Police Blotter: Defendant wins breathalyzer source code

[Tesha]

Source: ZDNet news

Police Blotter: Defendant wins breathalyzer source code

By Declan McCullagh
Published on ZDNet News: Aug 9, 2007 5:23:00 AM

What: Drunk driving defendant says he needs the source code to the
Intoxilyzer 5000EN to fight the charges in court.

When: Minnesota Supreme Court rules in his favor on July 26.

Outcome: Source code will be turned over to defense attorneys.

What happened, according to court records and other documents:
When Dale Lee Underdahl was arrested on February 18, 2006, on suspicion
of drunk driving, he submitted to a breath test that was conducted using
a product called the Intoxilyzer 5000EN.

During a subsequent court hearing on charges of third-degree DUI, Underdahl
asked for a copy of the "complete computer source code for the (Intoxilyzer)
currently in use in the state of Minnesota."

An article in the Pioneer Press quoted his attorney, Jeffrey Sheridan, as
saying the source code was necessary because otherwise "for all we know,
it's a random number generator." It is hardly new technology: One criminal
defense attorney says the Intoxilyzer is based on the antique Z-80 microprocessor.

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.

<more>


Read more: http://news.zdnet.com/2100-9595_22-6201632.html

---

This might have interesting ramifications for various
litigation and legislation going on re: Electronic
Voting machines.

Tesha

[ChairmanAgnostic]

1. brilliant!

good for him.

[robinlynne]

2. excellent. kicking.

[AX10]

3. EXCELLENT!

:thumbsup:

Could we work to get the source codes for electronic voting machines next?

[DaveJ]

4. There is no reason to hide the source code

For simple devices like this one, or voting machines, the only reason to hide the source code would be to cover up mistakes or intentional back doors put in to facilitate corruption.

[rurallib]

20. Just like the wingers always say "If you have nothing to hide."

[AngryAmish]

21. Companies say the source code is a trade secret

This is a dirty little secret. The machine is tested by the federal gov't. They say it works. Because the gov't says it works then the state say it works and the defendant has no way to challenge it or even see if the testing was on the up and up.

Ever wonder why the cops have you walk the line, touch your nose, follow a pen et al. Because the federal government says these are ways to detect intoxication. Was this done with peer reviewed research to see if it works? No, they just say so. Same thing.

A defendant has a right to confront each and every element of the charge.

My guess? The source code will never be turned over and so all the outstanding DUIs in Minn will be tossed. Then everyone has to buy new breathalyzers and we start the dance again.

[MLFerrell]

39. Actually, the pen-light test DOES have a medical basis.

It's called Nystagmus.

http://en.wikipedia.org/wiki/Nystagmus#Nystagmus_and_alcohol

In police work, testing for horizontal gaze nystagmus is one of a battery of field sobriety tests used by officers in the field to determine whether a suspect is driving under the influence of alcohol. The test involves observation of the suspect's pupil as it follows a moving object, noting (1) lack of smooth pursuit, (2) distinct and sustained nystagmus at maximum deviation, and (3) the onset of nystagmus prior to 45 degrees. A general rule of thumb is that a person's blood alcohol concentration can be estimated by subtracting the angle of onset from 50 degrees. Therefore, a person with an angle of onset of nystagmus at 35 degrees has a blood alcohol concentration of approximately 0.15%. Unfortunately, due to the complexities of eye conditions and physiology, police (with very limited ophthalmologic training) may be unable to disguish horizontal gaze nystagmus from other conditions such as occular flutter.

[AngryAmish]

43. But who said having nystagmus is predictive of having above .08?

The test is crap because 1. the testing conditions are awful 2. the training is awful

It is almost the definition of junk science.

[ProudDad]

25. Odds are the code is crap

9 out of a 10 computer programmers can't program...

[Tesha]

28. I always thought it was...

> 9 out of a 10 computer programmers can't program...

I always thought it was "11 out of 10 computer programmers..." ;)

If I had a 5.00001 cents for every "off by one" error...

Tesha

[ProudDad]

29. Don't forget

there are 10 kinds of people in this world

Those who know binary and those who don't... :hi: :evilgrin:

[EnricoFermi]

35. Typically the worst code

is written by the people holding the least desired jobs, just due to competition.

You can pretty much guarantee that this code is crap.

[Warpy]

5. I hope it does, but I'm not holding my breath(alyzer)

Anybody who fails a breathalyzer test needs to insist on a serum alcohol.

It will either convict you or acquit you.

[lildreamer316]

38. ? Uh-oh. What is a serum alcohol?

Husband recieved second DUI last year. Yes, he has a slight history. I am not arguing the point,believe me. However, he swears (and he is not known for lying to me) that he had his last drink nearly NINE hours before; if not almost ten. Yet, he blew a .12( I believe). No idea how this is really possible, even given all the variances of the human body, but there's no real way to prove that he wasn't drunk now.

[Tesha]

40. You can always ask that blood be drawn...

You can always ask that blood ("serum") be drawn
and analyzed for alcohol content; that's the final
word in whether or not you meet the legal standard
of "intoxicated" (since the law is written in terms
of "Blood Alcohol concentration").

Tesha

[Warpy]

45. It's a blood alcohol, requires a needle stick and lab fee

and if he blew that much, it would have confirmed it. Anyone who blows a .08 needs to get the test.

If his last drink was 9 hours earlier, it had to have been a pint and a half of vodka.

The male body metabolizes one drink per hour. That's one ounce of booze, one twelve ounce can of beer, or one four ounce glass of wine.

[notadmblnd]

6. it has me wondering....

how many drunk driving convictions are bogus.

[Mike Daniels]

13. Very few I imagine

Given that the breath test is just one process and given that road stops and sobriety checks are videotaped I doubt many people have been falsely convicted of being DUI or DWI.

People who protest DWI charges generally know they're guilty but they're trying to game some quirk in the law that they hope will get them off.

There was a case in Virginia about 10 years ago where some lawyers were trying to argue that the 1 week suspension of your license after being charged with DUI/DWI was effectively a conviction and that the court case was in effect double jeopardy. Every court shot down the argument and I imagine that there was no shortage of people who were without-a-doubt guilty of DWI/DUI who tied up a lot of court time by appealing cases until the matter was finally resolved.

[EnricoFermi]

30. Many, you'd be amazed

I know of many, personally.

A friend of mine was convicted of a DUI, for sleeping on the hood of the car, in front of his house, and having the keys in his pocket. Most people just don't fight it. Another friend was given a DUI for sleeping in her passenger seat during the winter, in this case with the keys in the ignition, to listen to the radio. How either of these were driving related, are just a mystery to me.

The system is a revenue machine, not a way to prevent people from getting hurt, in most cases. The entire traffic system, for that matter, is pretty much designed to get money for the police. Why can you double your fine to get a speeding ticket reduced to a non-moving violation? That doesn't even make sense.

Furthermore, the BAC required in each state has little to do with what adversely effects ones driving. Each person could also be effected differently, and I am pretty damn sure that they don't base this level (0.08 usually) on scientific studies. Lastly, the poor and minorities are usually the victims of the system. They are less likely to defend themselves and the cops even seek them out. Where I live, a policeman let me know that they are told not to pull cars over that are worth $60,000 or more, because those people will almost always successfully defend themselves in court, and it therefore will waste time and money.

There is a reason behind the term, driving while black.

[EnricoFermi]

31. Whenever they need more revenue, they either lower the BAC or increase the fine

It, as with most abuses of civil liberties, is under the guise of keeping the streets safe. As usual, there are often tragedies used as examples to justify their point. They use the public being appalled to their own benefit when passing legislation.

Take sex offender registrations as an example.

We pass countless laws each year, at the city, state, and federal levels, that further encroach on our freedoms. While I do not support drunk driving or consider it a "civil liberty", the way it is prosecuted and targeted, are further examples of how the court system is unfair to lower classes to the benefit of the upper class and corporations. Keeping people down, whether it is for their own mistakes or not, is the great tragedy within our justice system.

Revenue should never be considered when making laws, enforcing them, or prosecuting the public.

[tridim]

7. Runs on a Z-80 and probably costs thousands of dollars to purchase

I bet a dollar that there are still Z-80's on the space shuttle and Air Traffic Control systems too.

Massively obsolete hardware is a failure of infrastructure, just like bridges and levees.

Oh, and someone should check the source code in radar guns too. I know I've been screwed by those.

[Tesha]

9. I gotta tell you...

> I bet a dollar that there are still Z-80's on the space
> shuttle and Air Traffic Control systems too.
>
> Massively obsolete hardware is a failure of infrastructure,
> just like bridges and levees.

I gotta tell you...

I'd rather have some old and well-tested 8080, 8085, or Z-80
firmware than a lot of the new codes that I see nowadays.

Software (or firmware) in embedded applications doesn't
"rot" the way, for example, a Windows application does.
If it works, it keeps on working. How often do you need
to upgrade the firmware in your microwave oven?

Personally, I think this breathalyzer guy is probably
guilty and is just fishing for an excuse to be let off,
but I'm very glad of the legal precedents that are being
established in these cases; as I said in the original
posting, they'll be very helpful as we wage the Battle
of the Bogus Ballot Box.

Tesha

[tridim]

10. Well lets take my clock radio for example

It sucks, and has always sucked. The firmware is horrible, probably because it's the same chip they've been using for 30 years. All basic clock radios are the same, and they all suck. It's time for an update.

I realize the z80 is a tried and true platform, but that doesn't mean it's not old, slow, power hungry and generally obsolete. There are newer embedded platforms that are cheaper, more powerful and just as reliable as the older systems. However, I do worry about new embedded software that could easily be used to steal more of our privacy. I wouldn't want my clock radio calling "home" to report my sleeping habits, but I would like it to be open-source and mod-able.

[Tesha]

12. You do understand that most Z-80s these days aren't "real" Z-80s, right?

> I realize the z80 is a tried and true platform, but that
> doesn't mean it's not old, slow, power hungry and generally
> obsolete.

You do understand that most Z-80s these days aren't "real"
Z-80s, right? Instead, you go to an "IP" (Intellectual
Property) vendor, buy a Z-80 design (a "Z-80 core"),
and plug it into a very few percent of your gate array
or ASIC. Then the same gate array or ASIC that's doing
everything else in your design also contain a Z-80
processor for executing the control functions.

Because it fits in a very small part of your gate array,
it costs practically nothing. And it probably clocks 50
times faster than a real Z-80 as well. And if you don't
like the Z-80, buy an 8051 core instead. Or a PowerPC
core; they've got 'em all!

Tesha

[tridim]

16. No, I didn't know that.

But I'm just a lowly web developer. :)

The last z80 I worked with was a Timex Sinclair I bought for almost nothing after people realized they were crappy computers even by 1980's standards. My brother did assembly on z80's in school (1994), but by that time I had moved on to bigger and better things.

[sillyphoenix]

11. Radar guns

Radar guns are inaccurate because of the principle on which they work, not necessarily because of the code.

IIRC, they use the Doppler Effect (same thing that changes the pitch of an ambulance siren when it passes you) to determine how fast you're going. The problem is, the emitter has to 1) be perfectly stationary 2) have an uninterrupted line of sight to your car for it to work correctly.

Ergo, any readings taken from a moving car or a handheld unit are questionable at best. I heard you can point a radar gun at a tree and it'll read ~2 mph. No amount of source code modification is going to fix the laws of physics.

[tridim]

14. If they aren't accurate how can they be used as evidence?

I got clocked for 85 in a 65 once while my cruise control was set at 70.

The judge said the gun doesn't lie and I had to pay the fine or plead not-guilty, get a lawyer and go to trial.

[sillyphoenix]

15. Ignorance

I'm willing to bet that most judges and nearly all police officers have no idea what the Doppler effect is, or even how the guns work. They say "oooh, gadgetry.... if it's electronic, it must be infallible."

[EnricoFermi]

34. I would fight all of these if I had the time

and you have to be an "expert" to be taken seriously, which is wrong in principal.

Fortunately, I have the background to be considered one, but due to the system, I would have to take way too much time off of work to justify battling them. When you are standing in line in traffic court, it is pretty much just a line to pay the county, not a real court. Besides, I can usually just get it reduced to a non-moving violation, which I said made no sense anyway.

As for any of these devices, many of the components have tolerances in them that are usually 5% (and add up), with any number of them capable of failing or acting quirky at any time. Contrary to popular belief, components, microchips included, will do very strange things if faulty or nearing failure. They don't always just stop working and could easily throw incorrect readings. These devices are hardly perfect, and if you look at the circuit boards on some of these government products they are often at their 12th redesign or more. Something had to have been wrong in the previous versions, yet they were allowed to be used in court as evidence.

Breathalyzers don't directly measure alcohol anyway and can be effected by other chemicals, light conditions, and humidity.

Read about how the three types work, and you'll quickly see the places for error:
http://searchwarp.com/swa65174.htm

[whopis01]

22. That is not exactly correct.

Radar guns are not inaccurate because of the principle on which they work. You can measure speed with incredible accuracy using the Doppler effect. And you do not have to be stationary to do so. With that said, the radar gun still has to be properly built and calibrated to be accurate. It also has to be properly used.

The most common cause for an inaccurate reading is that the gun targeted a different vehicle (perhaps an oncoming vehicle in the other lane). That's not such a problem from a moving vehicle, but it can be from a parked vehicle.

And of course, the guns need to be calibrated properly. It isn't hard to do, but it does need to be done.

[benEzra]

44. Common sources of radar error:

http://www.motorists.org/fightticket/home/common-radar-errors/">Common Radar Errors

Explains the most common sources of radar error, with pictures.

[whopis01]

46. several of those are not common and are corrected for on most radar systems

Older style radar guns could be subject to those errors, but newer ones are not. That page references tests done in 1979 and the quality and performance of the systems has improved greatly since then. Of course, older guns are still in use, but my first post was disagreeing the idea that it was principle upon which they worked that caused the problem. The principles are fine, they just have to be implemented correctly.

The interference errors and errors caused by operating from a moving vehicle have effectively been eliminated on modern systems. You do still have to make sure you are targeting the right vehicle though.

[ProudDad]

26. I've been clocked at over 140 mph

on my motorcycle...

They don't work worth crap with bikes...

[flyingfysh]

18. Z-80s are actually good machines

They have been around so long that all the quirks and glitches are well-understood.

Another old machine, the 6502 (used in the Apple II) is used in heart pacemakers; there are no surprises with that processor, and software updates in that application are impossible.

[tridim]

19. Thanks, I'm learning a lot of new things about the old z80 today. nt

[ProudDad]

27. Worked great in my Osborne I (n/t)

[Canuckistanian]

24. Massively obsolete?

Why does a radar gun need to have the very latest Pentium?

Microprocessors like the Z80 and 8051 are indeed being used for avionics, space and other mission-critical applications. Why? Because they're robust, WELL-understood, easy to code for and they do the job. Simple applications call for simple solutions.

Plus, they're physically robust as well, withstanding conditions that modern processors weren't designed to do.

Even the International Space Station uses 386SX processors! For those that remember, it was two generations BEFORE the Pentium. Simply because they use less power.

[EnricoFermi]

36. I agree

The more complicated machine is more likely to fail because it is the sum of all of its parts. With more transistors and a more complex pipeline, it's that much more likely to have problems. However, there are other factors, including more advanced fabrication techniques that could skew it in favor of the Pentium like processors. I do remember a floating point error in the 90's though inside one of Intel's processors.

That said, the most important key, is how often the device was used, and essentially tested. A customized ASIC, even if it is based on the Z80, could easily have design specific flaws that weren't in the core. Calibration and statistical analysis of existing radar guns and breathalysers should be taken into account. A police department should have these records if they plan on convicting you.

From what I have read, they are usually uncalibrated and often used without proper training. I would say that unless they have proof that there wasn't interference (probably video of just you on the road), and that the device was recently calibrated and tested, that they can't convict you of any wrongdoing. Even NASA uses redundancy of electrical components for a reason. They can and do fail.

[Tandalayo_Scheisskopf]

8. There are cases waiting...

Here in NJ, because of this issue. In fact, IIRC, there have been some petitions for the source code as well.

If, after a professional source code audit(and I have no doubt that this code may well make Diebold's code look like a shining example of best practices...), this code is found to be faulty or allow operator manipulation of the results, this will be huge beyond huge here.

You might ask why?

NJ has a long history of stumblin', bumblin', fumblin' fiscal mismanagement. Endemic political corruption, "wheeling", "Christmas Tree Earmarks", just plain bribery(Hey, Sharp James! What up, homie?)...It has been called "The crucible for every bad and corrupt government practice in the nation". If it is corrupt, it has been prototyped in NJ.

One of the worst was our governoretta, Christie Whitman. Man, she left us a pile of debt and more fiscal messes than Carter has little pills. Back during her maladministration, traffic offenses were institutionalized as a leading revenue driver for every level of government in the state. The cornerstone of that was, of course, the DUI. Part of that plan was to give half of the take of DUI fines to police departments for an abomination called "The DUI Overtime Enforcement Fund". Long story short, that meant that for every DUI conviction, that meant money directly in the pockets of the cops on patrol. Of course, they went fucking wild. Since, at the time, they were using the Remington breathlyzer, a unit that allows the operator(usually the arresting officer) to dial in any number he likes, a lot of lawyers and consultants made a lot of gelt and a lot of people beat the rap. So then, the state changed to these machines. Now, this ruling.

The fact is that police departments, and their myriad toys, make really shitty revenue collection agencies, because that leads to public resentment and that erodes public respect for institutions. When their myriad toys are faulty and manipulable, in detriment to the citizens, that just compounds the problem exponentially.

[redqueen]

17. NOW can we get the DRE source code? PLEASE?

Christ almighty.

[The Sushi Bandit]

23. I want the source code for my wifes HOME PREGNANCY TESTER!!

jut kidding!

[rateyes]

32. Think it will change the results?

:rofl: Good one.

[Tesha]

37. Maybe he just wants to surprise his wife???

"We'll just change this "0" to a "1" right here at
good ole byte 0x375..."

Tesha

[rateyes]

42. It's amazing what technology can do these days!

:rofl: If only I had known when I was having my kids...I would have just mailed it in!

:rofl:

[Tellurian]

33. The whole point of getting the source code info...

Is, as all have agreed, the source code is the lock on how information is processed by the technology.

The defendant has the Right to Cross Examine that information first, seeing if the criteria used in the process is valid, second, then test it to subjectively and objectively ascertain the reliability of it's accuracy. Further and most importantly, it should be the very device used at the STOP site. If not, that would also be another problem needing to be addressed.

[Robb]

41. To my tired eyes, "5000EN" looked like "sodden"

...interesting coincidence. :)