If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(Some Guy)   DUI defendant finally gets access to breathalyzer code, ironically finds developers were probably drunk when they wrote it   (dwi.com) divider line 102
    More: Obvious  
•       •       •

13233 clicks; posted to Main » on 14 May 2009 at 12:56 PM (5 years ago)   |  Favorite    |   share:  Share on Twitter share via Email Share on Facebook   more»



102 Comments   (+0 »)
   

Archived thread

First | « | 1 | 2 | 3 | » | Last | Show all
 
2009-05-14 12:59:29 PM
Not necessarily a bad thing

imgs.xkcd.com
 
2009-05-14 01:00:04 PM
artfulwriter.com

After making an attempt to read that this is all I got from it.

/Should have taken more computer classes.
//and a longer attention span
 
2009-05-14 01:02:04 PM
Well I hope no one ever really thought that these tests were about finding the truth and public safety.

/To collect and serve
 
2009-05-14 01:02:07 PM
2. Readings are Not Averaged Correctly: When the software takes a series of readings, it first averages the first two readings. Then, it averages the third reading with the average just computed. Then the fourth reading is averaged with the new average, and so on. There is no comment or note detailing a reason for this calculation, which would cause the first reading to have more weight than successive readings. Nonetheless, the comments say that the values should be averaged, and they are not.

/facepalm
 
2009-05-14 01:02:35 PM
Now, I'm not a legal farker by any means....but isn't this pretty damning for all the DUI cases in which this particular model of breath tester has been used?
 
2009-05-14 01:03:50 PM
Theaetetus: 2. Readings are Not Averaged Correctly: When the software takes a series of readings, it first averages the first two readings. Then, it averages the third reading with the average just computed. Then the fourth reading is averaged with the new average, and so on. There is no comment or note detailing a reason for this calculation, which would cause the first reading to have more weight than successive readings. Nonetheless, the comments say that the values should be averaged, and they are not.

/facepalm


Worst. Mathfail. Evar. This will cause a nightmare in courts now, won't it?
 
2009-05-14 01:05:07 PM
i'd love to take a look at this code.

plus - translated comments? holy shiat, talk about an ass fark. comments written by people that speak english as a second language boggle the mind, too.

in general, i find that many logical solutions and naming conventions vary widely across cultures.
 
2009-05-14 01:06:58 PM
Well, roughly translated tje code allows for

**Catastrophic Error Detection Is Disabled**
 
2009-05-14 01:07:19 PM
Um, that means the last reading has 50% weight, the penultimate one 25%, etc...

Still, facepalm material.
 
2009-05-14 01:07:21 PM
BIGNICKEL: Worst. Mathfail. Evar. This will cause a nightmare in courts now, won't it?

One can hope. But I have a feeling it won't. If anything, police departments will now expend effort on making sure that the breathalyzer model in use is secret.

This illustrates the key problem with the breath test: a lack of transparency. In the American legal system, you are supposed to have the right to question your accuser. But if the accuser is an anonymous block of code that can't be released to the court because of "trade secrets"- how can you properly defend yourself?
 
2009-05-14 01:07:28 PM
So, when the company protested that their code was "proprietary", what they really meant was that it was a poorly-done port of a German Atari program written by people with little if any understanding of the terms precision and reliability. Color me shocked not at all surprised.
 
2009-05-14 01:07:55 PM
Bad: People drinking while impaired by alcohol.
Worse: Using unreviewed, proprietary, closed-source software to detect alcohol levels in drivers.
Worst: Finding out that the breathalyzers were coded for shiat and potentially all DUI convictions based on them will be thrown out.

/People shouldn't drive drunk.
//The government should use competent tools to prevent false convictions.
 
2009-05-14 01:08:27 PM
Sections of the original code and modified code show evidence of using an experimental approach to coding, or use what is best described as the "trial and error" method. Several sections are marked as "temporary, for now".

Good thing this machine is never used to ruin anyone's life. Oh wait.
 
2009-05-14 01:08:49 PM
Theaetetus:
2. Readings are Not Averaged Correctly: When the software takes a series of readings, it first averages the first two readings. Then, it averages the third reading with the average just computed. Then the fourth reading is averaged with the new average, and so on. There is no comment or note detailing a reason for this calculation, which would cause the first reading to have more weight than successive readings. Nonetheless, the comments say that the values should be averaged, and they are not.

/facepalm


Hell, I was more floored by this one:
"12. Defects In Three Out Of Five Lines Of Code"
 
2009-05-14 01:08:51 PM
superoogie: /People shouldn't drive drunk.
//The government should use competent tools to prevent false convictions.


This.
 
2009-05-14 01:08:54 PM
What I get from this article:

- He's probably right, the source code is flawed in ways that will push some borderline cases the wrong way over the line.
- He's still a farking scumbag drunk driver with 99% likelihood, and trying to wriggle out on a technicality. Though it's kind of him to audit software.
 
2009-05-14 01:09:26 PM
SwiftFox: Um, that means the last reading has 50% weight, the penultimate one 25%, etc...

Still, facepalm material.


What he said.

//Double facepalm
 
2009-05-14 01:09:42 PM
BIGNICKEL: Theaetetus: 2. Readings are Not Averaged Correctly: When the software takes a series of readings, it first averages the first two readings. Then, it averages the third reading with the average just computed. Then the fourth reading is averaged with the new average, and so on. There is no comment or note detailing a reason for this calculation, which would cause the first reading to have more weight than successive readings. Nonetheless, the comments say that the values should be averaged, and they are not.

/facepalm

Worst. Mathfail. Evar. This will cause a nightmare in courts now, won't it?


Well, it's not anyone's fault other than the judge for believing the trade secrets crap they were shouting all day. This could have been resolved years ago had it not been for that. Oh wait, this means how many cases umm, nationwide will soon be repealed?

we have something to hide: "TRADE SECRETS!"

/can't wait to see hundreds of thousands of DUI cases retried, maybe all, maybe a repeal of MADD
//DUI is a bad thing to do, MADD is worse
 
2009-05-14 01:09:44 PM
mister aj: He's still a farking scumbag drunk driver with 99% likelihood, and trying to wriggle out on a technicality.

The legal system doesn't work that way. Innocent until proven guilty. As he's pointing out via this audit: there's no proof against him.
 
2009-05-14 01:11:01 PM
chevydeuce: Now, I'm not a legal farker by any means....but isn't this pretty damning for all the DUI cases in which this particular model of breath tester has been used?

If the result of the breathalyser was the primary source of evidence, yes. If the police followed with a blood alcohol sample, then the cases are potentially sustainable.
 
2009-05-14 01:11:34 PM
jerkobson: After making an attempt to read that this is all I got from it.

/Should have taken more computer classes.
//and a longer attention span


It's a mixture of some serious points about sloppy logic and some probably silly nitpicking e.g. lint 3 errors per 5 lines of code likely means it was complaining about things like the return value of strncpy() being ignored. That said I won't check in code that doesn't at least compile clean under gcc -Wall
 
2009-05-14 01:13:05 PM
BIGNICKEL: Worst. Mathfail. Evar. This will cause a nightmare in courts now, won't it?

The software design detects measurement errors, but ignores these errors unless they occur a consecutive total number of times.

WTFF?

The whole thing has the quality of demoware or maybe a skunk works project. Not a production system that can seriously affect people's lives.
 
2009-05-14 01:13:22 PM
Cheops: chevydeuce: Now, I'm not a legal farker by any means....but isn't this pretty damning for all the DUI cases in which this particular model of breath tester has been used?

If the result of the breathalyser was the primary source of evidence, yes. If the police followed with a blood alcohol sample, then the cases are potentially sustainable.


Would the blood tests still be admissible if they were obtained due to evidence from faulty equipment?
 
2009-05-14 01:13:39 PM
The crux of this is that basically any case where these devices have been used is now going to be subject to review... I'd bet a bunch of cases where the main evidence was the reading from this machine will end up getting thrown out. Some of the people may actually have been driving under the influence or may be real threats to everyone else... but theres no way to sort them out from innocent people who are just a bit uncoordinated (failed field sobriety test) without more reliable evidence.
 
2009-05-14 01:15:14 PM
roughridersfan: Cheops: chevydeuce: Now, I'm not a legal farker by any means....but isn't this pretty damning for all the DUI cases in which this particular model of breath tester has been used?

If the result of the breathalyser was the primary source of evidence, yes. If the police followed with a blood alcohol sample, then the cases are potentially sustainable.

Would the blood tests still be admissible if they were obtained due to evidence from faulty equipment?


Nah, it's not a tainted tree thing if they did a blood test... all they need to request the blood test is probable cause (or consent). That said, blood tests are not done in all cases.
 
2009-05-14 01:15:32 PM
roughridersfan: Cheops: chevydeuce: Now, I'm not a legal farker by any means....but isn't this pretty damning for all the DUI cases in which this particular model of breath tester has been used?

If the result of the breathalyser was the primary source of evidence, yes. If the police followed with a blood alcohol sample, then the cases are potentially sustainable.

Would the blood tests still be admissible if they were obtained due to evidence from faulty equipment?


All the prosecution needs to do in order to get blood tests allowed as legally obtained evidence is proove that the arresting officer had probable cause. Erratic driving, slurred speech, and a failed roadside test would each individually qualify for probable cause. So long as the arresting officer didn't just shove a breathalyzer in the suspect's face without witnessing any likely indicators of drunken behaviour, it shouldn't be too hard to get probable cause accepted by the courts.
 
2009-05-14 01:16:01 PM
superoogie:
Hell, I was more floored by this one:
"12. Defects In Three Out Of Five Lines Of Code"

From what I can tell they used lint, so that's not surprising; according to lint, pretty much every application ever written has that many "defects."
 
2009-05-14 01:16:18 PM
roughridersfan: Cheops: chevydeuce: Now, I'm not a legal farker by any means....but isn't this pretty damning for all the DUI cases in which this particular model of breath tester has been used?

If the result of the breathalyser was the primary source of evidence, yes. If the police followed with a blood alcohol sample, then the cases are potentially sustainable.

Would the blood tests still be admissible if they were obtained due to evidence from faulty equipment?


I guess the question is- is it really faulty. If it does detect alcohol levels in a significantly correct manor the significant majority of the time, then it seems legitimate to go to a blood test. The blood test nails you, it's over. I'm not sure if this is actually a full-blown fruit-of-the-poison-tree thing, but then I'm not a lawyer.

/or have a GED in law
 
2009-05-14 01:16:49 PM
jerkobson: After making an attempt to read that this is all I got from it.

/Should have taken more computer classes.
//and a longer attention span


same here... I got sleepy while reading it. Too many words. I have to read a lot of cancer research papers at work & right now during my lunch break & I don't feel like reading that much.
 
2009-05-14 01:16:49 PM
Readings are Not Averaged Correctly: When the software takes a series of readings, it first averages the first two readings. Then, it averages the third reading with the average just computed. Then the fourth reading is averaged with the new average, and so on. There is no comment or note detailing a reason for this calculation, which would cause the first reading to have more weight than successive readings. Nonetheless, the comments say that the values should be averaged, and they are not.

Without seeing the code, I can't tell if they are averaging correctly or not. However, isn't the assertion that it would cause the first reading to have more weight incorrect?

Let's say reading 1 was a 10, and then each subsequent reading was a 2.

Avg readings 1,2: 6
Avg of that result and reading 3: 4
Avg of that result and reading 4: 3

Where if you had averaged them all it would be 4: (10+2+2+2)/4, showing that the first reading had less weight than it should have in the former.

However, it is still possible that the code was correctly calculating the average. If it keeps track of how many readings had been made and uses that, for exampe:

Avg of readings 1,2: 6
Avg of two times that result and reading 3: 14/3 = 4.667
Avg of three times that result and reading 4: 4
 
2009-05-14 01:16:50 PM
roughridersfan:
Cheops: chevydeuce: Now, I'm not a legal farker by any means....but isn't this pretty damning for all the DUI cases in which this particular model of breath tester has been used?

If the result of the breathalyser was the primary source of evidence, yes. If the police followed with a blood alcohol sample, then the cases are potentially sustainable.

Would the blood tests still be admissible if they were obtained due to evidence from faulty equipment?


IANAL, but I believe as long as the breathalyzer wasn't the sole source of evidence of intoxication before the blood test, it won't get thrown out.
 
2009-05-14 01:17:14 PM
With that said, yes you can convince a judge, but in order to convict you still need to convince a jury. A halfway competent defense attourney should be able to tear your case apart in front of a jury if you have one of these breathalysers on the evidence table.
 
2009-05-14 01:18:18 PM
roughridersfan: Would the blood tests still be admissible if they were obtained due to evidence from faulty equipment?

Yes.

Spontaneous Defenstration: From what I can tell they used lint, so that's not surprising; according to lint, pretty much every application ever written has that many "defects."

I don't do much C, so I don't know lint very well. I know most code audit tools are... generous with error messages. Based on my experience in other technologies- it still sounds like a lot of audit failures.
 
2009-05-14 01:19:06 PM
t3knomanser:The legal system doesn't work that way. Innocent until proven guilty. As he's pointing out via this audit: there's no proof against him.

Well, there **is** proof - flawed, untested, incomplete, and logically inconsistent proof - but still proof nonetheless.

/Glad I've never had to prove myself in that regard...
 
2009-05-14 01:19:34 PM
Cheops:
With that said, yes you can convince a judge, but in order to convict you still need to convince a jury. A halfway competent defense attourney should be able to tear your case apart in front of a jury if you have one of these breathalysers on the evidence table.

Not to mention that now that there is rather damning precedent, any competent lawyer defending a DUI will ask for the same investigation of any other model of breathalyzer as well.
 
2009-05-14 01:19:45 PM
Das Meister: but still proof nonetheless.

Well, evidence. Hardly "proof". And not terribly good evidence either.
 
2009-05-14 01:20:08 PM
So, is this for all breathalyzers? Like ones in Canada as well, cuz I have some buddies with some DWI charges who don't have court yet. If this affects all of em they might have a chance and not go to jail.

But still, big FAIL on the government's job. You would think a piece of equipment like that designed to save lives and an integral piece in a very serious conviction would have been better designed. Our lowest bidder mentality in action??
 
2009-05-14 01:20:20 PM
superoogie: DUI will ask for the same investigation of any other model of breathalyzer as well.

As they should.
 
2009-05-14 01:20:30 PM
Theaetetus

Agreed. This is epic fail.

superoogie

Hell, I was more floored by this one:
"12. Defects In Three Out Of Five Lines Of Code"

This is relative. You can run lint at different levels and suppress certain messages. Plus lint finds suspicious behavior and non-portable code. So calling this 60% of the lines of code have an error in it is a value call and not in any way empirical (having said that, lint is a highly valuable tool in the C programmer's arsenal....)
 
2009-05-14 01:23:13 PM
It's still probably better than a cop telling you to walk a straight line and touch your nose with your forefinger.
 
2009-05-14 01:23:52 PM
"the Computer Operating Property"

teehee, "the COP"
 
2009-05-14 01:26:39 PM
superoogie: Bad: People drinking while impaired by alcohol.

I find it's more fun that the alternative.
 
2009-05-14 01:29:46 PM
This ruling is a good thing. The evidence used for obtaining convictions must be above reproach or suspicion.
 
2009-05-14 01:30:32 PM
t3knomanser:
superoogie: DUI will ask for the same investigation of any other model of breathalyzer as well.

As they should.


If they are anything like this company, damn straight they should.
 
2009-05-14 01:34:13 PM
While Draeger's counsel claims that the "The Alcotest [7110] is the single best microprocessor-driven evidential breath tester on the market", Draeger has already replaced the antiquated 7110 with a newer Windows® based version, the 9510. The computer code in the 7110 is written on an Atari®-styled chip, utilizing fifteen to twenty year old technology in 1970s coding style.

This is reaching. First, what is an "Atari®-styled chip"? if they mean a 6502 type, they still work fine for embedded devices, even if they are not stylish.

Should concentrate on the real bugs, like the averaging one. Looks like that means only the samples from deepest in the lungs are significantly included.
 
2009-05-14 01:34:19 PM
As for blood tests....I know a couple of people who have gotten DUI's....additionally, I have done ride-alongs with the local cops a few times and have seen several people get hooked up for DUI...In none of those cases did anybody submit to a blood test....It was probable cause stop, detect alcohol smell, do the drunky roadside dance, go to jail, do a breathalyzer, if they blew more than .08, they got to spend some quality time in the family room of the county jail for a few hours.....three to six months later, pay an ass load of money and wait seven years for your record to clear.....

this is in California, I can't speak for other states...in fact, in Cali, the cop asks whether you would prefer to submit to a blood, breath or urine test (after he has placed the bracelets on you) and if you refuse all, you are automatically deemed guilty by the DMV and lose you license for some amount of time, drunk or not
 
2009-05-14 01:35:39 PM
Cheops: chevydeuce: Now, I'm not a legal farker by any means....but isn't this pretty damning for all the DUI cases in which this particular model of breath tester has been used?

If the result of the breathalyser was the primary source of evidence, yes. If the police followed with a blood alcohol sample, then the cases are potentially sustainable.


Yup... just as soon as we have had a chance to review the code for the blood testing machine.
 
2009-05-14 01:36:45 PM
There is a flip side here too that's just as scary, if not more so.

From TFA:
An incorrect breath test could lead to accidents and possible loss of life, because the device might not detect a person who is under the influence, and that person would be allowed to drive.

So this Atari breathalyzer had two ways to ruin lives!

Also, this analysis is two years old and it says the machine's been replaced with a new model. So what about this new version?

/Yeah, I read the whole thing and I don't even drink at all.
 
2009-05-14 01:37:10 PM
Sue Dunham: It's still probably better than a cop telling you to walk a straight line and touch your nose with your forefinger.

I disagree... at least that test measures actual impairment and not your BAC (which varies widely in relation to impairment from person to person).
 
2009-05-14 01:37:26 PM
Cheops: chevydeuce: Now, I'm not a legal farker by any means....but isn't this pretty damning for all the DUI cases in which this particular model of breath tester has been used?

If the result of the breathalyser was the primary source of evidence, yes. If the police followed with a blood alcohol sample, then the cases are potentially sustainable.


That may very state to state but I know that is not standard procedure in WI. You have to request a blood draw and pay for it.

Basically though the process of detailing the behavior of the accused will be enough to sustain most convictions. The breathalyzer is just the final nail in the coffin. Or they will turn around and charge you with OWI instead of PAC (in WI)
 
Displayed 50 of 102 comments

First | « | 1 | 2 | 3 | » | Last | Show all



This thread is archived, and closed to new comments.

Continue Farking
Submit a Link »






Report