Locked Out Of The Sixth Amendment By Proprietary Forensic Software

from the unconfrontable-witnesses dept

There are all sorts of proprietary code out there, unavailable for inspection and perfectly capable of sending people to prison, as Slate's Rebecca Wexler reports.

Defendant Martell Chubbs currently faces murder charges for a 1977 cold case in which the only evidence against him is a DNA match by a proprietary computer program. Chubbs, who ran a small home-repair business at the time of his arrest, asked to inspect the software’s source code in order to challenge the accuracy of its results. Chubbs sought to determine whether the code properly implements established scientific procedures for DNA matching and if it operates the way its manufacturer claims. But the manufacturer argued that the defense attorney might steal or duplicate the code and cause the company to lose money. The court denied Chubbs’ request, leaving him free to examine the state’s expert witness but not the tool that the witness relied on.
That's a starkly mercenary stance to take. The "trade secret privilege" invoked here basically states that the company's potential loss of income outweighs a person's potential loss of freedom. It also asks for a level of trust it hasn't earned: that the software is as close to infallible as it needs to be. Cross-examination is next to useless when the software itself can't be examined.

Worse, this closed-off software operates in a field where nearly every previous form of "indisputable" evidence has proven to be severely flawed.
Studies have disputed the scientific validity of pattern matching in bite marks, arson, hair and fiber, shaken baby syndrome diagnoses, ballistics, dog-scent lineups, blood spatter evidence, and fingerprint matching. Massachusetts is struggling to handle the fallout from a crime laboratory technician’s forgery of results that tainted evidence in tens of thousands of criminal cases. And the Innocence Project reports that bad forensic science contributed to the wrongful convictions of 47 percent of exonerees.
Everything tied to securing convictions seems to suffer from pervasive flaws compounded by confirmation bias. For four decades, the DOJ presented hair analysis as an unique identifier on par with fingerprints or DNA when it wasn't. A 2014 Inspector General's report found the FBI still hadn't gotten around to correcting forensic lab issues it had pointed out nearly 20 years earlier. This contributed to two decades of "experts" providing testimony that greatly overstated the results of hair analysis. All of this happened in the FBI's closed system, a place outsiders aren't allowed to examine firsthand.

That's the IRL version. The software version is just as suspect. Computers aren't infallible and the people running them definitely aren't. If the software cannot be inspected, the statements of expert witnesses should be considered highly dubious. After all, most expert witnesses representing the government have a vested interest in portraying forensic evidence as bulletproof. Without access to forensic software code, no one will ever be able to prove them wrong.

If a piece of software has the ability to deprive a member of the public of their freedom, its code should be open for inspection by the defense. "Trade secrets" should not take precedence over the public's right to defend themselves in court. Even in the highly unlikely event that Chubb's defense team would have copied the code and destroyed the company's future profits, it would still have the ability to seek redress through the court system. After all, that's the line the government uses when it argues for expanded "good faith exceptions" or warrantless searches and seizures: "Hey, if we screw up, you can always sue."

The judicial system is a remedy for wrongs, both criminal and civil. What it shouldn't be is a protective haven where ridiculous assertions like those made here are used to prevent an accused person from learning more about the evidence being used to convict them.


Reader Comments

Subscribe: RSS

View by: Time | Thread


  • icon
    Gary (profile), 8 Oct 2015 @ 3:45pm

    This seems pretty straightforward. It's under attorney privilege, it isn't reasonable to assume it will be stolen. They need to provide the evidence or toss it out.

    reply to this | link to this | view in chronology ]

  • icon
    Retsibsi (profile), 8 Oct 2015 @ 3:53pm

    Not quite as straightforward an argument as it seems. If I read the report correctly the argument put for forward by the defendant's own expert is that he cannot determine the accuracy of Chubb's system without having sight of the source code. But that's not really the issue here. What doesn't seem to be asked by the defendant is leave to carry out his *own* DNA testing using another system nominated by his expert. If there isn't a match then clearly he will have grounds to attack the state's expert testimony

    reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 8 Oct 2015 @ 4:09pm

    Intellectual Property Trumps Constitutional Rights?

    Amendment VI

    "In all criminal prosecutions, the accused shall enjoy the right to a speedy and public trial, by an impartial jury of the state and district wherein the crime shall have been committed, which district shall have been previously ascertained by law, and to be informed of the nature and cause of the accusation; to be confronted with the witnesses against him; to have compulsory process for obtaining witnesses in his favor, and to have the assistance of counsel for his defense."

    reply to this | link to this | view in chronology ]

    • identicon
      Anonymous Coward, 9 Oct 2015 @ 7:05am

      Re: Intellectual Property Trumps Constitutional Rights?

      OK, but the software is not the witness; the so-called expert is. Software is just a the thought process by which the witness came to his conclusion.

      (Remember all those debates about why software isn't patentable because it's just a thought process, which explicitly isn't patentable? Same thing applies here.)

      If the witness can't explain his thought process (as enacted through instructions to a general purpose computer) to the jury that is grounds for questioning his credibility. Same as if he declared the accused to be a witch but couldn't explain why.

      reply to this | link to this | view in chronology ]

      • icon
        lucidrenegade (profile), 9 Oct 2015 @ 11:03am

        Re: Re: Intellectual Property Trumps Constitutional Rights?

        Same as if he declared the accused to be a witch but couldn't explain why.

        V: There are ways of telling whether she is a witch.
        P1: Are there? Well then tell us! (tell us)
        V: Tell me... what do you do with witches?
        P3: Burn'em! Burn them up! (burn burn burn)
        V: What do you burn apart from witches?
        P1: More witches! (P2 nudge P1)
        (pause)
        P3: Wood!
        V: So, why do witches burn?
        (long pause)
        P2: Cuz they're made of... wood?
        V: Gooood.
        (crowd congratulates P2)
        V: So, how do we tell if she is made of wood?
        P1: Build a bridge out of her!
        V: Ahh, but can you not also make bridges out of stone?
        P1: Oh yeah...
        V: Does wood sink in water?
        P1: No
        P3: No. It floats!
        P1: Let's throw her into the bog! (yeah yeah ya!)
        V: What also floats in water?
        P1: Bread
        P3: Apples
        P2: Very small rocks
        (V looks annoyed)
        P1: Cider
        P3: Grape gravy
        P1: Cherries
        P3: Mud
        King: A Duck!
        (all look and stare at king)
        V: Exactly! So, logically...
        P1(thinking): If she ways the same as a duck... she's made of wood!
        V: And therefore,
        (pause & think)
        P3: A witch! (P1: a witch)(P2: a witch)(all: a witch!)

        reply to this | link to this | view in chronology ]

  • identicon
    Vic, 8 Oct 2015 @ 4:26pm

    Trade secret?

    A strange scenario just crept in. What if the manufacturer to claim it's a kind of DRM and it is covered by DMCA, then it would be even illegal to inspect the software, as I understand...

    So, it's your choice - go to prison for a murder or for a an attempt to decode a piece of software. And you never know ahead which sentence is longer...

    reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 8 Oct 2015 @ 4:45pm

    The "trade secret privilege" invoked here basically states that the company's potential loss of income outweighs a person's potential loss of freedom.

    The flag of the People's Republic of China goes up in front of our home tomorrow. No doubt later in the day the Stasi will kick in our door and we will never be seen or heard from again. Just drop all of your ordinance on my position, live free or die.

    reply to this | link to this | view in chronology ]

  • identicon
    Jyjon, 8 Oct 2015 @ 5:11pm

    What an interesting way to make a living, a string of heinous crimes so you can access the source code from various companies that you can sell to their competition.

    That totally makes sense.

    reply to this | link to this | view in chronology ]

  • identicon
    Song from the woods, 8 Oct 2015 @ 5:44pm

    there are no surprises

    there are no studies that verify ANY forensic "science" not even fingerprints never mind the rest of the "lies to convenience people in power", which state was it again in which %30 of death row in mates where definitively NOT GUILTY via DNA ?.. thousands have already been murdered by the state, time to stop caring what the state says.

    reply to this | link to this | view in chronology ]

    • identicon
      Anonymous Coward, 8 Oct 2015 @ 6:06pm

      Re: there are no surprises

      I doubt the many innocent people in prison are there solely due to so called "bad forensic science". I would guess that most are there because of a severe lack of public defenders, severe lack of funding for public defenders and a completely screwed up plea bargain system that is in place to enrich the private prison industry.

      reply to this | link to this | view in chronology ]

      • icon
        Starke (profile), 8 Oct 2015 @ 6:21pm

        Re: Re: there are no surprises

        And also "bad forensic science." There's an absurd willingness of juries in this country to accept anything presented in court as "forensic" fact, even when it's just the expert witness making an educated guess and then presenting that as fact.

        The underfunding of public defenders doesn't help. But, you tell a jury this guy in a suit can prove it through means they don't really understand, and then they'll think back to all those nights they fell asleep watching CSI, and think, "well I trust these guys, they've got magic science to back them up."

        reply to this | link to this | view in chronology ]

        • identicon
          Anonymous Coward, 8 Oct 2015 @ 7:34pm

          Re: Re: Re: there are no surprises

          I have read that judges have instructed juries to not expect CSI type investigation, data or analysis.

          Asking for a conviction based solely upon DNA is a stretch. Is there no circumstantial evidence to support the charges? For example, was the dude even in town at the time? Did he use a credit card? The lack of supporting evidence is ridiculous and s bit suspicious.

          reply to this | link to this | view in chronology ]

          • identicon
            Song from the woods, 8 Oct 2015 @ 8:14pm

            Re: Re: Re: Re: there are no surprises

            I said many where exonerated on dna evidence, not the reverse, of course you need other evidence to convict, but many where demonstrably NOT at the crimes they where convicted of.

            reply to this | link to this | view in chronology ]

          • icon
            lucidrenegade (profile), 9 Oct 2015 @ 11:07am

            Re: Re: Re: Re: there are no surprises

            I have read that judges have instructed juries to not expect CSI type investigation, data or analysis.

            That works about as well as instructing a jury to ignore something a prosecutor/defense attorney said that they weren't supposed to.

            reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 8 Oct 2015 @ 5:46pm

    Simple Math

    It is important to note that one breathalyzer did get its software analyzed and it could not even average four numbers correctly.

    https://www.schneier.com/blog/archives/2009/05/software_proble.html

    reply to this | link to this | view in chronology ]

    • identicon
      Song from the woods, 8 Oct 2015 @ 5:50pm

      Re: Simple Math

      Its important to realize that 50% of people are below average intelligence, and cops are not allowed to have an iq above 110

      reply to this | link to this | view in chronology ]

      • identicon
        Bengie, 9 Oct 2015 @ 5:29am

        Re: Re: Simple Math

        50% of people are below average IQ, but only because of how IQ is calculated. IQ does not directly represent intelligence.

        50% of people are below median intelligence, but we can't assume intelligence has a normal distribution. The way IQ is calculated assumes a normal distribution, but that is just a trick of the math.

        reply to this | link to this | view in chronology ]

        • identicon
          Anonymous Coward, 9 Oct 2015 @ 6:29pm

          Re: Re: Re: Simple Math

          Some people just do not understand math

          reply to this | link to this | view in chronology ]

        • icon
          John Fenderson (profile), 12 Oct 2015 @ 7:16am

          Re: Re: Re: Simple Math

          Also, IQ tests don't really measure intelligence. At best, they measure one specific manifestation of intelligence, but mostly they measure how good you are at taking tests.

          They are a good predictor of how well you'll do in college, and can identify people with certain cognitive problems, but that's about it.

          reply to this | link to this | view in chronology ]

    • icon
      beltorak (profile), 9 Oct 2015 @ 12:32pm

      Re: Simple Math

      it should be possible, absent a means to examine the actual code, to submit an examination of similar software as evidence of reasonable doubt...

      in other words "let us look at the code or we will assume it is just like this pile of crap"

      reply to this | link to this | view in chronology ]

  • icon
    Charles (profile), 8 Oct 2015 @ 5:53pm

    Perhaps an open source project could be developed for DNA testing done for evidentiary purposes in civil and criminal cases. The source code could then be looked at and tested by anyone.

    Just a thought.

    reply to this | link to this | view in chronology ]

  • icon
    Spaceman Spiff (profile), 8 Oct 2015 @ 7:07pm

    Aargh!

    If you cannot have an independent expert examine the tools that have convicted you, then those tools, and their results, should be summarily rejected by the courts! DNA is notoriously difficult to analyze without extensive lab work, even today. That some software tool some years ago could do this accurately with 100% certainty is ludicrous!

    reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 8 Oct 2015 @ 7:12pm

    Why not just use an NDA?

    reply to this | link to this | view in chronology ]

    • identicon
      Anonymous Coward, 8 Oct 2015 @ 7:15pm

      Response to: Anonymous Coward on Oct 8th, 2015 @ 7:12pm

      Yeah, there are multiple ways of handling it so the defense can see it but it's not published as part of the public record.

      reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 8 Oct 2015 @ 10:06pm

    What a fascinating article! I never considered software code used to convict people should be examinable for cross examination.

    Otherwise, it's only fair to allow the defense to use their own black box software. Potentially exonerating the defendant of any wrongdoing.

    reply to this | link to this | view in chronology ]

    • identicon
      Anonymous Coward, 10 Oct 2015 @ 12:50am

      Re:

      Your Honor, Members of the jury, as you can plainly see in exhibit R, The Guilt-o-tron has given a reading of not guilty when samples of my client's blood, skin, hair, saliva and 6th grade homework were examined.

      This outcome is irrefutable, which means you can't examine the machine or the results to refute them because of reasons.

      reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 8 Oct 2015 @ 11:33pm

    The shareholder salute

    When consideration for a company's potential for profit outweighs the constitutional rights of a citizen, in a court of law, that same constitution, on which the authority of the court rests, is in effect suspended.

    That would paradoxically result in rendering the same court invalid, which may be the goal of the corporate grab for power. Its the investor-state dispute settlement model, just transmogrifed into investor-citizen dispute settlement.

    reply to this | link to this | view in chronology ]

  • identicon
    not a phucking scientist, 9 Oct 2015 @ 12:32am

    believing is not science

    if it is not peer reviewed
    the it is NOT science
    like
    "social sciences"
    "women sciences"
    "political sciences"

    or in Peter Thiel words:
    “Whenever you can’t have a debate, I often think that’s evidence that there’s a problem,” Thiel said on The Glenn Beck Program. “When people use the word ‘science,’ it’s often a tell, like in poker, that you’re bluffing. It’s like we have ‘social science’ and we have ‘political science,’ [but] we don’t call it ‘physical science’ or ‘chemical science.’ We just call them physics and chemistry because we know they’re right.”

    reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 9 Oct 2015 @ 2:00am

    xkcd

    Why all the talks about dna evidence always remind meof a xkcd strip?
    https://xkcd.com/683/

    reply to this | link to this | view in chronology ]

  • identicon
    adrian, 9 Oct 2015 @ 2:16am

    What's wrong with a pen and a piece of paper?

    I don't understand the need to inspect the software - it's just a diversionary tactic. A DNA calculation between two samples can be double checked with a pen, a piece of paper and a calculator. There are various peer reviewed statistical databases which could be used, so it would be worth asking which database is used by the software but the rest of it can be done the good old fashioned manual way.

    reply to this | link to this | view in chronology ]

    • identicon
      Anonymous Coward, 9 Oct 2015 @ 3:14am

      Re: What's wrong with a pen and a piece of paper?

      The reason for examining the software is to check whether it does what you can do with pencil and paper. Unless the defence can get crime scene samples, they cannot repeat the matching to validate the results.

      reply to this | link to this | view in chronology ]

      • identicon
        adrian, 9 Oct 2015 @ 4:20am

        Re: Re: What's wrong with a pen and a piece of paper?

        All I was saying is that if you have DNA sample from the crime scene and the DNA sample from the suspect, which the defence should have as evidence then you don't need the software to perform the calculation. You can do the DNA match calculation manually, the answer will then either confirm the computer generated result or provide sufficient doubt on the result, without having to inspect the code.

        reply to this | link to this | view in chronology ]

    • icon
      nasch (profile), 9 Oct 2015 @ 6:53am

      Re: What's wrong with a pen and a piece of paper?

      What is the need to prevent the defense from examining the software? Note that the reason given for disallowing was not "you can accomplish the same end with your own testing" it was "you might cost us money".

      reply to this | link to this | view in chronology ]

  • icon
    That Anonymous Coward (profile), 9 Oct 2015 @ 3:59am

    Perhaps the concern isn't about the lawyer taking the code and releasing their own so much as it is the absolute fear that the code is wrong.
    The legal system REALLY hates admitting it made a mistake, to the point that they will execute someone proven to be innocent.
    The system punishes those who manage to prove they were railroaded. See also: Central Park Five, West Memphis Three, McMartin Preschool, and more.
    Imagine the pressure brought to bear that if you allow this person to examine the evidence you might undo convictions of all sorts of "bad people" who were found guilty... but only because of flawed evidence.

    We treat the legal system like a game, we reward for wins only. This is why we see prosecutions declined, the case isn't a slam dunk and a loss would hurt the score. We can't indict the officer, other officers might not play nice with us any more. Oh this idiot, we can just pile up the charges and tell him hes facing life in prison but if he takes our deal... and boom conviction and another gold star on the board.

    This is no longer justice for all, it is a game and one side has all of the power and we have seen they think nothing of condemning an innocent to the ultimate penalties to keep their records secure. This broken system is what we helped build, it will be hard but we need to tear it down and go back to what it was supposed to be.

    reply to this | link to this | view in chronology ]

    • icon
      nasch (profile), 9 Oct 2015 @ 6:54am

      Re:

      I regret that I have but one insightful vote to give.

      reply to this | link to this | view in chronology ]

    • identicon
      Anonymous Coward, 9 Oct 2015 @ 8:16am

      Re:

      Are you saying that plea bargains and everything like them that short circuit the judicial process should never happen? That would mean that at current case loads, we would need at least 50x the judges, courtrooms and everything that goes along with it to match the throughput of the current streamlined moneymaking incarceration system. There is no way juries would convict petty criminals to the number of years we are seeing. We would also allow that pesky jury nullification to rear its head and for the public to weigh in on these bogus laws.

      reply to this | link to this | view in chronology ]

      • icon
        That Anonymous Coward (profile), 9 Oct 2015 @ 12:38pm

        Re: Re:

        I am saying that when plea bargains are we just kept piling on charges until he buckled, its broken.

        Man shoots man who did something bad himself.
        There is nothing wrong with offering slightly better terms to hasten this case through.
        Justice can be served.
        However if there is a reward to just stack & stack & stack charges and turn what would have been a 5 yr sentence (guesstimate) into threats of a 50 yr sentence to get him to accept 10 so the prosecutor gets a better gold star its broken.

        Do you think the public would be perfectly okay with a system where they had to openly admit we found as many charges as we could to get him to accept these terms?

        The justice system is broken, the fact that we need to have deals to keep the system from collapsing raises questions about have we gone to fing far. Are we pursuing people in a silly game of we passed these stupid laws to make stupid people feel safer, yes we are. We have panic driven laws - see also powder cocaine vs rock cocaine, where there shouldn't be a damn difference but because of nonexistant crack baby hysteria sweeping the nation there is.
        Once they have served their time, we still have these awesome systems designed to keep punishing them, so we can feel safer. We have petty tyrants at every level inflicting more misery because their targets can't fight back. Look at the Judge who threw a domestic violence survivor in jail because she got away from him and didn't want to relive it one more time. I think it was 6 prior DV charges weren't enough for him to be found guilty and the Judge got on her high horse of respect my authoritah! and into jail the victim goes.

        When the legal system denies someone the ability to prove or disprove a tool works, it is failed. You can cross examine the expert but even he can't answer how the damn thing works... but the CSI effect means people believe the machines are perfect.

        reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 9 Oct 2015 @ 5:03am

    I'll just slip this little snipplet of code in here that will make Mr Doe 100% guilty of everything ...and no one can inspect it ...because "trade secret" .

    reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 9 Oct 2015 @ 5:11am

    He should have asked to cross examine the software , since it is the real witness not the manufacturer .. the manufacturer is just a third party as far as i can tell can not make the correct equations themselves or they would have done it , instead they rely solely on software.

    reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 9 Oct 2015 @ 5:28am

    Why only ask for the software? Why not insist on access to the lab and all lab equipment & chemicals on the basis that the mere extraction of the DNA from the sample(s) hasn't been proven to be valid and correct. Request all information from the manufacturers and insist on seeing their proprietary circuit diagrams, chemical reaction information, analysis of any and all reactants used in the test. And so on. Focussing on the software smells like a wookie tactic. The software is only the LAST step in the analysis of the DNA.

    reply to this | link to this | view in chronology ]

    • identicon
      adrian, 9 Oct 2015 @ 6:31am

      Re:

      Exactly, although the software isn't even the last step, just one step in the process. I know a DNA expert witness, after the computer software they use generates a probability of a match, they then go through a manual check of the results and then their work is independently double checked. It may surprise some people but at one stage they did actually check DNA sample matches without a computer.

      reply to this | link to this | view in chronology ]

      • icon
        That Anonymous Coward (profile), 9 Oct 2015 @ 12:41pm

        Re: Re:

        but is that the method used in THIS lab. Lab standards, sadly, are anything but standard.

        And humans can make mistakes, people know and accept this... the machine is always viewed as perfect and showing it had any flaw shatters trust in its findings.

        reply to this | link to this | view in chronology ]

  • icon
    John Fenderson (profile), 9 Oct 2015 @ 7:00am

    In cases like this

    Where a company is refusing to allow inspection of source code when confidentiality is ensued, the safe assumption is that the source code is in embarrassingly bad shape.

    reply to this | link to this | view in chronology ]

    • icon
      tqk (profile), 9 Oct 2015 @ 12:33pm

      Re: In cases like this

      Where a company is refusing to allow inspection of source code when confidentiality is ensued, the safe assumption is that the source code is in embarrassingly bad shape.

      Or, just basically flawed from the get go. Intel was selling processors for years which couldn't do floating point arithmetic correctly. Microsoft's calculator app, ditto.

      This sounds a lot like the abysmally flawed voting machine debacle. However with your life potentially on the line, the response in this case is nonsensical. I wonder if in ten years we'll hear of a class action lawsuit by the wrongfully convicted (or their heirs) suing this company into oblivion.

      reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 9 Oct 2015 @ 7:49am

    We're NOT talking about depriving a man of a few years of freedom. He faces the death penalty if convicted. In this case, there's NO reason he shouldn't have access to any and all possible mitigating data, including proprietary source code. As the response noted, the court can issue a protective order to prevent dissemination.

    I also agree that the only downside of the release of the source is that the program will prove to be seriously flawed and thus cost the company customers and revenue. A program used to convict and possibly kill a man must be open to examination by the defense experts.

    reply to this | link to this | view in chronology ]

    • identicon
      Anonymous Coward, 9 Oct 2015 @ 9:04am

      Re:

      ...He faces the death penalty if convicted...

      Which means that if the death penalty is given the case is automatically appealed. Then the higher court(s) can look at the issue and possibly judge a procedural error.

      reply to this | link to this | view in chronology ]

    • icon
      tqk (profile), 9 Oct 2015 @ 12:41pm

      Re:

      A program used to convict and possibly kill a man must be open to examination by the defense experts.

      In fact, that program if used, ought to be developed and controlled and owned by the gov't for use by the justice system. NIST, maybe? No defence lawyer should have to defend against, "'Cause we said so, so there." That's the opposite of blind justice. It's divine right of kings.

      reply to this | link to this | view in chronology ]

  • icon
    John Pettitt (profile), 9 Oct 2015 @ 8:34am

    Exhibit 1 VW

    In any future case where software is in question any good defense attorney should be citing VW as an example of why software can't be trusted and companies can't be trusted to certify their own software.

    reply to this | link to this | view in chronology ]

  • identicon
    Rekrul, 9 Oct 2015 @ 9:03am

    After all, most expert witnesses representing the government have a vested interest in portraying forensic evidence as bulletproof. Without access to forensic software code, no one will ever be able to prove them wrong.

    And that right there is the real reason they won't release the code. It's not because of trade secrets, it's because they (and the government) are afraid of losing their edge.

    reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 9 Oct 2015 @ 9:43am

    A great way to send political enemies to jail. charge them, then refuse to show the "evidence" that proves their guilt.


    If you have done nothing wrong you have a lot to fear.

    reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 9 Oct 2015 @ 10:28am

    Vangelder, Camblin, TrueAllele

    In California in 2013, California ruled you couldn't have the source code for a breathalyzer.

    People v. Vangelder


    In June 2015, Mass. ordered a hearing on the reliability of a newer blood alcohol testing device.

    Commonwealth v. Kirk P. Camblin

    Defendants have argued that the source code used by the device is flawed.


    This after Judge Maureen Clancy denied discovery of TrueAllele source code for DNA analysis in Ohio in 2014.



    ...

    reply to this | link to this | view in chronology ]

  • identicon
    Anonymous Coward, 9 Oct 2015 @ 11:18am

    entire script for their software:

    If (defendant)=not_white {goto guilty}
    if (defendant)=poor {goto guilty}
    else {goto innocent}
    end if

    function Guilty()
    {call prison-slavery}

    reply to this | link to this | view in chronology ]

    • identicon
      Anonymous Coward, 9 Oct 2015 @ 6:35pm

      Re:

      Private Prison prospectus probably states exactly that.

      reply to this | link to this | view in chronology ]

    • icon
      kehvan (profile), 10 Oct 2015 @ 8:37am

      Re:

      I don't know what computer language you think you're mocking here, but some things seem amiss... A more accurate coding analogy would be...

      if (defendant!=white) {
           guilty();
      } else if (defendant==poor) {
           guilty();
      } else {
           innocent();
      }

      function guilty() {
           prison-slavery();
      }

      function prison-slavery() {
           insert code
      }

      reply to this | link to this | view in chronology ]

      • identicon
        Anonymous Coward, 11 Oct 2015 @ 1:50am

        Re: Re:

        It wasn't meant to be a certifiable actually running piece of software, more of a joke I think.....

        Next you'll be asking how many beta testers ran his code and does he have a printable list of results... :)

        reply to this | link to this | view in chronology ]

      • icon
        kehvan (profile), 11 Oct 2015 @ 8:01am

        Re: Re:

        No, really I was wondering what language you're mocking, because I didn't recognize the structure of your if-then statement from ANY language I had ever seen... so, yeah, I took your malformed and disfigured if-then statement and made if more generically reflective of what a traditional if-then statement and function calls would look like.

        reply to this | link to this | view in chronology ]

        • icon
          nasch (profile), 11 Oct 2015 @ 12:46pm

          Re: Re: Re:

          He wasn't mocking a computer language, he was mocking that manufacturer. Thus he did not write in any computer language, it was pseudocode.

          reply to this | link to this | view in chronology ]

          • icon
            kehvan (profile), 11 Oct 2015 @ 1:38pm

            Re: Re: Re: Re:

            1. "Mocking" also means "to mimic, imitate, or counterfeit."

            and 2. Pseudocode is usually modelled after something, which is what I asked originally. I modelled my counter-example on PHP.

            reply to this | link to this | view in chronology ]

            • icon
              nasch (profile), 11 Oct 2015 @ 6:11pm

              Re: Re: Re: Re: Re:

              1. "Mocking" also means "to mimic, imitate, or counterfeit."

              True.

              Pseudocode is usually modelled after something, which is what I asked originally.

              Usually, but it needn't be. The original pseudocode was perfectly valid.

              reply to this | link to this | view in chronology ]

              • icon
                kehvan (profile), 12 Oct 2015 @ 10:17am

                Re: Re: Re: Re: Re: Re:

                1. I know.

                2. "The original pseudocode was perfectly valid." -- Smh... you have a weird definition of valid. Not only does this "valid" pseudocode not match any known programming language, but just in reference to itself, it's horribly malformed an illogical... not valid.

                For example, what is this garbage;
                If (defendant)=not_white {goto guilty}
                if (defendant)=poor {goto guilty}
                else {goto innocent}
                end if

                First off, the single equal sign used in these if-then statements, such as this line, "(defendant)=not_white," aren't comparing, which is two equal signs "==", but is assigning, so that's the first invalidity in this pseudocode, that makes it less pseudocode and more a muddled mess.

                Essentially the code as written is saying that "not_white" and "poor" are assigned as the "defendant", and so this code doesn't really testing to see the defendant race and income... I had to make the assumption it was comparing and not assigning.

                Secondly, is that a nested if-then or what?

                The first if-then, I assumed, was testing the defendant color, and the second if-then tests for income level, but it's ambiguous whether or not the income level test is nested within the color test... and determining nesting is as important as specifying AND versus OR when combining two search criteria.

                Third, the final "end if" comment is irrelevant and redundant, because the closing bracket, "}", is the logical end of an if-then statement, and the second closing bracket in an if-then-else statement.

                In the end though, the real test of validity is real life... you're being racist and bombastic to claim, either by assignment or by comparison, that all "not_white" or poor people are automatically the defendant and automatically guilty.

                There's not one shred of validity, coding-wise or in reality, to any of that quote-unquote, pseudocode.

                reply to this | link to this | view in chronology ]

                • icon
                  nasch (profile), 12 Oct 2015 @ 10:53am

                  Re: Re: Re: Re: Re: Re: Re:

                  Not only does this "valid" pseudocode not match any known programming language,

                  Again, there is no need for pseudocode to match a programming language.

                  First off, the single equal sign used in these if-then statements, such as this line, "(defendant)=not_white," aren't comparing, which is two equal signs "==", but is assigning

                  You are making assumptions based on actual programming languages. Since the conventions of this pseudocode were not specified, it makes more sense to assume conventions that yield a sensible result, such as = meaning comparison.

                  Secondly, is that a nested if-then or what?

                  No, it's an if followed by a jump. If the goto executes, the next if isn't evaluated. That's generally what goto means. The "else" and "end if" aren't actually necessary, so "perfectly valid" was maybe an overstatement. But personally I didn't find it confusing, and the quality of the communication of ideas is what's important with pseudocode, not syntactic correctness - since there is no such thing.

                  In the end though, the real test of validity is real life...

                  I'll let him defend his own claim if he wants, though I suspect it was more hyperbole than a claim of fact.

                  reply to this | link to this | view in chronology ]

                  • icon
                    kehvan (profile), 12 Oct 2015 @ 12:19pm

                    Re: Re: Re: Re: Re: Re: Re: Re:

                    "Again, there is no need for pseudocode to match a programming language."

                    I do agree, but it is important for pseudocode to follow logical conventions.
                    "You are making assumptions based on actual programming languages. Since the conventions of this pseudocode were not specified, it makes more sense to assume conventions that yield a sensible result, such as = meaning comparison."

                    No, not at all. I'm trying to NOT make assumption, which is THE REASON why I asked which language it was based on. Why is this important?

                    Because many programming languages allow for the use of an assignment operator within the conditions of an if-then statement.

                    For example...
                    if (!$results=db_query("SELECT * FROM table")) {
                         Instruction for when the $results array is empty or the query failed
                    } else {
                         Instructions for displaying the $results array
                    }

                    Not only is the if-then statement testing to see if the query output is null, but it's also assigning whatever values were returned from that query to the variable $results.

                    The way that Anonymous Coward wrote that pseudocode could very well be interpreted to say one of two things... either...

                    1. If a person is not white, they're the defendant and the defendant is guilty, or...
                    2. If a defendant is not white they're guilty.

                    Now, why is this important? It's important, because it's important to me. I want to understand exactly what that Anonymous Coward meant, and not make any assumptions.

                    Another Anonymous Coward said it was a joke, but given the nature of anonymous cowards, it's hard to tell if it's the same one or someone else. Either way, the pseudocode doesn't feel like a joke, but instead is making the statement that all minorities and poor people are found guilty, and if you're rich and white, you're not... oh, and that prison is slavery. That's some very racist comments to toss around as a joke.
                    "No, it's an if followed by a jump. If the goto executes, the next if isn't evaluated. That's generally what goto means."

                    Oh come now, just admit you have no idea what I'm talking about instead of feigning understanding... the fact is the OP wrote pseudocode that suggests if a person is not white, they'll be found guilty. That's it and no conditional branch for not guilty. The next if-then suggest that if the defendant is poor they're guilty, but this time it gives the "else" statement for innocence.


                    As you've been saying, I can't take pseudocode literally and I'm not, which is exactly why I asked for clarification.

                    And given that you keep saying I can't take it literally, I can easily derive two meanings from that malformed attempt at pseudocode, and that is either...
                    1. All black defendants are found guilty, and all poor defendants are found guilty, or...
                    2. All poor black defendants are found guilty.

                    I do believe the writer literally meant #1, and that's how I rewrote my own pseudocode statement.
                    "The "else" and "end if" aren't actually necessary, so "perfectly valid" was maybe an overstatement."

                    ROFLMAO... not a programmer, are you?

                    The only people I know who use pseudocode are programmers, but whatever... the fact is, the only reason to use if-then statements in real life or in pseudocode is for conditional branching, and if you don't think a conditional branch for "innocent" is necessary, then we have a bigger argument than one over the interpretation of someone's pseudocode.
                    "But personally I didn't find it confusing, and the quality of the communication of ideas is what's important with pseudocode, not syntactic correctness - since there is no such thing."

                    What part of "uses the structural conventions of a programming language" do you not understand about pseudocode. I couldn't care less about syntax, and the proof of that is that I haven't once mentioned closing tags, now have i? But there is conceptual correctness, and that's where my focus is at, so if you're going to invoke the if-then paradigm in trying to explain something, you better fu_king well respect the concepts of nesting, assignments and comparisons when using pseudocode if-then statements.

                    But on a broader point... if you didn't find anything the OP wrote confusing and you're scolding me for requesting clarification, then I feel it's safe for me to assume you either agree with the OP or have no opinion on what he meant by his pseudocode.

                    I do have problems with it, but I want to know exactly what the OP meant, so I could better debate the point.

                    reply to this | link to this | view in chronology ]

                    • identicon
                      Anonymous Coward, 12 Oct 2015 @ 12:40pm

                      Re: Re: Re: Re: Re: Re: Re: Re: Re:

                      It was obvious what the OP meant, what wasn't obvious to you is that it was the legal system being mocked and not programming language.

                      It's clear that you think you are very clever, it's also clear (from this and other posts you've made) that you're a woeful misanthrope who doesn't play well with others.
                      Perhaps it would be better for us all if you didn't. You Bore.

                      reply to this | link to this | view in chronology ]

                      • icon
                        kehvan (profile), 12 Oct 2015 @ 1:08pm

                        Re: Re: Re: Re: Re: Re: Re: Re: Re: Re:

                        1. No... the OP wasn't clear, otherwise I wouldn't have asked.

                        2. I've already address the usage of the word "mocking".

                        And 3. as for the rest of you ad hominem... to suggest I somehow dislike humanity is ignorant, and to a large extent, hypocritical, considering you don't trust your fellow human beings enough to not act as an anonymous coward who's running around trolling people.

                        So really, not only are you a bore, but you're boorish.

                        reply to this | link to this | view in chronology ]

                        • identicon
                          Anonymous Coward, 12 Oct 2015 @ 1:27pm

                          Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re:

                          Says the dude who just won an argument by attrition.

                          reply to this | link to this | view in chronology ]

                          • identicon
                            Anonymous Coward, 12 Oct 2015 @ 1:34pm

                            Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re:

                            "Now, why is this important? It's important, because it's important to me. I want to understand exactly what that Anonymous Coward meant, and not make any assumptions."

                            How's that autism working out for you dude?

                            reply to this | link to this | view in chronology ]

                    • icon
                      nasch (profile), 12 Oct 2015 @ 1:07pm

                      Re: Re: Re: Re: Re: Re: Re: Re: Re:

                      Honestly, tl;dr. This argument isn't worth the time to read through that. If you would like to take that as a victory, feel free.

                      reply to this | link to this | view in chronology ]

                      • icon
                        kehvan (profile), 12 Oct 2015 @ 1:30pm

                        Re: Re: Re: Re: Re: Re: Re: Re: Re: Re:

                        Victory?

                        Whatever.

                        I only win if I gain something, but all I've done is waste my time explaining my request for clarification.

                        Now you can kindly go screw yourself... I won't be responding to you anymore... or any of the Anonymous Cowards on this thread.

                        reply to this | link to this | view in chronology ]

            • identicon
              Anonymous Coward, 11 Oct 2015 @ 6:25pm

              Re: Re: Re: Re: Re:

              You really know how to suck the life out of a room, you must be a hoot at parties.

              reply to this | link to this | view in chronology ]

  • icon
    lfroen (profile), 9 Oct 2015 @ 9:28pm

    Source code have nothing to do here

    As already pointed out here, DNA pattern match can be checked by pen and paper, there's no need for _ANY_ software.
    Since when forensic evidence require such a thing? Now what, for fingerprints to be valid, one must present blueprints of microscope? And how about photos/video? Will police have to present source code for camera? And in tax fraud cases lawyer will require to examine source code of Excel or calculator, because who knows what's in there?

    Of cause all this is ridiculous. Tools and software should be checked and certified before use in judgement. But, after certification, tool results are assumed to be valid, unless counterexample is presented. I don't see counterexample here; and except for this DNA sample there must be _other_ supporting evidence.

    reply to this | link to this | view in chronology ]

    • icon
      That Anonymous Coward (profile), 10 Oct 2015 @ 11:13am

      Re: Source code have nothing to do here

      The VW diesels were checked & certified, and well that didn't result in accurate results.

      If they are solely relying on the magic box, they need to pony up the source code. They need to pony up the logs & evidence the machine was used correctly and maintained properly.

      While they might be able to check it with pen and paper, there is a case where they went from the standard 1 in a kajillion chance of it being someone elses DNA to a 1 in 50 chance. Perhaps more evidence would have been required if the jury was told it was 1 in 50 and not 1 in an impossibly large number.

      There were those breathalyzers where they got caught with bad code after relying on the hidden code for so long that they went out of their way to block people convicted using the flawed data from demanding their cases be heard again.

      reply to this | link to this | view in chronology ]

      • icon
        nasch (profile), 10 Oct 2015 @ 12:11pm

        Re: Re: Source code have nothing to do here

        I think it went from 1 in 1000 to one in 30 or 40, if I'm thinking of the same case. But yeah, big difference.

        reply to this | link to this | view in chronology ]

        • icon
          lfroen (profile), 10 Oct 2015 @ 8:45pm

          Re: Re: Re: Source code have nothing to do here

          >> But yeah, big difference.
          No, not really. "1 of 1000" is very high (or low, depends how you look) number. You don't convict people in murder on such basis. Other evidence must be presented.

          reply to this | link to this | view in chronology ]

          • icon
            nasch (profile), 10 Oct 2015 @ 9:56pm

            Re: Re: Re: Re: Source code have nothing to do here

            Whether or not 1 in 1000 is enough to convict, surely you would agree that there's a big difference between 1 in 1000 and 1 in 40. To deny that there is a difference because neither should be sufficient to convict is to miss the point that this indicates a serious problem in their procedures. This was two tests done on the same evidence, and if I understand correctly, it was done in the same lab both times. There should not be such a wide variance between the two runs.

            reply to this | link to this | view in chronology ]

      • icon
        lfroen (profile), 10 Oct 2015 @ 9:06pm

        Re: Re: Source code have nothing to do here

        >> The VW diesels were checked & certified
        No they were not. They were checked by certified equipment, which _does_ produced correct result. Oh, you don't want to check equipment source code, you want one of VW? That's irrelevant. When you check suspension, you are not going to examine blueprints, right?

        reply to this | link to this | view in chronology ]

        • icon
          nasch (profile), 10 Oct 2015 @ 9:58pm

          Re: Re: Re: Source code have nothing to do here

          No they were not. They were checked by certified equipment, which _does_ produced correct result.

          So you agree they were checked. So you must be denying that they were certified. So you're saying VW was permitted to sell vehicles that had not been certified to pass emissions testing?

          (yes I know this is off topic)

          reply to this | link to this | view in chronology ]

        • icon
          That Anonymous Coward (profile), 11 Oct 2015 @ 6:06am

          Re: Re: Re: Source code have nothing to do here

          Pornoscanners were touted (still are) at being able to detect a multitude of things, they rolled out "proof" to support the claims. When people got their hands on a retired scanner they showed that they did not work as advertised.

          While there might be more ways to examine decades old DNA evidence, a good easy place to start is the source code. Having someone check code for errors is sadly cheaper than trying to run new DNA tests on small samples that the chain of care and custody could be very questionable of given the wide range.

          I have a better car example, everyone swore the brakes always worked that it was operator error. They offered up all of these tests & proof to support it. Then someone finally got the source code and showed that it was possible for it to enter a race situation that did/could (because I am still sure its being litigated somewhere) cause the accidents.

          You have someone facing a jury who hear the magic words DNA, Magic DNA machine. They all watch CSI and know that you can get results in 3 minutes and the computers are always right. It is impossible to convince people that what they know is wrong (see also AntiVax, Climate Change, etc etc.) What can raise issues is showing that the machine we've imbued with magical powers can make mistakes.
          We all know the examples of flawed CPUs, but how many of us suspect that could happen today... but I bet it has happened again.
          We all know that "sometimes" code can be wrong & still generate correct results... how often do we suspect it.
          We have heard stories of getting different results based on racial background, but we assume they account for that now... but do we bother to check?

          While it might seem silly to you to want to examine the source code, it should be the right of the accused. They didn't even offer up (mind you we have limited information) evidence of any certifications and testings done with the unit. This was well you might build your own and steal our customers as a viable answer to well the magic box said you murdered someone and you can not ask how the magic box got there.
          The code prints SSSSS on your boarding pass, but no one gets to ask how the code does that. They are just supposed to accept being treated like a lesser citizen because some coded metrics say you could be a terrorist. Pretty sure we sometimes put to much faith in the code and hold it in to much reverence rather than trust but verify it works as claimed.

          reply to this | link to this | view in chronology ]

  • icon
    That Anonymous Coward (profile), 11 Oct 2015 @ 7:49am

    One REALLY important detail that everyone seems to have missed in the discussion of well its DNA you can do it with paper.

    ""TrueAllele, created by Dr. Perlin and in its current version since 2009, is the only computer software system of its kind that interprets DNA evidence using a statistical model. It can single out individuals in a complex DNA mixture by determining how much more probable a match is versus mere coincidence. Complex mixtures can involve multiple people, as well as degraded or small DNA samples. ... "

    This could be 1 crackpots math in a black box, and he gets to use trade secrets to keep the ONLY thing doing this from being examined when its using a model of statistics not actual DNA comparison.

    http://science.slashdot.org/story/15/10/10/125249/source-code-on-trial-in-dna-matching-ca se

    reply to this | link to this | view in chronology ]

  • identicon
    Max Houck, 14 Oct 2015 @ 8:43am

    Comment on hair analysis

    The statement, "For four decades, the DOJ presented hair analysis as an unique identifier on par with fingerprints or DNA when it wasn't", is false. Microscopic hair comparisons have never been considered as a form of positive identification. If anyone testifies to that, it is false; the literature in science journals supports that. The DOJ and the FBI never taught, sanctioned, or supported that statement.

    reply to this | link to this | view in chronology ]


Add Your Comment

Have a Techdirt Account? Sign in now. Want one? Register here
Get Techdirt’s Daily Email
Use markdown for basic formatting. HTML is no longer supported.
  Save me a cookie
Follow Techdirt
Techdirt Gear
Shop Now: Techdirt Logo Gear
Advertisement
Report this ad  |  Hide Techdirt ads
Essential Reading
Techdirt Deals
Report this ad  |  Hide Techdirt ads
Techdirt Insider Chat
Advertisement
Report this ad  |  Hide Techdirt ads
Recent Stories
Advertisement
Report this ad  |  Hide Techdirt ads

Close

Email This

This feature is only available to registered users. Register or sign in to use it.