is ask "can you demonstrate that that particular breathalyser on that particular night was 100% accurate under those particular conditions? Can you prove that no other medical condition, even a transient one could cause those readings?
In the case of a blood test, the prosecution certainly can prove all of those things beyond reasonable doubt. The methodology is well known, widely peer reviewed and generally accepted as accurate by the medical community. If they have properly retained the sampled blood, the defense can even have a second lab re-test the sample. The blood test is NOT hearsay because the lab submits an affidavit containing their findings and methodology. Because of that, the defense won't likely want to bring it up since it would only serve to further convince the jury of the prosecution's case.
With the breathalyser, the police honestly have no idea how the machine should work and even less idea if it actually does work that way. They can't testify to the methodology of the measurement or adequate controls because that's proprietary information. They can't testify that the machine actually does practice that methodology correctly because the source code is not available for their examination. The sample can't be re-tested by an independant lab because none is retained (nor is it practical to retain one).
Because of all of that, a breathalyser test is an adequate screening test only. If it says you're not under the influence, it's not likely worthwhile to get the blood test. It COULD be wrong about that, but if the officer sees no other evidence that you're impaired, it's close enough. It can help when an officer is genuinely trying to decide if the suspect is intoxicated or if they really are "just tired" or "not feeling well".
Like any screening test, a positive reading only indicates that it is worthwhile to perform the definitive test. Never anything more than that. That is, the screening test IS good enough to establish probable cause for the more invasive test.
Notably, if the police had correctly used the breathalyzer only as a screening test and then performed the definitive blood test, none of this would be before the court today since the breathalyzer would only need to meet the standards for probable cause and the actual conviction would be based on the much stronger evidence of the blood test.