Yes, it's very difficult, if not impossible, to prove whether a scientific theory is true or not. In fact, when does something count as a scientific theory at all?
My favorite answer to this is Karl Popper's response: that something is a scientific theory when it is falsifiable. By this standard, creationism isn't a scientific theory because there's (I believe) no evidence or argument which you can produce to convince creationists that their theory is incorrect.
You say that effectiveness, or how much you can explain using a theory, is a better judge of a theory's value. Doesn't this imply that you believe that there is some sort of reality or truth (so to speak) with which you can evaluate theories by?
But back to maths versus science. My point is that treating maths as a simpler case of science, as you seem to be doing, is invalid, because they a really 2 different beasts: unlike science, maths does not make use of reality as a benchmark, and science does not suffer from the problems associated with defining axioms that maths does.
Truth is a weak foundation. Mathematics is on no firmer ground than evolution. Godel and Tarski showed that mathematics isn't about truth, it's about logical relationships between statements. Your system is never any better than your axioms, and no finite set of axioms can ever suffice.
What you've described is exactly why mathematics isn't a science. Mathematicians define axioms, scientists make empirical measurements -- whether you like it or not, there is an independent reality distinct from how you define your axioms.
Yes. Imagine what would happen if a credit card company using the Linux name turned out to be a scam. Linus should start requiring that these companies obtain explicit permission from him, and give permission only after thoroughly investigating the company's background -- perhaps some separate or existing organisation can be setup to do this (funded using fees paid by the applicants of course).
Uh, if you don't have a good credit rating, it means that you either don't buy much with your credit card (and hence don't make much money for the credit card company), or don't pay your bills on time (or both). In either case, you're not a good buy for the credit card company i.e. a credit risk.
I first read Neuromancer only many years after it had been released. By that time, the whole idea of cyberpunk had already been widely popularized and even somewhat stale. Reading Neuromancer, because it had been so widely copied, the cyberpunk ideas in Neuromancer itself seemed incredibly cliched. Without originality to sustain it, the plot came across as weak and confused, and the characters cardboard-like. I think this is the problem with cyberpunk, and the reason why writers have moved away from it so rapidly: after the initial shock value, the genre doesn't really have enduring quality.
Re:Proof that it's optimal? I don't believe that.
on
David Huffman is Dead
·
· Score: 2
Yes, so you code more than one symbol at a time. An example:
Symbols A, B. P(A)=0.8, P(B)=0.2 Code A=0, Code B=1, Expected length=1
Code 2 symbols at a time: P(AA)=0.64, P(AB)=0.16, P(BA)=0.16, P(BB)=0.04 Code AA=0, AB=10, BA=110, BB=111 Expected length=0.78
Code more symbols at a time if you want (but returns are diminishing)
The shortest possible expected length is given by the entropy which is equal to -P(A)logP(A)-P(B)logP(B) =0.72
I'm not sure how arithmetic coding works, but I highly doubt that it gives you the theoretically optimal coding just like that.
Not at all. For every decodable code, there's an equivalent prefix free code. So it's sufficient to consider only prefix free codes. So the Huffman code is optimal among all decodable codes. Note that you can actually encode more than one symbol at a time i.e. if you've symbols A & B, you can code it as AA,AB,BA,BB. By coding more and more symbols at a time, you can eventually reach the limit determined by the entropy of your signal source.
While I'll hate to see a world with only one browser that matters, supporting an inferior product doesn't really make sense to me either: both from a practical point of view and as a matter of principle i.e. why should I support a company that makes a shoddy, inferior product? -- it just rewards mediocrity..
I'll support Mozilla based on its potential to compete with IE, but I expect a good product eventually (soon!); goodwill by the community will only carry them so far.
In what way will Netscape 5 be a quantum leap over Netscape 4?
More stable? Faster? Less bloaty? Wow, some quantum leap..
These are basically fixes for bugs that should never have been present had Netscape been written well the first time around. Hardly a quantum leap. NS 1->NS 2 was a quantum leap. NS 5 is just a bug fix release that has been mired down and long delayed.
The problem with the argument that it's kinder to euthanise a disabled child is this: if the child really felt that he would be better off dead than alive, he could very well kill himself later on.
Sounds absurd doesn't it? Then it's ridiculous to think that you're doing a disabled child a favor by killing him -- after all, even the most severely disabled people show a fierce desire to live.
The problem with the Turing Test is that it tries to make a computer human and that's not really what AI is all about - it's more about trying to solve problems using various techniques in order to make programs useful. (Maybe making a computer human is not all that useful;) )
What you describe is the viewpoint of those who have given up on developing a general intelligence because the problem has proved to be too difficult. "Solving problems using various techniques in order to make programs useful" isn't AI. It's standard algorithm development. It's AI in its most limited sense, where highly specific and limited intelligence is applied to specific problems.
AI more generally is about developing an artificial intelligence i.e. a computer that can convince us that it's conscious. It's a higher goal than a purely utilitarian one (but will no doubt prove to be far more useful in the long run). This means that the AI has to have human characteristics, and a test along the lines of the Turing test is the only way to measure this.
The guy broke in because of a faulty CGI script! No fair, the test was supposed to test the underlying OS alone! Well bugs in Linux helped! Well they should have known to apply the patches! ... blah blah blah. etc.etc.etc.
The lesson to be learnt from this is that there's no such thing as the "security of this or the security of that alone". Security is a holistic concept, and a weak link phenomena. Everything has to be considered when designing a secure system, including human factors (how easy is it to make the system secure? How likely is it that people will make errors? etc.).
It's pointless to say: We're testing the security of the OS alone -- because there's no such thing. The PCWeek test is meaningful in the sense that it reveals how difficult it may be to make the a Linux system, as a whole, secure.
A more general question: Is there a good way of writing GUI based apps that are at least source code compatible across multiple platforms (e.g. unix, mac, windows)? Java is one way, are there any others? And preferably free.
The article is very strange. Sometimes it speaks in past tense, as though the device has already been developed: "It claims it has developed a hand-held device that can break the code in 12 microseconds." While at other points it speaks as though the device has yet to, but will be developed: "While quantum computers may be some time off, when they are available no communication will be secure unless it is quantum."
The article also confuses Quantum computing with Quantum Cryptography. The whole thing may just be the work of a confused journalist confusing a proposed, speculative design with something that has been built.
It sounds like a bad idea to use elephants as hosts. Mammoths and elephants are probably sufficiently alike so that in the case of contamination, it'll be difficult to tell how much elephant and how much mammoth the creature is?
While it's true that people could theoretically use crypto with some sort of stenography even if crypto were outlawed, this isn't really the point.
The point is to hinder the widespread deployment of cryptography. If crypto were outlawed (or heavily restricted as it is now), legitimate corporations can't release cryptographic software. This is good enough for the government since it means that most people won't be able to use cryptography since they don't have the know how to code it up themselves, or scour the net for some hacked up crypto software (and they probably wouldn't know how to use it even if they could find it).
This is good for the government from a law enforcement stand point since if they're looking for encrypted communications by criminals, they only have to sift through a fairly small number of encrypted messages. Or as other posters have suggested, the government's objective may just be to spy on the general populace.
Slashdot Mirror
Yes, it's very difficult, if not impossible, to prove whether a scientific theory is true or not. In fact, when does something count as a scientific theory at all?
My favorite answer to this is Karl Popper's response: that something is a scientific theory when it is falsifiable. By this standard, creationism isn't a scientific theory because there's (I believe) no evidence or argument which you can produce to convince creationists that their theory is incorrect.
You say that effectiveness, or how much you can explain using a theory, is a better judge of a theory's value. Doesn't this imply that you believe that there is some sort of reality or truth (so to speak) with which you can evaluate theories by?
But back to maths versus science. My point is that treating maths as a simpler case of science, as you seem to be doing, is invalid, because they a really 2 different beasts: unlike science, maths does not make use of reality as a benchmark, and science does not suffer from the problems associated with defining axioms that maths does.
Truth is a weak foundation. Mathematics is on no firmer ground than evolution. Godel and Tarski showed that mathematics isn't about truth, it's about logical relationships between statements. Your system is never any better than your axioms, and no finite set of axioms can ever suffice.
What you've described is exactly why mathematics isn't a science. Mathematicians define axioms, scientists make empirical measurements -- whether you like it or not, there is an independent reality distinct from how you define your axioms.
Even so, this shows that the bible does not contain the literal, exact, truth.
Yes. Imagine what would happen if a credit card company using the Linux name turned out to be a scam. Linus should start requiring that these companies obtain explicit permission from him, and give permission only after thoroughly investigating the company's background -- perhaps some separate or existing organisation can be setup to do this (funded using fees paid by the applicants of course).
Uh, if you don't have a good credit rating, it means that you either don't buy much with your credit card (and hence don't make much money for the credit card company), or don't pay your bills on time (or both). In either case, you're not a good buy for the credit card company i.e. a credit risk.
I first read Neuromancer only many years after it had been released. By that time, the whole idea of cyberpunk had already been widely popularized and even somewhat stale. Reading Neuromancer, because it had been so widely copied, the cyberpunk ideas in Neuromancer itself seemed incredibly cliched. Without originality to sustain it, the plot came across as weak and confused, and the characters cardboard-like. I think this is the problem with cyberpunk, and the reason why writers have moved away from it so rapidly: after the initial shock value, the genre doesn't really have enduring quality.
Yes, so you code more than one symbol at a time.
An example:
Symbols A, B. P(A)=0.8, P(B)=0.2
Code A=0, Code B=1, Expected length=1
Code 2 symbols at a time:
P(AA)=0.64, P(AB)=0.16, P(BA)=0.16, P(BB)=0.04
Code AA=0, AB=10, BA=110, BB=111
Expected length=0.78
Code more symbols at a time if you want (but returns are diminishing)
The shortest possible expected length is given by the entropy which is equal to -P(A)logP(A)-P(B)logP(B) =0.72
I'm not sure how arithmetic coding works, but I highly doubt that it gives you the theoretically optimal coding just like that.
Not at all. For every decodable code, there's an equivalent prefix free code. So it's sufficient to consider only prefix free codes. So the Huffman code is optimal among all decodable codes. Note that you can actually encode more than one symbol at a time
i.e. if you've symbols A & B, you can code it as AA,AB,BA,BB. By coding more and more symbols at a time, you can eventually reach the limit determined by the entropy of your signal source.
While I'll hate to see a world with only one browser that matters, supporting an inferior product doesn't really make sense to me either: both from a practical point of view and as a matter of principle i.e. why should I support a company that makes a shoddy, inferior product? -- it just rewards mediocrity..
I'll support Mozilla based on its potential to compete with IE, but I expect a good product eventually (soon!); goodwill by the community will only carry them so far.
In what way will Netscape 5 be a quantum leap over Netscape 4?
More stable? Faster? Less bloaty?
Wow, some quantum leap..
These are basically fixes for bugs that should never have been present had Netscape been written well the first time around. Hardly a quantum leap. NS 1->NS 2 was a quantum leap. NS 5 is just a bug fix release that has been mired down and long delayed.
The problem with the argument that it's kinder to euthanise a disabled child is this: if the child really felt that he would be better off dead than alive, he could very well kill himself later on.
Sounds absurd doesn't it? Then it's ridiculous to think that you're doing a disabled child a favor by killing him -- after all, even the most severely disabled people show a fierce desire to live.
The problem with the Turing Test is that it tries to make a computer human and that's not really what AI is all about - it's more about trying to solve problems using various techniques in order to make programs useful. (Maybe making a computer human is not all that useful ;) )
What you describe is the viewpoint of those who have given up on developing a general intelligence because the problem has proved to be too difficult. "Solving problems using various techniques in order to make programs useful" isn't AI. It's standard algorithm development. It's AI in its most limited sense, where highly specific and limited intelligence is applied to specific problems.
AI more generally is about developing an artificial intelligence i.e. a computer that can convince us that it's conscious. It's a higher goal than a purely utilitarian one (but will no doubt prove to be far more useful in the long run). This means that the AI has to have human characteristics, and a test along the lines of the Turing test is the only way to measure this.
The guy broke in because of a faulty CGI script!
No fair, the test was supposed to test the underlying OS alone!
Well bugs in Linux helped!
Well they should have known to apply the patches!
...
blah blah blah. etc.etc.etc.
The lesson to be learnt from this is that there's no such thing as the "security of this or the security of that alone". Security is a holistic concept, and a weak link phenomena. Everything has to be considered when designing a secure system, including human factors (how easy is it to make the system secure? How likely is it that people will make errors? etc.).
It's pointless to say: We're testing the security of the OS alone -- because there's no such thing. The PCWeek test is meaningful in the sense that it reveals how difficult it may be to make the a Linux system, as a whole, secure.
Maviglio said that the anti-spam bill, known as the Can Spam Act , had picked up a half-dozen additional sponsors in a 48-hour time span.
hmm..It might be possible that the aide deliberately sent out the spam to gain support for the anti-spam bill.
A more general question: Is there a good way of writing GUI based apps that are at least source code compatible across multiple platforms (e.g. unix, mac, windows)? Java is one way, are there any others? And preferably free.
The article is very strange. Sometimes it speaks in past tense, as though the device has already been developed: "It claims it has developed a hand-held device that can break the code in 12 microseconds." While at other points it speaks as though the device has yet to, but will be developed: "While quantum computers may be some time off, when they are available no communication will be secure unless it is quantum."
The article also confuses Quantum computing with Quantum Cryptography. The whole thing may just be the work of a confused journalist confusing a proposed, speculative design with something that has been built.
It sounds like a bad idea to use elephants as hosts. Mammoths and elephants are probably sufficiently alike so that in the case of contamination, it'll be difficult to tell how much elephant and how much mammoth the creature is?
While it's true that people could theoretically use crypto with some sort of stenography even if crypto were outlawed, this isn't really the point.
The point is to hinder the widespread deployment of cryptography. If crypto were outlawed (or heavily restricted as it is now), legitimate corporations can't release cryptographic software. This is good enough for the government since it means that most people won't be able to use cryptography since they don't have the know how to code it up themselves, or scour the net for some hacked up crypto software (and they probably wouldn't know how to use it even if they could find it).
This is good for the government from a law enforcement stand point since if they're looking for encrypted communications by criminals, they only have to sift through a fairly small number of encrypted messages. Or as other posters have suggested, the government's objective may just be to spy on the general populace.