That would be fair enough, except for this: The FBI has tried to block distribution of the Arizona results and is blocking people from performing similar searches using CODIS. If this is correct, then the FBI does not have confidence in its own claims, which is worrisome. If they had confidence, they would have no need to block distribution of anything, indeed they would increase confidence in the system by having others validate that the probability of a false match is staggeringly low. I do not trust those who will not trust their own integrity.
You are probably correct, but fundamentally there are several questions here that that does not answer: First, is it ethical - or even legal - to withhold evidence from the defense? Essentially, if the defense cannot question the validity of the prosecution's match, then that is what the FBI is doing.
Secondly, if the frequency of matches is such as to raise questions, should the FBI be gathering more detailed genetic information? If you do the birthday problem again, but this time instead of matching the same day, match the same hour, you get much lower probabilities of a match. Still high, though. So you go to the same minute. Probabilities drop further. You can keep going, as far as you like, until you run out of data or you reach an acceptable precision.
Thirdly, should the way in which the evidence is presented be changed? Juries are going to be swayed by "experts" that tout unimaginably low probabilities of a wrong result, especially in horrific cases when the jury WANTS someone to be punished, but if that claim is invalid, then juries should be presented by something far more honest.
Lastly, if the testing method is at fault, then more stringent testing should be insisted upon for all DNA tests ordered by, or on behalf of, appeals. If the trial was in doubt but the trial presented DNA evidence at a thousand markers, then go ahead and test a million. This should also be considered for all death penalty cases where the convicted person claimed innocence up to and including being punished, but DNA evidence seemed to back the conviction.
If we are to have confidence in the system, the system must ALSO have confidence in the system, and that includes the real experts who understand the statistics, the experimental errors and the uncertainties involved.
Hell, it's hard enough to get a final price for a hamburger in the US. Prices are invariably given minus sales tax, where sales tax (and what sales tax applies to) varying from State to State. Maybe it's some sort of Uncertainty Principle: The more certainty you have regarding quantity of hassle you'll get in trying to buy something, the less certainty you have regarding the cost.
...absolutely superb novels that accurately depict the different dialects in England. That has nothing to do with skin colour, origin, or ethnicity. Since it is typically the author's own dialect that is depicted accurately, it is very likely not prejudice against that region, either. Oh, certainly, I'd question some use of dialect, but even someone as prejudice as Enid Blyton did not use them maliciously. Oh, some American authors might, but even some of the best-known examples of American authors who have used dialects in writing have done so to be historically or geographically correct, or to highlight prejudice, not to exacerbate it.
Well, I'd argue it's not the only reason death is a bad idea. Ghosts don't do a very good job at restitution, for starters, and that's ignoring all the ethical arguments. But you're absolutely correct that whilst the probability of error is indeterminate (well, nobody's being allowed to validate the database) and quite possibly excessive (there are currently one million SNPs tested at deCODEme, not sure how many SNPs or markers are looked at by the FBI - assuming they're that sophisticated and aren't just using the archaic chromatographic techniques for producing rather dull and uninformative photographic images).
Now, one might argue that if the FBI started doing a thorough analysis of DNA, it could do all kinds of nasty things to privacy. True enough, if they store the raw data. However, that would be insanely dumb and they wouldn't do that, would they...? Far more efficient to store multiple cryptographic hashes (eg: SHA512, Whirlpool and Tiger), so that the raw values are not present but the ability to match with no additional risk of aliasing is still present.
The other question is: since the database was started, our ability to study DNA, examine genetic markers and recognize which areas are better for identifying individuals has drastically improved. Has the database improved to match this knowledge, or are we relying on outmoded methods? If the former, then it's as good as it is going to get - for now. If the latter, the probability of a false positive could be massively slashed. That would surely be desirable for a crime-fighting system... wouldn't it?
The software guys in the team have some neat utilities for analyzing and generating parallel code - but it's old and clearly not maintained. Apart from the stuff that has become commercial. Either way, a good workman keeps their toolset in good condition, so the lack of maintenance does bother me.
Since all identifying documents are time-sensitive, it might be best to express it as dollars per year. How long does a New Zealand passport last, these days?
Easy way to test that theory. Egotistical managers who don't give a damn about the company are usually former Intel employees or former IBMers. Anyone know if Childs' boss was ever employed by either of those?
It depends on flight risk, the probability he has implanted mind-control software that uses the router lights to hypnotize people, the financial status of the county, and whether or not there's a league table for bail collected that the judge wants to top.
Be fair. Psychologists have been pointing out for some time that the same traits that make for good managers and administrators are listed in psychology texts as the traits of full-blown schizophrenics. I'm serious.
Yes, he may very well have bordered on "pathological criminal behaviour", but that is what is expected of employees. It is necessary to exhibit exactly that behaviour if you wish to successfully rise through the ranks.
I dislike that, and believe it is one incredibly unhealthy attitude - tied utterly to America's Puritan "Work Ethic". However, that is neither here nor there. The guy is a product of such attitudes. If he is a monster, then Herr Frankenstein bears the greater responsibility.
That does not make him blameless. It means that he is culpable only to some degree below 100%. Punish him for that percentage he is culpable. Fine. But it means there is also a non-zero component of responsibility elsewhere, which should not go unpunished, and that there is a non-zero element of illness the guy has developed as a result, for which he should be treated.
In any such system, blaming one person is extremely easy but utterly futile. It doesn't fix the underlying problems which made the failure possible, and the ultimate problem is invariably the mental illness prevalent in modern management methods.
You are correct for how it works right now, but the complaint was that all CAs should be considered equal, that a self-rolled cert should not throw up warnings. Your argument is entirely correct provided the web of trust is maintained. As soon as self-signed certs are acceptable, provided you can find an unpatched DNS server that would allow you to re-target all packets destined for the bank to the proxy, you could snoop on traffic to the bank without anyone being the wiser.
Now, I accept CAs don't always do a good job of validating who someone is. I also accept that they charge 5-10 times as much (per year) than the passport office, for what is basically the same thing. These are not the fault of the web of trust, but a fault in the way the industry is operated and regulated. These things won't be fixed by breaking the web of trust.
All possible attacks against certificates are purely hypothetical at this time. These would include:
A poor, seeded PRNG being used where the seed is somehow exposed or part of the key - such as a simple hashed value of the same information that is made public, where the PRNG algorithm can be determined and reproduced in some way
Someone has figured out a solution to the factoring problem, breaking RSA
The effective key length is so short that the private key can be brute-forced
There are also two attacks against infrastructure which can compromise a key:
The machine generating the key pair has been compromised in advance, with private keys intercepted and copied elsewhere
Any machine subsequently storing the private key has been compromised, allowing the private key to be stolen
Of all of these, the last one is the only one anyone needs to take seriously. Even then, there are plenty of ways of making directories and files very secure, and making sure that potential exploits like buffer overflows are blocked in advance. (Just use a malloc replacement that prevents them.) The other attacks are so improbable that you can ignore them.
This leave one other attack vector:
Social Engineering
This, according to reports, was used to obtain Microsoft's private keys from Verisign. Most reputable cert vendors have established better practices now. Simply choose one that will only deliver keys to an authorized contact point and only after a call-back check or some other authentication scheme.
The summary says it's 4th generation. Most bittorrent clients are in C or Java, not Forth. (Ok, that was bad. But even so, I'd call bittorrent a first generation true peer-to-peer protocol, same with Gnutella and ED2K. Freenet, Tor and X-Bone might be considered second generation, and that's a push, but there frankly isn't anything out there I'd consider third generation, never mind fourth. You need to have something that is significantly different to qualify for a whole generation in tech jargon, which is why even the latest multi-core multi-way hybrid architecture CPUs are not considered to be running millionth-generation languages.)
Let's start with a Man-in-the-Middle attack. Attacker finds an unpatched DNS and points www.somebank.com to their proxy that has SSL support. A user connects, thinking it is their bank. It looks like it, because it really is the bank's website that is being displayed, and the URL is correct. The user enters their account login information, because it's a secure site. The proxy, of course, decrypts the inbound user SSL traffic, stores username/password information, re-encrypts using the bank's SSL session and forwards to the bank. The bank never knows it's not the user - it's encrypted, after all, and it is all correct.
The idea of certificates is to authenticate the connection, make it impossible to someone in the middle to pretend to be the server to the client, and the client to the server. Actually, it would be better to require users to have certificates as well, in many cases, as passwords tend to be too trivial.
Now, the price of certificates is horrendous. The passport office provides a document as good, or better, than many certificates, but it doesn't cost many hundreds of dollars to obtain a passport. In fact, as digital certificates are essentially the same as a passport with electronic information, it might be better if the passport office issued digital certificates along with physical passports as a combined package. The added cost to them would be practically nil, and the certificates would have a much greater credibility level than those by most corporations, at least for personal certs.
Heh. Actually, it's through science fiction that formed the basis of my ideas. (Books: weirdstone of brisingamen, moon of gomrath, eight keys to eden, starship medic, caves of steel, stars like dust, citizen of the galaxy; Short stories: The Tales of the Hexadecimal Kid; Television: doctor who, blake's 7, sapphire and steel, space: 1999, children of the stone, the tomorrow people.) Since then, I've explored different avenues of my own (nothing terribly original) but have basically come to the same premise you described, which is that of classic sci-fi.
This is why I am firmly convinced that society cannot develop lopsidedly - if technology is progressing rapidly, then those individuals who govern the technology should not be permitted to stagnate. As I see it, this loosely divides up into ethics (how you deal with what you have got) and artistic expression (how you deal with you dealing with what you've got). Ethics evolves by questioning old ideas that are no longer valid, no matter how cherished they may be. Expression evolves by safeguarding such expression but also by encouraging people to leave "safe" formulaic art and venture into new ground. That would give society a lot more of a voice, but also a more effective safety-valve. Change produces tensions, but tensions can be grounded.
If the development of science, technology and society was kept in a dynamic equilibrium, with whichever one is faster pulling the others to where they should be, then a 1984 scenario becomes impossible. 1984-type dystopias are only possible if one of those three prongs exceeds the ability of the other two to keep it in balance. If society exceeds science or technology by too much, society's natural fears will cripple the other two and balance becomes impossible. If science exceeds society or technology, then what you understand exceeds what you can control and paranoia results. The cult of Pythagoras is a good example. If technology is king, then what you can do exceeds what you can control or understand, and the results are often self-destructive.
Well, if you're doing something like that, you would probably start with LynxOS or one of the carrier-grade Linux distros, remove absolutely all unnecessary kernel options, replace all software (including the init code) with the smallest and fastest code you can find, and place a lot more emphasis on tuning of that kind from a known highly stable start point.
You were trying to read my mind, weren't you? Well, it won't work. My tinfoil hat is the latest deluxe model. The nice gentleman from the NSA told me it was even checked for listening devices beforehand.
Grid computing has largely fixed the limitations on protocols and authentication, making "cloud computing" look, well, just a little bit wet. (There is now a very nice SASL layer for Globus, for example.) The main problem I have with the term "cloud computing" is that the term "cloud" already has a well-defined meaning in computer networks - it's any topology where you don't care about where things are or what things look like, you pass packets in and you get packets out. To use programming jargon, it is a black box. The internals don't matter. This fits with the idea of grid computing, but does NOT fit with the idea of a single, centralized off-site system. If you have to care about it being centralized somewhere, then it is not black box, because you DO care about what is on the inside.
Slashdot Mirror
That would be fair enough, except for this: The FBI has tried to block distribution of the Arizona results and is blocking people from performing similar searches using CODIS. If this is correct, then the FBI does not have confidence in its own claims, which is worrisome. If they had confidence, they would have no need to block distribution of anything, indeed they would increase confidence in the system by having others validate that the probability of a false match is staggeringly low. I do not trust those who will not trust their own integrity.
You are probably correct, but fundamentally there are several questions here that that does not answer: First, is it ethical - or even legal - to withhold evidence from the defense? Essentially, if the defense cannot question the validity of the prosecution's match, then that is what the FBI is doing.
Secondly, if the frequency of matches is such as to raise questions, should the FBI be gathering more detailed genetic information? If you do the birthday problem again, but this time instead of matching the same day, match the same hour, you get much lower probabilities of a match. Still high, though. So you go to the same minute. Probabilities drop further. You can keep going, as far as you like, until you run out of data or you reach an acceptable precision.
Thirdly, should the way in which the evidence is presented be changed? Juries are going to be swayed by "experts" that tout unimaginably low probabilities of a wrong result, especially in horrific cases when the jury WANTS someone to be punished, but if that claim is invalid, then juries should be presented by something far more honest.
Lastly, if the testing method is at fault, then more stringent testing should be insisted upon for all DNA tests ordered by, or on behalf of, appeals. If the trial was in doubt but the trial presented DNA evidence at a thousand markers, then go ahead and test a million. This should also be considered for all death penalty cases where the convicted person claimed innocence up to and including being punished, but DNA evidence seemed to back the conviction.
If we are to have confidence in the system, the system must ALSO have confidence in the system, and that includes the real experts who understand the statistics, the experimental errors and the uncertainties involved.
Hell, it's hard enough to get a final price for a hamburger in the US. Prices are invariably given minus sales tax, where sales tax (and what sales tax applies to) varying from State to State. Maybe it's some sort of Uncertainty Principle: The more certainty you have regarding quantity of hassle you'll get in trying to buy something, the less certainty you have regarding the cost.
...absolutely superb novels that accurately depict the different dialects in England. That has nothing to do with skin colour, origin, or ethnicity. Since it is typically the author's own dialect that is depicted accurately, it is very likely not prejudice against that region, either. Oh, certainly, I'd question some use of dialect, but even someone as prejudice as Enid Blyton did not use them maliciously. Oh, some American authors might, but even some of the best-known examples of American authors who have used dialects in writing have done so to be historically or geographically correct, or to highlight prejudice, not to exacerbate it.
Using the Older or Younger Futhark runic system?
I can see it now. Sparticus II: "We Are Sarcasticus!"
Now, one might argue that if the FBI started doing a thorough analysis of DNA, it could do all kinds of nasty things to privacy. True enough, if they store the raw data. However, that would be insanely dumb and they wouldn't do that, would they...? Far more efficient to store multiple cryptographic hashes (eg: SHA512, Whirlpool and Tiger), so that the raw values are not present but the ability to match with no additional risk of aliasing is still present.
The other question is: since the database was started, our ability to study DNA, examine genetic markers and recognize which areas are better for identifying individuals has drastically improved. Has the database improved to match this knowledge, or are we relying on outmoded methods? If the former, then it's as good as it is going to get - for now. If the latter, the probability of a false positive could be massively slashed. That would surely be desirable for a crime-fighting system... wouldn't it?
The software guys in the team have some neat utilities for analyzing and generating parallel code - but it's old and clearly not maintained. Apart from the stuff that has become commercial. Either way, a good workman keeps their toolset in good condition, so the lack of maintenance does bother me.
Since all identifying documents are time-sensitive, it might be best to express it as dollars per year. How long does a New Zealand passport last, these days?
Easy way to test that theory. Egotistical managers who don't give a damn about the company are usually former Intel employees or former IBMers. Anyone know if Childs' boss was ever employed by either of those?
It depends on flight risk, the probability he has implanted mind-control software that uses the router lights to hypnotize people, the financial status of the county, and whether or not there's a league table for bail collected that the judge wants to top.
Yes, he may very well have bordered on "pathological criminal behaviour", but that is what is expected of employees. It is necessary to exhibit exactly that behaviour if you wish to successfully rise through the ranks.
I dislike that, and believe it is one incredibly unhealthy attitude - tied utterly to America's Puritan "Work Ethic". However, that is neither here nor there. The guy is a product of such attitudes. If he is a monster, then Herr Frankenstein bears the greater responsibility.
That does not make him blameless. It means that he is culpable only to some degree below 100%. Punish him for that percentage he is culpable. Fine. But it means there is also a non-zero component of responsibility elsewhere, which should not go unpunished, and that there is a non-zero element of illness the guy has developed as a result, for which he should be treated.
In any such system, blaming one person is extremely easy but utterly futile. It doesn't fix the underlying problems which made the failure possible, and the ultimate problem is invariably the mental illness prevalent in modern management methods.
Now, I accept CAs don't always do a good job of validating who someone is. I also accept that they charge 5-10 times as much (per year) than the passport office, for what is basically the same thing. These are not the fault of the web of trust, but a fault in the way the industry is operated and regulated. These things won't be fixed by breaking the web of trust.
There are also two attacks against infrastructure which can compromise a key:
Of all of these, the last one is the only one anyone needs to take seriously. Even then, there are plenty of ways of making directories and files very secure, and making sure that potential exploits like buffer overflows are blocked in advance. (Just use a malloc replacement that prevents them.) The other attacks are so improbable that you can ignore them.
This leave one other attack vector:
This, according to reports, was used to obtain Microsoft's private keys from Verisign. Most reputable cert vendors have established better practices now. Simply choose one that will only deliver keys to an authorized contact point and only after a call-back check or some other authentication scheme.
They're not multicasting it as well? Damn. Thanks for the stream info, tho.
The summary says it's 4th generation. Most bittorrent clients are in C or Java, not Forth. (Ok, that was bad. But even so, I'd call bittorrent a first generation true peer-to-peer protocol, same with Gnutella and ED2K. Freenet, Tor and X-Bone might be considered second generation, and that's a push, but there frankly isn't anything out there I'd consider third generation, never mind fourth. You need to have something that is significantly different to qualify for a whole generation in tech jargon, which is why even the latest multi-core multi-way hybrid architecture CPUs are not considered to be running millionth-generation languages.)
The idea of certificates is to authenticate the connection, make it impossible to someone in the middle to pretend to be the server to the client, and the client to the server. Actually, it would be better to require users to have certificates as well, in many cases, as passwords tend to be too trivial.
Now, the price of certificates is horrendous. The passport office provides a document as good, or better, than many certificates, but it doesn't cost many hundreds of dollars to obtain a passport. In fact, as digital certificates are essentially the same as a passport with electronic information, it might be better if the passport office issued digital certificates along with physical passports as a combined package. The added cost to them would be practically nil, and the certificates would have a much greater credibility level than those by most corporations, at least for personal certs.
Heh. Actually, it's through science fiction that formed the basis of my ideas. (Books: weirdstone of brisingamen, moon of gomrath, eight keys to eden, starship medic, caves of steel, stars like dust, citizen of the galaxy; Short stories: The Tales of the Hexadecimal Kid; Television: doctor who, blake's 7, sapphire and steel, space: 1999, children of the stone, the tomorrow people.) Since then, I've explored different avenues of my own (nothing terribly original) but have basically come to the same premise you described, which is that of classic sci-fi.
I don't know. Aaaaaaaaaaaaaaaaaaaaaaaaaaaaarrrrrrrrrrrrrrrrrgh!
If the development of science, technology and society was kept in a dynamic equilibrium, with whichever one is faster pulling the others to where they should be, then a 1984 scenario becomes impossible. 1984-type dystopias are only possible if one of those three prongs exceeds the ability of the other two to keep it in balance. If society exceeds science or technology by too much, society's natural fears will cripple the other two and balance becomes impossible. If science exceeds society or technology, then what you understand exceeds what you can control and paranoia results. The cult of Pythagoras is a good example. If technology is king, then what you can do exceeds what you can control or understand, and the results are often self-destructive.
Well, if you're doing something like that, you would probably start with LynxOS or one of the carrier-grade Linux distros, remove absolutely all unnecessary kernel options, replace all software (including the init code) with the smallest and fastest code you can find, and place a lot more emphasis on tuning of that kind from a known highly stable start point.
More to the point, what is their favourite colour, and what is the airspeed of an unladen swallow?
You were trying to read my mind, weren't you? Well, it won't work. My tinfoil hat is the latest deluxe model. The nice gentleman from the NSA told me it was even checked for listening devices beforehand.
Grid computing has largely fixed the limitations on protocols and authentication, making "cloud computing" look, well, just a little bit wet. (There is now a very nice SASL layer for Globus, for example.) The main problem I have with the term "cloud computing" is that the term "cloud" already has a well-defined meaning in computer networks - it's any topology where you don't care about where things are or what things look like, you pass packets in and you get packets out. To use programming jargon, it is a black box. The internals don't matter. This fits with the idea of grid computing, but does NOT fit with the idea of a single, centralized off-site system. If you have to care about it being centralized somewhere, then it is not black box, because you DO care about what is on the inside.