Accused Murderer Wins Right To Check Source Code of DNA Testing Kit (theregister.com) 167
"A New Jersey appeals court has ruled that a man accused of murder is entitled to review proprietary genetic testing software to challenge evidence presented against him," reports The Register.
Long-time Slashdot reader couchslug shared their report: The maker of the software, Cybergenetics, has insisted in lower court proceedings that the program's source code is a trade secret. The co-founder of the company, Mark Perlin, is said to have argued against source code analysis by claiming that the program, consisting of 170,000 lines of MATLAB code, is so dense it would take eight and a half years to review at a rate of ten lines an hour. The company offered the defense access under tightly controlled conditions outlined in a non-disclosure agreement, which included accepting a $1m liability fine in the event code details leaked. But the defense team objected to the conditions, which they argued would hinder their evaluation and would deter any expert witness from participating...
Those arguing on behalf of the defense cited past problems with other genetic testing software such as STRmix and FST (Forensic Statistical Tool). Defense expert witnesses Mats Heimdahl and Jeanna Matthews, for example, said that STRmix had 13 coding errors that affected 60 criminal cases, errors not revealed until a source code review. They also pointed out, as the appeals court ruling describes, how an FST source code review "uncovered that a 'secret function...was present in the software, tending to overestimate the likelihood of guilt.'"
EFF activists have already filed briefs in multiple courts "warning of the danger of secret software being used to convict criminal defendants," reports an EFF blog post.
"No one should be imprisoned or executed based on secret evidence that cannot be fairly evaluated for its reliability, and the ruling in this case will help prevent that injustice."
Long-time Slashdot reader couchslug shared their report: The maker of the software, Cybergenetics, has insisted in lower court proceedings that the program's source code is a trade secret. The co-founder of the company, Mark Perlin, is said to have argued against source code analysis by claiming that the program, consisting of 170,000 lines of MATLAB code, is so dense it would take eight and a half years to review at a rate of ten lines an hour. The company offered the defense access under tightly controlled conditions outlined in a non-disclosure agreement, which included accepting a $1m liability fine in the event code details leaked. But the defense team objected to the conditions, which they argued would hinder their evaluation and would deter any expert witness from participating...
Those arguing on behalf of the defense cited past problems with other genetic testing software such as STRmix and FST (Forensic Statistical Tool). Defense expert witnesses Mats Heimdahl and Jeanna Matthews, for example, said that STRmix had 13 coding errors that affected 60 criminal cases, errors not revealed until a source code review. They also pointed out, as the appeals court ruling describes, how an FST source code review "uncovered that a 'secret function...was present in the software, tending to overestimate the likelihood of guilt.'"
EFF activists have already filed briefs in multiple courts "warning of the danger of secret software being used to convict criminal defendants," reports an EFF blog post.
"No one should be imprisoned or executed based on secret evidence that cannot be fairly evaluated for its reliability, and the ruling in this case will help prevent that injustice."
Good (Score:5, Insightful)
"No one should be imprisoned or executed based on secret evidence that cannot be fairly evaluated for its reliability, and the ruling in this case will help prevent that injustice."
It doesn't matter how good your are, or how rich you are. This, is a good thing.
"...argued against source code analysis by claiming that the program, consisting of 170,000 lines of MATLAB code, is so dense it would take eight and a half years to review at a rate of ten lines an hour."
Code complexity, is a bullshit defense. Protect your assets and liability, yes. But stand by your damn code.
Re: (Score:3)
The other reason it's a bullshit defense is they're pretending the way to find flaws in code is to read through it linearly scrutinizing each line for 6 minutes. There are strategies and tools to identify problem spots, and test cases to be written.
Re: (Score:2)
Re: (Score:3)
In my limited experience, ie. what I've gathered during the years, MATLAB is surprisingly often to be what is used to start a prototype/research type program which may then become an entrenched in the product. My guess would be that MATLAB is used because it was used in studying times and therefore it's familiar. Then it's built upon, more layers upon more layers and then after twenty years it's 170,000 lines of code. Then it's a lot of effort to change the programming language.
Cybergenetics says [cybgen.com] about M
Re: (Score:2)
Im not familiar with DNA sequences, but 170k is a lot of logic, i would have thought finding patterns while complex is generally recursive and repeated attempts at finding sequences. Shouldnt the code be significantly less statements thatn 170k worth ?
Going to make another statement, but something of that size with a limited or platform not designed with structure and tooling sounds like a plain dumb choice.
Re: (Score:2)
I worked at a startup where this was exactly the case. Started by a professor who came up with a very neat solution to an image processing task. Anything algorithmically complex was written by him in MATLAB, translated to C through coder and compiled, but development continued in MATLAB. His stuff was called from some C programs on a linux box. The GUI was written in Java and ran on a laptop or tablet.
I actually inherited something similar at this startup. We were working with another researcher who gave us
Re: (Score:3, Informative)
Im not familiar with Matlab but i wonder why anyone would use such a system with its very limited developer tools (compared to say java) to write such a large app.
Depends on what you're developing. I'm going to neither defend nor criticise MATLAB per-se. What it's built for is doing numerical and/or scientific code aimed at people whose expertise is the analysis, not programming and it has a ton of tools for that. It has a lot of the analysis functions built in, and the MathWorks do care about recurring cus
Re: (Score:2)
I think you just agreed with me, if its a bad development platform then it must have quality issues.
> So if your expertise is the scientific side, not general software engineering, it fills a pretty good niche by isolating you from all those de
Re: (Score:2)
Im not familiar with Matlab but i wonder why anyone would use such a system with its very limited developer tools (compared to say java) to write such a large app. It almost sounds like the provider themelves dont know what their own code does.
Because Matlab is really good at maths, statistics, etc - which is the major part of what they are doing here. For some examples, look at a basic tutorial [mathworks.com]. There is an open source tool that is very similar in many ways - GNU Octave [gnu.org].
Re: (Score:2)
For the last fifteen years of my career my job was to take systems that no one knew how they worked and back-figure the design.
From experience I can tell you that you don't go in and examine every line.
Re: (Score:2)
Plus, the defendant shouldn't even have to show the code is flawed.
The burden of proof is on the prosecution to show that it is NOT flawed.
If the code is opaque crappy spaghetti code (as the prosecution seems to be claiming), that, in itself, should mean the conviction should be overturned.
A production system implemented in MATLAB?!!?! Oh, God.
Re:Good (Score:5, Funny)
If the code is opaque crappy spaghetti code
Lol IF hahahaha. Dude it's written in MATLAB.
Re: (Score:2)
Re: (Score:2)
The best kind of proof is making accurate predictions. If this tool can repeatedly pick a needle out of a haystack with high precision and recall, then it does work. Bottom line.
Re: (Score:2)
here's my algorithm:
for each item in haystack:
accuse item of being a needle
send item to needle jail
this will always find the needle if there is one to be found. every time. guaranteed. so, it "does work. bottom line."
who gives a shit about false-positives, anyway?
Re: (Score:2)
Re:Good (Score:4, Insightful)
A magnet will pick all the needles out of a hay stack, both guilty and innocent, and if they stop at the first needle, out of many, well it looks like they can pick the needle out of the hay stack.
So the question then becomes, how many needles are/were there in the haystack.
One way to test, and how it is done in my country for things like blood tests for DUI, is to let the defendant use a different independent testing agency.
Re: (Score:2)
A magnet will pick all the needles out of a hay stack, both guilty and innocent, and if they stop at the first needle, out of many, well it looks like they can pick the needle out of the hay stack.
And how does a casual observer tell the difference between needles and hay? If the badly coded magnet also collects hay it looks like it's working too.
Re: (Score:2)
Re:Good (Score:5, Insightful)
Code complexity, is a bullshit defense. Protect your assets and liability, yes. But stand by your damn code.
Someone telling me their code is complex and dense and thus hard to understand is basically telling me that it is almost certainly unreliable and noone really knows what it's doing in a lot of places
Good lord... here it is again. (Score:2)
"Someone telling me their code is complex and dense and thus hard to understand is basically telling me that it is almost certainly unreliable and noone really knows what it's doing in a lot of places."
You realize this is the same stupid opinion used to invalidate everything scientific by those unqualified to judge it, right?
"If I can't understand it, it must not be rightt." Well, what if the code is really complex, dense, and hard because the problem itself is complex, dense, and hard?
Yeah, yeah, yeah. The
Re: (Score:2)
"If I can't understand it, it must not be rightt."
No. That is not what the GPP said at all.
If a scientist claims that X is true, saying that I don't understand it is not a valid counter-argument.
Showing that THE SCIENTIST doesn't understand it, is certainly a valid counter-argument.
If the people who designed the device don't understand how it works, then we shouldn't be using it to send people to prison.
Re: (Score:2)
You do what this defense would do, get an expert to look at it. Your bogus argument hinges on the actual defendant doing the examining. He wouldn't be. It would be an expert.
Re: (Score:2)
Re: (Score:3)
Code complexity, is a bullshit defense. Protect your assets and liability, yes. But stand by your damn code.
The question is how the f*ck did they manage to reach such code complexity in the first place. Genetic analysis is Bayes. In fact, it was the first major applicaiton of Bayes statistics before insurance, financial markets and everything else. It is not particularly complex and I would really like to know how can one manage to belch 78K lines of Matlab on what is generally expressed in less than 6K (including data handling).
Further to this, the tool should be inadmissible as a tool to start off with. In th
Matlab code is usually complex (Score:4, Interesting)
Well, it is Matlab after all. Matlab discourages you to refactor code, both because of a performance penalty with function calls and because each function needs to be in a separate file, sharing the same namespace with thousands of built-in functions.
The Matlab language and builtin functions have bizarre ways of handling both edge cases and reasonable use cases:
* Spaces act as commas: [1, 2 -3 + 4] is a 3-element array.
* If a function returns an array, you cannot index the return value without an temporary variable: tmp=foofunc(123); a=tmp(1). Not a=foofunc(123)(1).
* The character ' is both a string delimiter and a unary conjugate transpose operation.
* i and j are predefined as the imaginary constant, but you can use them as a loop variable. Printing a complex variable can result in '1 + i', but don't try to copy that into a variable initialization.
* If d=[2, 3, 4] then zeros(d) will initialize a 3D array. But if d=2 or d=[2], then you'll get a 2x2 array. Code that needs to handle input data with unknown dimensions must be sprinkled with if statements to check whether it is the special case of a 1D array.
* A string a='xyz\n' has a backslash. Only some string functions will convert it to newline.
* Operations involving floats and ints will demote the result to the lower precision datatype. int8(25)*10.0 == 127 WTF.
* Functions fail silently. If you save data in a format that allows 2 GB per record and you try more, you end up with files having missing records and a warning on stdout. Great fun in batch jobs that take 24 hours to finish and that work fine with small test datasets. The workaround made me run into the next one.
* The only built-in way to write data to a file can only write records with unique names matching variables in the present scope. There is no way to write an array row to a file without first assigning it to a variable.
* Though Matlab is prized for vectorized handling of array data, it cannot do linear operations on high-dimensional arrays; where in numpy you'd use c=np.einsum('ijkl,kij->il', a, b), you'll spend 5 lines of code reordering the array layout or writing out an explicit nested loop.
I sometimes have to deal with legacy Matlab code and each and every time I'm screaming "WTF you can't be serious" as I discover another quirk like these.
Re: (Score:2)
Re:Matlab code is usually complex (Score:5, Insightful)
Matlab has some quirks, not gonna lie.
The main problem though is that the code authors are generally experts in their domain of something scientific or engineering related, not programming. Mostly it's used by people who self-taught or had a couple of courses then self taught the all they needed to know to do their science, which can get quite large and complex. Such people generally don't have any professional software experience from the point of view of software engineering.
If you do have the right background then you can write good matlab code. It's very rare to find someone who can (a) write actual production quality code and (b) has the experience and knowledge to do so to solve some niche scientific problem. And those people are often extremely expensive because it's such a rare combined skillset.
Re: (Score:2)
Can't the defence have their own DNA analysis done?
In the UK people convicted on fingerprint evidence have been subsequently found innocent by getting their own experts to do an analysis saying they do not match, contradicting the prosecution's expert.
Re: (Score:2)
Can't the defence have their own DNA analysis done?
In the UK people convicted on fingerprint evidence have been subsequently found innocent by getting their own experts to do an analysis saying they do not match, contradicting the prosecution's expert.
Matching DNA if you have a good quality sample is easy, and gives a high quality match. Of course, there are always weird counterexamples [scientificamerican.com]. Tools like this are used when you don't have high quality samples - and you end up with a lot of statistics for making things more or less likely. An interesting article on this [nih.gov].
Re: (Score:2)
Cases have collapsed here because of dodgy techniques like "amplification" producing results that are unreliable. In one case it emerged in court that the match pointed to at least 5 other people in the police database alone.
Re: (Score:3)
you have no idea do you, of the levels of shittiness in code, both in quantity and quality
i have seen a single function that is on the order of tens of thousands of lines of code
which does "serialize object to xml"
by hand.
and what about 3rd party code they can't give out (Score:2)
and what about 3rd party code they can't give out they have there own NDA that says they can't do this.
Re: (Score:3)
Then the prosecution shouldn’t have built their case on evidence that is either inadmissible or indefensible. Relying on such software is like relying on a process for managing evidence that doesn’t maintain a proper chain of custody: you just don’t do it.
Re: (Score:3)
All NDAs I've ever seen include a clause allowing for the release of information if a court orders it. There probably is another clause that requires the third party to be notified and allowing that third party to make its own case to the court that they shouldn't have to release it. But the courts definitely override NDAs.
Re: (Score:3)
Then that evidence is thrown OUT. If you sign such an NDA and insist on keeping it, you are declaring that your software will NEVER be used in a court of law.
The rights of the accused outweigh ALL business interests. You can not say "I have proof that he is guilty, but I can't show it to you because I signed an agreement."
The Constition says you can confront your accuser/witnesses. This obviously includes software that analyzes evidence, otherwise corrupt people will write some software for DNA tests w
self driving cars will be an issue unless owners c (Score:2)
self driving cars will be an issue unless owners can not face any lawsuits or jail / prison time for the car doing something.
Re: and what about 3rd party code they can't give (Score:2)
Re: (Score:2)
Countries can override IP laws in other countries. Here in Canada, we've considered the nuclear option of just declaring all American IP law null and void every time America screws us in favour of China, like Trump especially liked to do.
Re: (Score:2)
and what about 3rd party code that the company only uses as a closed-source library? Especially if the owner of that library is a Chinese (for example) company that will happily show the middle finger in response to any demand to open its code from an American court?
Re: (Score:3)
That's fine. Then any results from it can be ruled as inadmissible in court.
Self driving cars need this in case there is an (Score:2)
Self driving cars need this in case there is an criminal case and the car owner / driver?? is facing changes.
Also the Uber self-driving car operator Needs to demand ALL source code, logs, even the code of the car that uber modded.
How is this a software problem? (Score:5, Insightful)
Surely the issues are with the sample collection, preparation, and chemical analysis.
What comes out is a pile of data. The defense should be able to check that with different software. If the results don't match then why needs to be investigated.
But it is a concern when a vast amount of MatLab code is required that would take "eight and a half years" to review -- how long did it take to write? And how do you unit test such code?
Or better, test the entire system. Put lots of samples from one end to the other, with the source not known to the testers, and see how they go.
Re: (Score:3)
That's probably a good approach, but the source code of any software used to make decisions by the justice system should be open to examination. Period. No excuses. And no reason needing to be given.
Re: (Score:2)
That makes a lot of sense if you're looking for the truth, that's not the point of an adversarial legal system
i.e., The defense wouldn't want to actually do that because, more likely than not, the test cases would work properly (even if there is some issue with the code, which I suspect there is, especially considering how seemingly unnecessarily voluminous it is, i
Re: (Score:2)
This could be a software problem because DNA sequencing requires assembling and matching of smaller DNA sequences. To a certain degree DNA matching is a probabilistic task, and different software can yield different results.
Re: (Score:3)
Not to mention the model which looks at the assembled matches and generates a likelihood ratio for the two samples being from the same person vs not. I once worked on a project to implement a frontend for some fingerprint matching analysis written by an academic: it was impossible to test properly because the backend analysis had a Monte Carlo component and the LR could vary by an order of magnitude between submissions of identical data. As you might imagine, I'm now rather cynical about forensic science.
In DUI case that this happened & the code suck (Score:2)
In DUI case that this happened & the code sucked with an lot of bugs and I think they got off.
Outrageous (Score:5, Insightful)
The notion that a decision to put someone away, possibly executed, for murder, being entrusted to closed-source software is outrageous on the face of it. And the defense that 'it's a trade secret' is doubly outrageous.
Re: Outrageous (Score:3)
Q: "how do you know they are guilty?"
A: "I just know".
Well case closed I guess. No arguing that.
How is the code certified / qualified (Score:3)
What other software should be made public? (Score:2, Troll)
Let's see. Oh, hell, I don't know, perhaps every voting machine? Not just the source code but also a complete build procedure including the entire toolchain so you could compare the binary with what's actually installed on the machine to ensure that there are no shenanigans going on. What? You don't like that? You think everyone should just trust it? What are you afraid of? If you have nothing to hide, you should have no problem with this. That goes for hockey stick boy's climate data and methodologie
Re: (Score:3)
Let's see. Oh, hell, I don't know, perhaps every voting machine?
What if they could verify the actual results not the software.
Which is exactly what happened with voting. They were verified by hand recounts. Now fuck off.
abysmal injustice (Score:4, Insightful)
The co-founder of the company [Cybergenetics], Mark Perlin, is said to have argued against source code analysis by claiming that the program, consisting of 170,000 lines of MATLAB code, is so dense it would take eight and a half years to review at a rate of ten lines an hour.
Consider the complete idiocy of convicting someone according to evidence generated by software too complex to be audited for correctness.
The company president's own statements are admission that there can be no confidence in the correctness of its software.
That alone should result in exoneration. "Should result", not in the sense that it is an expected outcome according to conventional judicial practice, but in the hypothetical sense of if the courts actually worked according to justice and reason.
170,000 lines of MATLAB code (Score:4, Insightful)
!?
That much Matlab code means it's probably cobbled together bits and pieces from a hundred grad students who just needed something to work so they could write their thesis.
I once worked in a lab where the prof owned a company that did clinical trials. He was showing me some results from a particular trial and saying "the therapy is great! Every single patient got better!"
*Every* patient? 100% response is a giant red flag. A little digging turned up two things: 1) the code they were using for the analysis was written by a student learning to program, and 2) it discarded the sign of all the results. I recommended that they hire actual software engineers and developers to write, and test, their critical software, rather than using some grad student's learn-to-code-in-Perl project.
Re: (Score:3)
Re: (Score:2)
Why trade secrets? (Score:2)
Why not rely on copyright to protect their code? Are they worried about pirated versions being bought by *checks notes* the government? Is DNA testing some mysterous science? Or is it far more likely that they don't want to release the code because it is either shit or actually infringes on someone else's IP.
Seem obvious (Score:5, Interesting)
If you can't check the code then the defendant can present his own proprietary codebase that the prosecution isn't allowed to see which completely exonerates him.
Re: (Score:2)
Either it is evidence, OR it is secret. (Score:2)
It cannot be both.
The whole definition of evidence is, that you reveal information that was previously not known or secret.
Any self-respecting judge should ask ...
Judge: This just says 'guilty'. Can you back up how you came to that conclusion?
Company: We do not share our code with you!
Judge: Then you ain't got evidence!
Company: But you are supposed to trust our judgement!
Judge: *Who's the judge here, motherfuckers?!*
Re: (Score:2)
Has Sam L. Jackson ever played a judge?
I want to see that.
I mean a courtroom judge, not like Dread or something.
170000 LOC?!? (Score:2)
Re: (Score:2)
The other 169000 loc are for the UI.
Get a 2nd opinion (Score:2)
If you don't like the results of a DNA test, just get a second opinion. In fact, for criminal cases, there should be automatic multiple tests from different services. The services can be selected at random from a list of same, including services in Europe and China. Double-blind tests and best 3-out-of-4 wins.
Russian matlab code (Score:2)
Re:Exonerations (Score:5, Insightful)
While it's nice to want to audit the code, this code also represents someone's years of research...
If it was research, the surely it's been published and this is not an issue. If it hasn't been published then it's worthless as research anyway since you can't review it.
Re:Exonerations (Score:5, Insightful)
Guilty criminals may have been falsely exonerated by faulty software too. The difference is the victims or prosecutors didn't have good enough lawyers or sufficient motive to properly audit the software. Someone who is facing a significant sentence is obviously more motivated.
Either way, lives are at stake and the methods used for evidence collection should be transparent.
100 guilty Persons should escape than that one inn (Score:2)
That it is better 100 guilty Persons should escape than that one innocent Person should suffer
Re: (Score:3)
That it is better 100 guilty Persons should escape than that one innocent Person should suffer
He also has the right to a speedy and fair trial. If the deciding evidence is DNA analysis that will take 8 years to audit, it shouldn't even be admissible.
Re: 100 guilty Persons should escape than that one (Score:2)
Does the entire case hinge on only the DNA results? So much for preponderance of evidence and beyond reasonable doubt. DNA, at best, only proves the blood or skin cells belonged to you. They do not tell the story of how or when it got there.
Re: 100 guilty Persons should escape than that one (Score:2)
But if 101 guilty suffer.... Have to draw the line somewhere
Re: (Score:3, Informative)
That it is better 100 guilty Persons should escape than that one innocent Person should suffer
No, that is nonsense. Our system of justice doesn't work that way.
When DNA evidence first became viable, the Innocence Project did an analysis of archived evidence used to convict and incarcerate defendants. They found that in more than 10% of cases, the defendant couldn't possibly have committed the crime they went to prison for, and in an additional 10%, the DNA evidence indicated likely innocence.
There is no reason to believe we are any more accurate today, or in crimes that don't involve DNA evidence.
Re: (Score:2)
That it is better 100 guilty Persons should escape than that one innocent Person should suffer
No, that is nonsense. Our system of justice doesn't work that way.
When DNA evidence first became viable, the Innocence Project did an analysis of archived evidence used to convict and incarcerate defendants. They found that in more than 10% of cases, the defendant couldn't possibly have committed the crime they went to prison for, and in an additional 10%, the DNA evidence indicated likely innocence.
There is no reason to believe we are any more accurate today, or in crimes that don't involve DNA evidence.
The false conviction rate is WAY higher than 1%. It is likely between 10% and 20%.
What you are looking at is a severely biased sample, so I don't think you can use this to extrapolate the false conviction rate.
Re: 100 guilty Persons should escape than that one (Score:2)
They also argued that DNA testing, at the time, was better suited to prove a negative rather than a positive. Meaning that when the dna isnt even close, you definitely couldnt have done this. Whereas a match could be off enough that relatives would also match since my match it was not a 100% exact copy. However, this was during the OJ trials. I cannot attest to modern collection and extraction methods. Analysis and conclusions might be software, but at some level you have to start digging into the alleles u
Re: Exonerations (Score:3)
Re: (Score:2)
by the government which will not use pirated copies
Yep, that would definitely never happen. [slashdot.org]
Re: (Score:2)
Reading that article, do you think it was intentional piracy? Because I read a licensing fuck-up and a commitment to fix it going forward.
Re: Exonerations (Score:5, Insightful)
Re:Exonerations (Score:5, Insightful)
The government is free to either buy the right to "open" the software if the author/copyright owner is amenable.
The government is also free to develop their own open source software to perform the necessary functions.
The government, however, is not free to use black boxes that are not open to audit (and that even the government themselves have not rigorously audited) to convict criminal defendants.
Re: (Score:2)
If you haven't seen the code, that bolded word is pulled out yer ass.
Re: (Score:3)
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
> Which one are we using today?
A few things.
1. Tens of Millions of vaccines have been given out with a few people having adverse effects. The "adverse effect" rate is at least 3 orders of magnitude lower.
2. The vaccine's adverse effects are mild
3. Tens of Millions of vaccines have a positive effect on people's lives by reducing their infection rate, the sever
Re: (Score:2)
Sorry, but anaphylactic shock is not mild. The rest of your argument is correct.
OTOH, if you watch for it and are ready to treat it, then anaphylactic shock can easily be handled. But it's not mild, it's often (usually?) life threatening.
Re: (Score:2)
First, the government in the US is not requiring anyone to take vaccines. In some cases a failure to take, for example, a measles vaccine may preclude your kid from going to a public school, but not a private one or to receive an education via homeschooling. A criminal defendant, on the other hand, has no choice but to submit to the analysis of this software.
Second, the standard for criminal convictions is very high: a unanimous jury decision of "guilty (beyond reasonable doubt)" and the defendant is guaran
Bad argument [Re:Which logic are we using today?] (Score:2)
Thousands and thousands of samples have gone through this system and 60 were found to have issues
Bad argument. You state thousands and thousands of samples have gone through the system, but the question is not how many samples were processed, it is how many of these samples were were checked by independent methods.
which means we should throw out every case involving the software.
Bad argument. The conclusion was we should be able to examine the software, not "we should throw out every case".
Thousands and thousands of vaccines have been given and a few people had adverse reactions which means we should stop using the vaccine.
Bad argument. The conclusion was for examination, not "we should stop using it."
Which one are we using today?
If the software is used to produce evidence in court, the defense should be allowed to examine i
Re: (Score:2)
Wow, you can't tell the difference between that? You are that devoted to bad logic?
Getting a rash counts as a reaction. 60 reactions out of many thousands is minute and unimportant. If there were 60 deaths, then that would be worth talking apart, despite the vaccine saving more than 60 lives.
But 60 cases thrown out of many thousands is HUGE. Barely one in a thousand make appeals, and often the judges ignore evidence of wrong doing. 60 cases being thrown out is a tidal wave.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
More people are dying now than at any point when trump was in charge. What happened?
ManGod Trump is not going from bed to bed giving 5D CPR anymore.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
As well you shouldn't.
Re: (Score:3)
I don't think that's true.
If the defense requests an extension of eight and a half years to review the code they may have to waive their right to a speedy trial in order for the court to grant the request, but there's nothing illegal about making the request or the court granting the request.
Re: (Score:2)
the PD only has 15 min an hour for your case and they can do 10 lines in that time.
Re: (Score:2)