As Staton says, Thompson's methods found that Pac Man was 62% violent, Dig Dug was 67% violent, and Centipede was 97% violent (!). These results (which, not so coincidentally, were expunged from the final report) indicate that the whole method is flawed. This only begs the question - why were these numbers removed? Perhaps because it would have signaled to anyone reading the study that it was hopelessly flwed?
"We don't need the death penalty. We've got the tossed salad man. Shit, if I had a choice right now between the electric chair and tossing a salad - I'd be like, 'so where do you plug it in? shouldn't I be wet?'
Everyone's talking about public education. Kids are outta control. We need tougher rules. We need prayer in schools. We don't need that shit. We just need the tossed salad man. He'd straighten those kids out. Hey, Jimmy. You got a D. You know what that means. NOOOO! NOOOO! I don't wanna toss a salad! I don't wanna toss a salad! I'm gonna read! I'm gonna learn to read" -- Chris Rock
Tor has three intermediate hops between you and the destination; this only has one - so you get lower latency. Also, with Tor, your download speed is the minimum of the 4 intermediate connections' bandwidths. If one of those people happens to be a dial up user, you will be getting dial-up speeds.
High Times is certainly more reputable than the Washington Times (which is to the newspaper industry what Fox News is to the TV news industry - part of the propaganda wing of the Repulican party).
My computer engineering group works rather extensively with IBM's T.J Watson research lab in New York (off the top of my head, we're working with them on two new architectures they are designing, and they used us as guinea pigs to test a new multi-threaded programming language they are developing). I can say first-hand that they do some really great work.
Incorrect. Nature released the data about a month after they ran the story. If you google around, you can find the pdf which contains the exact errors per article.
Did you know that Britannicaresponded to less than half of the mistakes picked up by Nature? And to date Britannica has not fixed any of them? (Wikipedia fixed every last one) And that, of the ones they did respond to, a number of Britannica's responses were laughable. One nature reviewer said that a particular Britannica article on a particular plant genus could apply to any of 10 other genesus. Britannica's response - "It's OK, because we're not an encyclopedia of botany and don't claim to be". In other words, "it's not a bug, it's feature"
"As far as whether you can free it from the GPL, I believe the answer is no. While the data is arguably merely facts, and therefore not protected by copyright law, I think there was a copyright amendment recently that made a particular compilation of data subject to copyright." - I am not a lawyer either, but I think I can help you out. Copyright law protection compilations where subjective taste is used to distinguish what gets included from what does not. I believe this was the distinction the Surpreme court used in Feist V Rural to determine that anthologies (the "The best of Edgar Allan Poe", for example) are protected by copyright, whereas complications of data that have no subjective criteria for inclusion (the phone book, in that particular case) are not subject to protection. In this case, I suspect they will take any CD's data; meaning that there is no subjective critera for inclusion, meaning that no copyright can be applied.
In May, I was a speaker at the ACM Conference on Computers, Privacy, and Freedom (CFP). On the last day of the conference, one of the speakers was the guy in charge of digization efforts at the Smithsonian Musuem of the American Indian. (Granted, a different branch of the government than the National Archives, which this story pertains to). He said that digization efforts are hampered by a number of issues, not the least of which are the sheer size of the collection, the relatively small budget available, the extreme difficulty of digitizing some parts of the collection (like a 16-ton statue, for example). At this point, even getting an electronic catalogue of the entire collection would be a huge step forward.
"What the censors were arguing is that if you obtain a legal copy of something, you've got the right to make and distribute a derivitve work from it." - incorrect. There original owner (the person who actually paid for it) is the only person who gets that second copy. It's exactly equivalent to making a personal backup copy, which most copyright lawyers believe is not a violation of copyright law. Or, to use your scenario, it's as if EvilCorp took the code, and modified it, and used the code only internally. There is no 'distribution'.
I remember back in the first quarter of 2003, when Wikipedia lost one of its three servers. Editing became unbearable - 2 minutes to load a page, 5 minutes to save one. Wikipedia's gotten better because it has upwards of 150 (or more) servers now:)
Commons - perfectly categorized? I'm sorry, but I have to strongly disagree. I like commons a lot (I'm one of the 140-odd admins there) but, IMO, the categorization system is far from perfect. The categorization system for music is an unmitigated disaster. Photos are better, but inconsistent.
(1) The distinction isn't clear cut. There are some people who write mostly or exclusively on a small number of articles to try to get them up to featured status, there are some people who divide their time and once in a great while might try to get something featured, and there are people who have never been to the featured article candidates page. (Full disclosure - I'm the person who oversees the whole system); and (2) The distinction is entirely self-selected. Nobody is forced to work on anything in particular, so if someone never wants to do a featured article, that is his choice. It's not as if people are being prevented from writing FAs (quite the opposite).
One Wikipedia contributor, David Gerard, did a nice job of summarizing Wikipedia's dirty little secret of how it works: 'On Wikipedia, the reward for a job well done is another three jobs'. Once someone establishes himself as being reliable, trustworthy, comptentent, 'etc, they tend to get handed a lot of responsability in short order. A relatively small number of people tend to wear many hats. (Myself included - I'm an administrator, burecreat, arbitrator, checkuser-weilder, member of the communications and press committees, handler of email via OTRS, 'etc)
Result in a nutshell: If I own a DVD, I cannot pay someone to make a copy of that movie for me sans parts I might find offensive. It's not censorship, because *I'm the one asking him to do it for me*. But in yet another defeat for personal freedom (and another win for the moneyed interests), the courts have found that this is a violation of copyright law.
...the write up here is awfully vague. The incident it refers to happened on Big Brother Australia - one of the guys in the house held a girl down while the other rubbed his penis on her face. Both are now facing charges.
I don't think the integer units are based on anything. The whole chip is being custom designed from scratch. (Interestingly enough, the VHDL code for the chip is being written by only one guy - the project leader himself).
As far as instruction re-ordering -- for parallel computation, the big peformance hits occur with waits, synchronizations/barriers, and locks/mutexes. Making these cheap and reducing the number of them is the biggest way to increase performance.
What you are describing has already been done, and was done quite a while ago. Around 1990, NASA realized that the way we do parallel benchmarks sucks. The way most benchmarks (including hte parallel ones) work is that some organization posts the code, and people have to compile and run the code as-is. There's not much room there for optimization (other than tweaking the compiler flags, some trivial hardware settings, 'etc), which is essential to getting good parallel performance (because parallel machines vary so widely). So performance was tied very closely to the implimention over which nobody had any control.
NASA approached the problem differently. Their numerical analysis group put out a set of "paper and pencil" benchmarks (based on real world problems that one would encounter, for example, fluid dynamics). The actual implimentation was left up to the individual companies. This is what we know today as the NAS benchmark suite.
The compiler [the current version, at any rate] is based on gcc. So it sports the same out-of-order execution you would expect to get from compile-time optimization. I am not sure if it has hardware-based re-ordering. My guess would be that no, it does not, but without the Principles of Operation in front of me, I couldn't say (the advisor borrowed my paper copy for IPDPS 2006 and hasn't given it back yet).
You have to ask yourself - 57% of what? Of the peak theoretical performance. Peak theoretical performance is the performance you get under ludicrious circumstances - a continious stream of (only) multiply-and-accumulate instructions. (Each MDAC counts as two floating point operations, even though many architectures impliment it as a single instruction). Not only is it impossible to reach 100% effeciency with any program, but it's nearly impossible to even approach 100% if the program is supposed to do something useful (as opposed to being a toy program designed to get to 100% utilization).
Cell was designed around one single objective - to get a clock rate as sickeningly high as possible, because clock speed cells. Trust me when I say that programmability was not (at all) a consideration (I should mention - my research group got one of the very first Cell processor's sent to the US. We are currently in the process of implimenting OpenMP on it to make it a little nicer to program).
As far as writing multi-threaded code, I've spent the last 5 months rewriting the NAS CG benchmark to work effeciently on Cyclops64, which will probably play some part of my PhD thesis. (A sidenote: All of NASA's NAS implimentations are written in Fortran (except Integer Sort), which would have necessitated me rewriting NAS-CG in C. Fortunately, I didn't have to start from scratch, because the Japanese had already done the hard part).
Slashdot Mirror
I believe the whole point of my post is that the numbers are meaningless, as anyone familiar with those games can attest to.
As Staton says, Thompson's methods found that Pac Man was 62% violent, Dig Dug was 67% violent, and Centipede was 97% violent (!). These results (which, not so coincidentally, were expunged from the final report) indicate that the whole method is flawed. This only begs the question - why were these numbers removed? Perhaps because it would have signaled to anyone reading the study that it was hopelessly flwed?
"We don't need the death penalty. We've got the tossed salad man. Shit, if I had a choice right now between the electric chair and tossing a salad - I'd be like, 'so where do you plug it in? shouldn't I be wet?'
Everyone's talking about public education. Kids are outta control. We need tougher rules. We need prayer in schools. We don't need that shit. We just need the tossed salad man. He'd straighten those kids out. Hey, Jimmy. You got a D. You know what that means. NOOOO! NOOOO! I don't wanna toss a salad! I don't wanna toss a salad! I'm gonna read! I'm gonna learn to read" -- Chris Rock
But, as some people have found out, it can be subpoenaed by someone suing you and used against you in court.
One logical hop; any wide-area-connection has (almost by defintion) more than one physical hop. Yeesh... picky picky.
Tor has three intermediate hops between you and the destination; this only has one - so you get lower latency. Also, with Tor, your download speed is the minimum of the 4 intermediate connections' bandwidths. If one of those people happens to be a dial up user, you will be getting dial-up speeds.
High Times is certainly more reputable than the Washington Times (which is to the newspaper industry what Fox News is to the TV news industry - part of the propaganda wing of the Repulican party).
My computer engineering group works rather extensively with IBM's T.J Watson research lab in New York (off the top of my head, we're working with them on two new architectures they are designing, and they used us as guinea pigs to test a new multi-threaded programming language they are developing). I can say first-hand that they do some really great work.
Incorrect. Nature released the data about a month after they ran the story. If you google around, you can find the pdf which contains the exact errors per article.
Did you know that Britannicaresponded to less than half of the mistakes picked up by Nature? And to date Britannica has not fixed any of them? (Wikipedia fixed every last one) And that, of the ones they did respond to, a number of Britannica's responses were laughable. One nature reviewer said that a particular Britannica article on a particular plant genus could apply to any of 10 other genesus. Britannica's response - "It's OK, because we're not an encyclopedia of botany and don't claim to be". In other words, "it's not a bug, it's feature"
"As far as whether you can free it from the GPL, I believe the answer is no. While the data is arguably merely facts, and therefore not protected by copyright law, I think there was a copyright amendment recently that made a particular compilation of data subject to copyright." - I am not a lawyer either, but I think I can help you out. Copyright law protection compilations where subjective taste is used to distinguish what gets included from what does not. I believe this was the distinction the Surpreme court used in Feist V Rural to determine that anthologies (the "The best of Edgar Allan Poe", for example) are protected by copyright, whereas complications of data that have no subjective criteria for inclusion (the phone book, in that particular case) are not subject to protection. In this case, I suspect they will take any CD's data; meaning that there is no subjective critera for inclusion, meaning that no copyright can be applied.
In May, I was a speaker at the ACM Conference on Computers, Privacy, and Freedom (CFP). On the last day of the conference, one of the speakers was the guy in charge of digization efforts at the Smithsonian Musuem of the American Indian. (Granted, a different branch of the government than the National Archives, which this story pertains to). He said that digization efforts are hampered by a number of issues, not the least of which are the sheer size of the collection, the relatively small budget available, the extreme difficulty of digitizing some parts of the collection (like a 16-ton statue, for example). At this point, even getting an electronic catalogue of the entire collection would be a huge step forward.
"What the censors were arguing is that if you obtain a legal copy of something, you've got the right to make and distribute a derivitve work from it." - incorrect. There original owner (the person who actually paid for it) is the only person who gets that second copy. It's exactly equivalent to making a personal backup copy, which most copyright lawyers believe is not a violation of copyright law. Or, to use your scenario, it's as if EvilCorp took the code, and modified it, and used the code only internally. There is no 'distribution'.
I remember back in the first quarter of 2003, when Wikipedia lost one of its three servers. Editing became unbearable - 2 minutes to load a page, 5 minutes to save one. Wikipedia's gotten better because it has upwards of 150 (or more) servers now :)
Commons - perfectly categorized? I'm sorry, but I have to strongly disagree. I like commons a lot (I'm one of the 140-odd admins there) but, IMO, the categorization system is far from perfect. The categorization system for music is an unmitigated disaster. Photos are better, but inconsistent.
Two things I have observed:
(1) The distinction isn't clear cut. There are some people who write mostly or exclusively on a small number of articles to try to get them up to featured status, there are some people who divide their time and once in a great while might try to get something featured, and there are people who have never been to the featured article candidates page. (Full disclosure - I'm the person who oversees the whole system); and (2) The distinction is entirely self-selected. Nobody is forced to work on anything in particular, so if someone never wants to do a featured article, that is his choice. It's not as if people are being prevented from writing FAs (quite the opposite).
One Wikipedia contributor, David Gerard, did a nice job of summarizing Wikipedia's dirty little secret of how it works: 'On Wikipedia, the reward for a job well done is another three jobs'. Once someone establishes himself as being reliable, trustworthy, comptentent, 'etc, they tend to get handed a lot of responsability in short order. A relatively small number of people tend to wear many hats. (Myself included - I'm an administrator, burecreat, arbitrator, checkuser-weilder, member of the communications and press committees, handler of email via OTRS, 'etc)
Result in a nutshell: If I own a DVD, I cannot pay someone to make a copy of that movie for me sans parts I might find offensive. It's not censorship, because *I'm the one asking him to do it for me*. But in yet another defeat for personal freedom (and another win for the moneyed interests), the courts have found that this is a violation of copyright law.
...the write up here is awfully vague. The incident it refers to happened on Big Brother Australia - one of the guys in the house held a girl down while the other rubbed his penis on her face. Both are now facing charges.
I honestly have no idea about inter-chip communication issues. All of our work has focused on intra-chip issues.
I don't think the integer units are based on anything. The whole chip is being custom designed from scratch. (Interestingly enough, the VHDL code for the chip is being written by only one guy - the project leader himself).
As far as instruction re-ordering -- for parallel computation, the big peformance hits occur with waits, synchronizations/barriers, and locks/mutexes. Making these cheap and reducing the number of them is the biggest way to increase performance.
What you are describing has already been done, and was done quite a while ago. Around 1990, NASA realized that the way we do parallel benchmarks sucks. The way most benchmarks (including hte parallel ones) work is that some organization posts the code, and people have to compile and run the code as-is. There's not much room there for optimization (other than tweaking the compiler flags, some trivial hardware settings, 'etc), which is essential to getting good parallel performance (because parallel machines vary so widely). So performance was tied very closely to the implimention over which nobody had any control.
NASA approached the problem differently. Their numerical analysis group put out a set of "paper and pencil" benchmarks (based on real world problems that one would encounter, for example, fluid dynamics). The actual implimentation was left up to the individual companies. This is what we know today as the NAS benchmark suite.
The compiler [the current version, at any rate] is based on gcc. So it sports the same out-of-order execution you would expect to get from compile-time optimization. I am not sure if it has hardware-based re-ordering. My guess would be that no, it does not, but without the Principles of Operation in front of me, I couldn't say (the advisor borrowed my paper copy for IPDPS 2006 and hasn't given it back yet).
You have to ask yourself - 57% of what? Of the peak theoretical performance. Peak theoretical performance is the performance you get under ludicrious circumstances - a continious stream of (only) multiply-and-accumulate instructions. (Each MDAC counts as two floating point operations, even though many architectures impliment it as a single instruction). Not only is it impossible to reach 100% effeciency with any program, but it's nearly impossible to even approach 100% if the program is supposed to do something useful (as opposed to being a toy program designed to get to 100% utilization).
Cell was designed around one single objective - to get a clock rate as sickeningly high as possible, because clock speed cells. Trust me when I say that programmability was not (at all) a consideration (I should mention - my research group got one of the very first Cell processor's sent to the US. We are currently in the process of implimenting OpenMP on it to make it a little nicer to program).
As far as writing multi-threaded code, I've spent the last 5 months rewriting the NAS CG benchmark to work effeciently on Cyclops64, which will probably play some part of my PhD thesis. (A sidenote: All of NASA's NAS implimentations are written in Fortran (except Integer Sort), which would have necessitated me rewriting NAS-CG in C. Fortunately, I didn't have to start from scratch, because the Japanese had already done the hard part).