The point I'm trying to make isn't about what people know. If the questions were asked in an unbiased way, then amongst the people who have no idea half should answer "true" and half should answer "false". The fact that our population would be outperformed by a population who was just randomly guessing shows a flaw in the study. Or perhaps this shows that our education system is systematically teaching people that electrons are bigger than atoms (but I doubt this).
According to this study less than half of Americans know whether electrons or atoms are bigger. Doesn't this indicate a flaw in the study? My lucky nickel is smarter than that.
My bet is that the way the questions were worded affected the answers people gave.
Out of curiousity, if someone came up with a means of distributing analog software (or does this already exist?), am I correct in assuming that it would not be covered by this legislation. Of course, if someone created something just to circumvent the legislation, the legislation would probably be edited rather quickly.
The computer technology is the one area of our economy where the government has been, for the most part, hands-off. As a result, technology has improved by leaps and bounds while prices have dramatically decreased.
The government keeping out of technology causes improvements??? While, government policies can slow technological innovations (see the protection of rot-13 under DMCA), I would argue that rapid technological improvements are due more to the relative newness of computers than to lack of governmental restrictions.
I probably don't speak for everyone in this case, but my experience with biology inspiring artifiical intelligence is as follows:
1. Start out with something vaguely like something biological (a neural net)
2. Talk about how you're going to use the biological model because evolution is powerful.
3. Make changes to the model that don't follow the biological model because it's easier to get reasonable results quickly if you're not restrained by biological constraints.
So, in short I thought the "biology inspires AI" idea was mostly smoke and mirrrors (at least in the classes I've taken). Biology did inspire the original idea, but biology got dropped after that.
The biggest difference between terrorist attacks and things like smoking deaths in my mind is that people who die of smoking related causes have made the choice to be subjected to smoking (I'm assuming that most deaths are not caused by second hand smoke). People subjected to terrorist attacks have made no decision to be subjected to the terrorist attack.
This is not to say that we shouldn't be concerned about the number of deaths due to smoking, just that the comparison of smoking deaths and terrorist attack deaths may not be very helpful.
Is there any network connection to the outside world in this system? If not, then I don't see how this could be any less secure than other voting methods.
Since I'd imagine blowing things up in the middle of the desert is not particularly good for the desert, there might also be some environmental concern here.
Am I just out of it today, or does that map only show the Western hemisphere?
Re:Why is this then worthy...
on
Share The Pi!
·
· Score: 1
The way the previous poster means repeating is this: an exact sequnece of digits is repeated over-and-over. So, the number 0.12341234124... is repeating since 1234 is repeated forever.
The example the previous poster used is not repeating in this sense.
I am not sure what the technical mathematical definition of repeating is (or even if one exists and/or is agreed on by mathematicians). However, I would lean toward the previous poster's usage. My reasoning: if we define repeating to be the existence of a pattern, then by definition pi is repeating because its decimal expansion has the pattern of the digits of pi. And if we use the definition in this way the defintion is useless because all numbers are repeating.
Yes, I'm assuing that histograms fall into some sort of non-uniform distribution. But, with the regularity of language I would think that this is a reasonable assumption. It may well be that the histograms fall randomly enough that this method works. However, without any evidence, arguing over who's intutition is right is pretty useless.
Using exact matches would not work very well. Spammers could just make slight changes to every email (change the name of the person it's addressed to) to get spam past the system. So, you would have to use some sort of similarity measure.
Which raises the questions: what similarity function would you use? What criteria determine when a group of histograms is similar enough and frequent enough to constitute spam?
To me it seems that checking the histogram is going to take at least O(N) time (you're going to have to read how many occurences of each hash value there were at some point). Thus, there is a practical limit on how big you can make N since the historgram checking algorithm has to be run on every piece of email.
I would think the distribution of histograms is not uniform. Common two word phrases like "if the", "i think", etc. would be more common than phrases like "zebra battleship" (same for 3 word phrases). Granted the hash values of the common phrases may all hash to different numbers, but even then some hash values will probably be more frequent than other hash values since certain phrases are more frequent than others. If anyone has any actual distributions, I'd be interested in how uniform they are.
Also, the spammers wouldn't have to send all the messages. They could look at the histogram of a bunch of regular emails and just send the spam messages whose histograms are close to a lot of the histograms of the regular emails. This assumes that spammers would have access to the hash function though.
There may be something I don't know about here, but I am one of those stupid people who ask to be removed from just about every spam email I get. This results in me getting virtually no spam.
I'm not sure, but I'd guess that certain histogram patterns would be common in non-spam email messages. If spammers wanted to get spam through, then all they would have to do is send out messages that have histograms that are common for non-spam email.
Also, I would think that creating such messages is easier than it sounds. Here's a quick way to generate lots of mesages that all say the same thing:
{Hi, Hello, Howdy, Good Day}, would you {like to, enjoy, want to, be interested in} trying a free trial subscription,...
Just pick one phrase out of each set of brackets and you can generate a lot of different messages. (I saw something like this in a cryptography class where the prof got something like 2^30 messages all saying the same thing.) With all those messages, one of them is likely to match a common histogram and if so the system is broken.
Doesn't this make it possible to assign an extremely large number to N - say, a billion times bigger than we've ever calculated pi out to before - and see if that resolves to a zero? And, if so, proving that pi eventually ends? and working backwards from this theoretical point, couldn't you quite easily find the last digit of pi basically using a binary tree method?
No. To show pi ends you would have to show that every digit after that 0 was also a 0. By the way, this cannot be done since pi is irrational and thus does not end.
What meaning could there possibly be in a pattern that exists only in a subset of the inifinite number of different representations, and which changes from base to base within that subset? As one of my math professors once put it, this is just an exercise in numerical masturbation.
The meaning might not be in the pattern itself. The existence of a pattern for pi in any given base could be the important part. Who knows, maybe this will give us more understanding of how different bases work which could give us an efficient means of factoring (not to say that base representations and factoring are linked at all, but they might be). And an efficient means of factoring is certainly important.
I have a blind friend who owns a speech synthesis card. Getting the "book" out of encrypted format would allow him to use other tools to extract the text and have the machine read it aloud for him.
According to Barnes & Noble's eBook comparison page, you can get spoken text from Adobe's reader with the book publisher's permission (under "Cool Features").
They could have tried strengthening Ebooks until it was strong enough to actually secure book content, but that was apparently too difficult or troublesome for them.
Does anyone know of any means to make an encryption system that works how the EBooks reader is supposed to? It seems to me that Adobe's setup is pretty hard to implement. Adobe has to put a key somewhere in the Ebooks reader (which could probably be found by looking at the executable file). And once someone gets the key, writing an alternate decrypter should be fairly easy. This is all assuming you know what method of encryption Adobe is using, but I thought that was a basic assumption of cryptography.
Yes it's a news graphic, but it should still be accurate. I think the suspicion that the media is biased, perhaps unintentionally, on this issue is supported by the loose checking of such things. Now, if all that is wrong is the labelling of 1900 instead of 2000 (or whatever it's supposed to be), the factual evidence of the graph is not hurt (not to say I agree or disagree with the predictions). However, the fact that the mislabelling (or perhaps more serious problems) was missed indicates that the graphs are not being checked very closely. Now I may be wrong, but if that graph was evidence against global warming, I would guess that it would have been scrutinized rather closely.
More evidence of a media bias: check out what one of the author's of the NAS's June report had to say about the media's treatment of the report.
Sure it would be nice if all foods came labeled with whether they contained genetically modified organisms, but this costs money. There is at least costs for labelling (probably pretty minor) and keeping food separate (I don't know how much this costs, but it seems that it might be significant).
Currently, companies can spend the money to label food as not genetically modified. However, it seems that few companies actually do this. This is conjecture, but I would guess that's because most of the consumers who say they want labelling are unwilling to pay for it.
Not mandating labeling seems to be the more free market way to go.
Slashdot Mirror
Interesting concept, but if the speed is 20 mph at the junctions, won't traffic bottleneck at the junctions and reduce the whole system to 20 mph.
I'm also having a hard time visualizing how there wouldn't be massive traffic jams at the junctions (from vehicles getting on and off).
The point I'm trying to make isn't about what people know. If the questions were asked in an unbiased way, then amongst the people who have no idea half should answer "true" and half should answer "false". The fact that our population would be outperformed by a population who was just randomly guessing shows a flaw in the study. Or perhaps this shows that our education system is systematically teaching people that electrons are bigger than atoms (but I doubt this).
According to this study less than half of Americans know whether electrons or atoms are bigger. Doesn't this indicate a flaw in the study? My lucky nickel is smarter than that.
My bet is that the way the questions were worded affected the answers people gave.
Out of curiousity, if someone came up with a means of distributing analog software (or does this already exist?), am I correct in assuming that it would not be covered by this legislation. Of course, if someone created something just to circumvent the legislation, the legislation would probably be edited rather quickly.
The computer technology is the one area of our economy where the government has been, for the most part, hands-off. As a result, technology has improved by leaps and bounds while prices have dramatically decreased.
The government keeping out of technology causes improvements??? While, government policies can slow technological innovations (see the protection of rot-13 under DMCA), I would argue that rapid technological improvements are due more to the relative newness of computers than to lack of governmental restrictions.
I probably don't speak for everyone in this case, but my experience with biology inspiring artifiical intelligence is as follows:
1. Start out with something vaguely like something biological (a neural net)
2. Talk about how you're going to use the biological model because evolution is powerful.
3. Make changes to the model that don't follow the biological model because it's easier to get reasonable results quickly if you're not restrained by biological constraints.
So, in short I thought the "biology inspires AI" idea was mostly smoke and mirrrors (at least in the classes I've taken). Biology did inspire the original idea, but biology got dropped after that.
The biggest difference between terrorist attacks and things like smoking deaths in my mind is that people who die of smoking related causes have made the choice to be subjected to smoking (I'm assuming that most deaths are not caused by second hand smoke). People subjected to terrorist attacks have made no decision to be subjected to the terrorist attack.
This is not to say that we shouldn't be concerned about the number of deaths due to smoking, just that the comparison of smoking deaths and terrorist attack deaths may not be very helpful.
Is there any network connection to the outside world in this system? If not, then I don't see how this could be any less secure than other voting methods.
Since I'd imagine blowing things up in the middle of the desert is not particularly good for the desert, there might also be some environmental concern here.
Am I just out of it today, or does that map only show the Western hemisphere?
The way the previous poster means repeating is this: an exact sequnece of digits is repeated over-and-over. So, the number 0.12341234124... is repeating since 1234 is repeated forever. The example the previous poster used is not repeating in this sense.
I am not sure what the technical mathematical definition of repeating is (or even if one exists and/or is agreed on by mathematicians). However, I would lean toward the previous poster's usage. My reasoning: if we define repeating to be the existence of a pattern, then by definition pi is repeating because its decimal expansion has the pattern of the digits of pi. And if we use the definition in this way the defintion is useless because all numbers are repeating.
Yes, I'm assuing that histograms fall into some sort of non-uniform distribution. But, with the regularity of language I would think that this is a reasonable assumption. It may well be that the histograms fall randomly enough that this method works. However, without any evidence, arguing over who's intutition is right is pretty useless.
Using exact matches would not work very well. Spammers could just make slight changes to every email (change the name of the person it's addressed to) to get spam past the system. So, you would have to use some sort of similarity measure.
Which raises the questions: what similarity function would you use? What criteria determine when a group of histograms is similar enough and frequent enough to constitute spam?
To me it seems that checking the histogram is going to take at least O(N) time (you're going to have to read how many occurences of each hash value there were at some point). Thus, there is a practical limit on how big you can make N since the historgram checking algorithm has to be run on every piece of email.
I would think the distribution of histograms is not uniform. Common two word phrases like "if the", "i think", etc. would be more common than phrases like "zebra battleship" (same for 3 word phrases). Granted the hash values of the common phrases may all hash to different numbers, but even then some hash values will probably be more frequent than other hash values since certain phrases are more frequent than others. If anyone has any actual distributions, I'd be interested in how uniform they are.
Also, the spammers wouldn't have to send all the messages. They could look at the histogram of a bunch of regular emails and just send the spam messages whose histograms are close to a lot of the histograms of the regular emails. This assumes that spammers would have access to the hash function though.
There may be something I don't know about here, but I am one of those stupid people who ask to be removed from just about every spam email I get. This results in me getting virtually no spam.
I'm not sure, but I'd guess that certain histogram patterns would be common in non-spam email messages. If spammers wanted to get spam through, then all they would have to do is send out messages that have histograms that are common for non-spam email.
...
Also, I would think that creating such messages is easier than it sounds. Here's a quick way to generate lots of mesages that all say the same thing:
{Hi, Hello, Howdy, Good Day}, would you {like to, enjoy, want to, be interested in} trying a free trial subscription,
Just pick one phrase out of each set of brackets and you can generate a lot of different messages. (I saw something like this in a cryptography class where the prof got something like 2^30 messages all saying the same thing.) With all those messages, one of them is likely to match a common histogram and if so the system is broken.
Doesn't this make it possible to assign an extremely large number to N - say, a billion times bigger than we've ever calculated pi out to before - and see if that resolves to a zero? And, if so, proving that pi eventually ends? and working backwards from this theoretical point, couldn't you quite easily find the last digit of pi basically using a binary tree method?
No. To show pi ends you would have to show that every digit after that 0 was also a 0. By the way, this cannot be done since pi is irrational and thus does not end.
What meaning could there possibly be in a pattern that exists only in a subset of the inifinite number of different representations, and which changes from base to base within that subset? As one of my math professors once put it, this is just an exercise in numerical masturbation.
The meaning might not be in the pattern itself. The existence of a pattern for pi in any given base could be the important part. Who knows, maybe this will give us more understanding of how different bases work which could give us an efficient means of factoring (not to say that base representations and factoring are linked at all, but they might be). And an efficient means of factoring is certainly important.
Does it matter that Sklyarov hold's the patent for the program? (according to the program itself)
To me this means he is not just an employee of a company.
I have a blind friend who owns a speech synthesis card. Getting the "book" out of encrypted format would allow him to use other tools to extract the text and have the machine read it aloud for him.
According to Barnes & Noble's eBook comparison page, you can get spoken text from Adobe's reader with the book publisher's permission (under "Cool Features").
They could have tried strengthening Ebooks until it was strong enough to actually secure book content, but that was apparently too difficult or troublesome for them.
Does anyone know of any means to make an encryption system that works how the EBooks reader is supposed to? It seems to me that Adobe's setup is pretty hard to implement. Adobe has to put a key somewhere in the Ebooks reader (which could probably be found by looking at the executable file). And once someone gets the key, writing an alternate decrypter should be fairly easy. This is all assuming you know what method of encryption Adobe is using, but I thought that was a basic assumption of cryptography.
I think you're confusing not random and repeating.
Not random does not necessarily mean repeating. The number 0.101001000100001... is not random, but it doesn't repeat.
Without the assumption that a non-random number is repeating (and thus rational), the proof doesn't work.
Yes it's a news graphic, but it should still be accurate. I think the suspicion that the media is biased, perhaps unintentionally, on this issue is supported by the loose checking of such things. Now, if all that is wrong is the labelling of 1900 instead of 2000 (or whatever it's supposed to be), the factual evidence of the graph is not hurt (not to say I agree or disagree with the predictions). However, the fact that the mislabelling (or perhaps more serious problems) was missed indicates that the graphs are not being checked very closely. Now I may be wrong, but if that graph was evidence against global warming, I would guess that it would have been scrutinized rather closely.
More evidence of a media bias: check out what one of the author's of the NAS's June report had to say about the media's treatment of the report.
This is a bit off-topic, but check out some graphs from the BBC's global climate change evidence.
Notice that emission reductions today affect what the temperature was 50 years ago.
Hmmm, methinks there's some bias here.
Sure it would be nice if all foods came labeled with whether they contained genetically modified organisms, but this costs money. There is at least costs for labelling (probably pretty minor) and keeping food separate (I don't know how much this costs, but it seems that it might be significant).
Currently, companies can spend the money to label food as not genetically modified. However, it seems that few companies actually do this. This is conjecture, but I would guess that's because most of the consumers who say they want labelling are unwilling to pay for it.
Not mandating labeling seems to be the more free market way to go.
Here's some pictures and more info on previous experiments.