I cannot believe you people. So, how many of you believe that Bush caused 9/11 by "missing obvious signs" that an attack was imminent? Now, when they go looking for those signs, everyone complains about losing freedom. You cannot have it both ways. Personally, I like freedom, thus I don't blame Bush for 9/11. However, since just about everyone did blame the government, now we have these idiotic systems put in place by the government to try to find that impossible needle in a haystack. Congratulations, everyone who blamed the government for 9/11 is now getting what they deserve. The funniest part is how people will quote security experts like Bruce Schneier, but only look at part of it and miss the whole message. He advocates that if we want to live with privacy, we have to accept that terrorist attacks will happen, and we would spend our resources most efficiently on emergency preparation rather than attack prevention. Since terrorist attacks do far more damage in the media than they do to actual lives, this actually makes sense.
On to the particular issue at hand, the whole point of feeding in all these variables into a statistical model is that a computer can do this better than a human. It's called data-mining, and its been going on in industry for ten years at least. While a human may just end up saying "He looks arab", the computer will base its decision only on what the variables actually suggest based on past data. This is more fair, not less. You can even set it up so that unless a person is flagged, no human ever gets to look at the personal data. You all joke about what the person is eating, but statistical models work best when you toss them all the possible features, and let the system figure it out. Anyone who's ever worked on machine learning would realize that humans cannot always judge what is a relevant feature and what isn't, and that sometimes a computer can learn a model that works well, but at first seems completely off the wall. Google "mushroom dataset" and get some background to this sort of thing with machine learning. There are rules which can calculate with very high probability if a mushroom is poisonous, yet the rules make no sense to the human observer. Welcome to data-mining; If you want to find a needle in a haystack, that's the way to go.
And about that data... the airline already had it, and probably already gave it to their marketing, who may have sold it to "partners". Other companies have been collecting data on people for many years. There is really nothing new about this information being collected. The only "outrage" is that a computer program gets to see it rather than only airline security personel.
Also, of course will not be allowed to see the model. If you have the model, you know how to avoid being flagged. That's like asking the local police for their driving schedule, or expecting highway patrol to reveal to you where all their speed traps are on a given day. That is simply not going to happen, and the reasons are obvious. An automated version such screening does not really represent anything new.
Now, of course, you can still be outraged if you want. Personally I'd rather have the privacy, and ditch such a system. However, don't be a hypocrite and expect the government to stop the next attack. Accept the 1 in 50,000 risk of being directly affected by an attack, and get on with your life... Worry about the real dangers that are likely to affect your life (bad diet, disease, accidents), rather than what Bruce Schneier calls "movie plot threats".
Believe it or not, but Katamari Damacy was a challenging game to make work from a graphics standpoint. There was a good article about it in Game Developer magazine a while back. The katamari ball builds up a huge number of objects, thus necessitating the use of low-poly models, even after fancy depth-based culling was implemented for the ball. What's funny is that they turned that "limitation" into a feature by going for a retro look, and nailed it. Along with the simple controls, an excellent soundtrack, and a very different concept of success/failure, it was an instant classic. It does push a playstation to its limits on a triangle-rate basis though.
Sure, but you'll have to start referring to satellite phones as "mobile phones" as well, and "fixed-line telephones" as immobile phones. Nobody even says "mobile phone", as they use the shortened name "mobile", which is dumb, as it could refer to just about anything (how about a car, that moves too).
If land-line telephones are on their way out, just call it a "phone" and be done with it. Mobility/wireless will just be assumed.
So the question is, with 16M colors, how much information can be stored there? 24 bits. "256 GB" is measuring bytes, and bytes are made of bits. Thus, ultimately we care about the number of bits, not the number of colors. So, he's right, you're wrong.
That's awesome. A supposed "genius" of data encoding and compression, and he uses GIFs for all of his photographs. Yep, definitely an image processing and compression expert...
Sorry, you are wrong. I recommend looking up "Information Theory" or "Encoding Theory", in particular stuff from Shannon.
Your matrix example is basically saying you can make a small amount of information into a larger, less efficient encoding. Of course you can. but that's not the direction that is interesting. I can repeat a bit a thousand times, but that doesn't mean I can take an arbitrary 1000-bit sequence, and store it in a lossless form with one bit. Thus the question is if you can always go both ways - take an arbitrary NxN matrix of values, store it as two 1xN vectors (encoding), and then recreate the original matrix (decoding). That's compression.
Another way to look at the fundamental limit is the following: You cannot compress all possible values of a data block. If you could, you could keep re-running the compression on the result, and eventually you will compress all possible data streams to a single bit - obviously that makes no sense, as you can't recover much from a single bit. Real compression works by making "common" sequences, with a lot of repetition, into shorter sequences, while uncommon ones get slightly longer. There will not be some "layer of abstraction" that can universally compress data.
Spend some time at Wikipedia and this will make a little more sense. It's actually pretty interesting stuff, as there really are hard limits you hit when sending or encoding information, and they can be calculated. The early work on this topic formed the basis for the modern digital world.
While my Athlon T-bird 900 could probably have cooked eggs (and it eventually *did* cook itself), it pales in comparison to Intel's Pentium4 NukeBurst architecture. I'm glad that both companies eventually got the message. Now we cook mainly with our graphics cards.
His research is using radiation exposure 4000-100000 times greater than what you'll get from an access point. Now, by all means be concerned about cell phones if you want, and follow the outcomes of "accelerated" studies using large doses. However, there is no substitute for long-term studies with meaningful doses. Most of this research is 10-15 years old, and no such long-term, normal-dosage study has emerged since then to demonstrate risk from low-power devices.
Drinking 1000 glasses of water will kill you quite easily; Does that mean drinking water demonstrates an adverse effect?
A cell phone is operating up to 1 watt a few inches from much of your brain, while a WiFi base station is under 100mw eight feet away. The difference in power reaching your brain is ~10000x. So, while I don't think its too much of a knee-jerk reaction to worry some about cell phones and to pursue research on it, worrying about something four orders of magnitude less powerful is ridiculous at this point. If it turns out cell phones kill people, this could be revisited. Doing anything before then is a misuse of resources that could be better spent on almost any kind of safety improvement.
So, what you're saying is basically: If you drink 1000 glasses of water in one day you will die, therefore each glass of water is dangerous. You're comparing apples and oranges, or rather pebbles and boulders. A microwave is so much more powerful than WiFi equipment that this is a conservative comparison (10,000x - 100,000x). I do think it is perfectly fine to be concerned about something like a cell phone, which is operating 2 inches from your brain. Any RF equipment not directly on your body will be minuscule by comparison however, thanks to our friendly inverse square law. Uninformed people fear the cell towers and WiFi base stations because they don't understand this. All they have to do to "protect" their child is to make sure their own child doesn't have a transmitter on them.
Well, for cooling it really depends on the case. For one of my Lian-Li's, taking off the side panel will mess with the airflow, but then I bought those cases for their well designed airflow. I've owned cheap cases in the past where they seemed to think that air could come from nowhere: "We'll pair a rear exhaust fan with a front fan, but to keep it quiet, we'll seal off the front entirely, except for a tiny bottom slit to suck up dust."
Are you trying to say that Apple didn't invent the non-beige computer case? Clearly the SGI indigo or onyx computers would only came in beige, right? What's next, are you going to tell me Apple didn't create the first mp3 player with their ultra-revolutionary ipod?
Niagra 1's FP indeed sucks, but I thought they were fixing that for Niagra 2. I heard about an FPU-per-core design being done (maybe that's a later follow-on?).
Google are experts at scaling out, not up. Running applications to serve 10000 users at the same time is different from executing one massive program. The main way in which Google and HPC (High Performance Computing, i.e. Supercomputers) overlap is in enormous file systems.
I disagree entirely. Consumers are willing to change on a dime. As recent history has shown, delays in Windows Vista have driven huge increases in people switching to Mac or Linux.
First off, don't treat me like an idiot. After 10 years of college and graduate school, I know perfectly well how to use google and wikipedia.
most of the controversy that I'm aware of right now does not involve taking stem cells from aborted fetuses, or aborted embryos. It involves taking embryos that were already created and frozen in a lab for in vitro fertilization, but never used.
The only reason the "current controversey" is about embryos is that there has already been a fight over other sources. When stem-cell research first started, they were using every available source, and some on the fringe even talked about deliberately getting women pregnant to generate cells. That sort of thing is what started the controversey, they pushed fast and hard on the boundary, until people realized what was going on, and some finally pushed back. Which brings us to where we are today, where the decision is over harvesting embryos. Don't think for a minute that the medical research community did this on their own; It was only after significant uproar that they relented.
So, given where we are, is it wrong to not want to pay for that sort of research? I'm not completely against privately funded research into such things, but I find it distasteful enough that I would not want to help support it. It's much like buying products not tested on animals - you aren't trying to put the animal testers in jail, but you try not to reward it by paying them. I realize that many in Congress, and the White House do not see this middle ground, and are less reasonable. Don't take that to mean everyone exists on that side only, or completely the opposite.
Furthermore, most in vitro clinics destroy the unwanted embryos after the couple has successfully conceived.
Do the couples get asked what to do with their embryos? I bet many clinics are not upfront about it.
Right now, these embryos are just being destroyed, but instead, they could be use to cure people!
If I'm guilty of misinformation, then you are just as guilty. Many of the current stem-cell treatments can use adult stem cells, which don't involve embryo/fetus harvesting (you can get adult cells from umbilical cords, for example). Researchers cry that we are targeting stem-cell research as a whole, when really its just embryonic stem-cell research which is targeted. They also cry that we are "banning research" while in many cases, we're just not publicly funding it. It seems to me that promising research should be able to justify funding from drug companies, which are constantly awash in funds while our government is broke. We've also got C. Reeves and M.J. Fox asking for embryo research, even though the most successful treatment to date for spinal injuries used adult stem cells. If that isn't trying to appeal to raw emotions, I don't know what is. It's also being deliberately misleading, since they don't mention the non-controversial approach even exists while they push for embryonic stem-cell research.
Is it really that bad to expect that when developing targeted treatments, the partially differentiated adult stem-cells are going to work adequately? The embryonic cells can do almost anything, which makes them more dangerous in that sense than a more targeted treatment. It's true that the embryonic cells are better for basic research, but is it too much to ask to either (1) do that with your own money, or (2) do that in another country.
I don't see the problem either. Laws are meant to be interpreted by a judge. They can appeal it if they don't like it.
The exact medium is now has become irrelevant, considering that automatic gateways exist between email, text messaging, IMs, and html (webmail,web im). Online is online, and in particular, IMs with delayed delivery (ICQ,AIM), and websites with accounts and "personal messages", can function exactly like traditional email. I don't see why someone should get off the hook for any non-RFC-conforming email message: "The message was missing required headers, therefore it was not email, and you must acquit."
Much of the progress in modern social government is due to judges willing to think beyond what legislatures are willing to consider. I am not a huge fan of the legal system, but congress/legislatures are much worse. The judges are almost always better informed when they make decisions compared to legislators when they pass laws.
So how long until we have some person who is opposed to harvesting aborted fetuses wielding this nugget of information in his crusade against taxpayer funding of stem-cell research?
There, I fixed it for you.
Why don't we start harvesting organs from prisoners against their will, and carry out various risky medical research on the long-term prison population? At least then we would be consistent. It's sad when the average person can watch a movie such as The Island, and yet not see any parallels to the choices we face in current society.
Slashdot Mirror
I cannot believe you people. So, how many of you believe that Bush caused 9/11 by "missing obvious signs" that an attack was imminent? Now, when they go looking for those signs, everyone complains about losing freedom. You cannot have it both ways. Personally, I like freedom, thus I don't blame Bush for 9/11. However, since just about everyone did blame the government, now we have these idiotic systems put in place by the government to try to find that impossible needle in a haystack. Congratulations, everyone who blamed the government for 9/11 is now getting what they deserve. The funniest part is how people will quote security experts like Bruce Schneier, but only look at part of it and miss the whole message. He advocates that if we want to live with privacy, we have to accept that terrorist attacks will happen, and we would spend our resources most efficiently on emergency preparation rather than attack prevention. Since terrorist attacks do far more damage in the media than they do to actual lives, this actually makes sense.
On to the particular issue at hand, the whole point of feeding in all these variables into a statistical model is that a computer can do this better than a human. It's called data-mining, and its been going on in industry for ten years at least. While a human may just end up saying "He looks arab", the computer will base its decision only on what the variables actually suggest based on past data. This is more fair, not less. You can even set it up so that unless a person is flagged, no human ever gets to look at the personal data. You all joke about what the person is eating, but statistical models work best when you toss them all the possible features, and let the system figure it out. Anyone who's ever worked on machine learning would realize that humans cannot always judge what is a relevant feature and what isn't, and that sometimes a computer can learn a model that works well, but at first seems completely off the wall. Google "mushroom dataset" and get some background to this sort of thing with machine learning. There are rules which can calculate with very high probability if a mushroom is poisonous, yet the rules make no sense to the human observer. Welcome to data-mining; If you want to find a needle in a haystack, that's the way to go.
And about that data... the airline already had it, and probably already gave it to their marketing, who may have sold it to "partners". Other companies have been collecting data on people for many years. There is really nothing new about this information being collected. The only "outrage" is that a computer program gets to see it rather than only airline security personel.
Also, of course will not be allowed to see the model. If you have the model, you know how to avoid being flagged. That's like asking the local police for their driving schedule, or expecting highway patrol to reveal to you where all their speed traps are on a given day. That is simply not going to happen, and the reasons are obvious. An automated version such screening does not really represent anything new.
Now, of course, you can still be outraged if you want. Personally I'd rather have the privacy, and ditch such a system. However, don't be a hypocrite and expect the government to stop the next attack. Accept the 1 in 50,000 risk of being directly affected by an attack, and get on with your life... Worry about the real dangers that are likely to affect your life (bad diet, disease, accidents), rather than what Bruce Schneier calls "movie plot threats".
Believe it or not, but Katamari Damacy was a challenging game to make work from a graphics standpoint. There was a good article about it in Game Developer magazine a while back. The katamari ball builds up a huge number of objects, thus necessitating the use of low-poly models, even after fancy depth-based culling was implemented for the ball. What's funny is that they turned that "limitation" into a feature by going for a retro look, and nailed it. Along with the simple controls, an excellent soundtrack, and a very different concept of success/failure, it was an instant classic. It does push a playstation to its limits on a triangle-rate basis though.
Sure, but you'll have to start referring to satellite phones as "mobile phones" as well, and "fixed-line telephones" as immobile phones. Nobody even says "mobile phone", as they use the shortened name "mobile", which is dumb, as it could refer to just about anything (how about a car, that moves too).
If land-line telephones are on their way out, just call it a "phone" and be done with it. Mobility/wireless will just be assumed.
Yes, you can prove it. This comment explains it quite well. Universal compression cannot exist.
What's sad is that you get the units right, but then get the exponentiation/bits wrong.
log2(2^18 * 2^2) = log2(2^18) + log2(2^2) = 18 + 2
i.e. 18 color bits + 2 shape bits = 20 bits.
So the question is, with 16M colors, how much information can be stored there? 24 bits. "256 GB" is measuring bytes, and bytes are made of bits. Thus, ultimately we care about the number of bits, not the number of colors. So, he's right, you're wrong.
P.S. 2^6 = 64, NOT 32
You are off by a factor of 8845.
Hint: log2(N) != N
You make the same error numerous people have made. 16M colors is only 24 bits, 16K colors is only 16 bits, so you are off by a factor of 1024.
(1200^2 * 8.5*11 * 16 / 8)/(1024^2) = 257 MB
That's awesome. A supposed "genius" of data encoding and compression, and he uses GIFs for all of his photographs. Yep, definitely an image processing and compression expert...
Sorry, you are wrong. I recommend looking up "Information Theory" or "Encoding Theory", in particular stuff from Shannon.
Your matrix example is basically saying you can make a small amount of information into a larger, less efficient encoding. Of course you can. but that's not the direction that is interesting. I can repeat a bit a thousand times, but that doesn't mean I can take an arbitrary 1000-bit sequence, and store it in a lossless form with one bit. Thus the question is if you can always go both ways - take an arbitrary NxN matrix of values, store it as two 1xN vectors (encoding), and then recreate the original matrix (decoding). That's compression.
Another way to look at the fundamental limit is the following: You cannot compress all possible values of a data block. If you could, you could keep re-running the compression on the result, and eventually you will compress all possible data streams to a single bit - obviously that makes no sense, as you can't recover much from a single bit. Real compression works by making "common" sequences, with a lot of repetition, into shorter sequences, while uncommon ones get slightly longer. There will not be some "layer of abstraction" that can universally compress data.
Spend some time at Wikipedia and this will make a little more sense. It's actually pretty interesting stuff, as there really are hard limits you hit when sending or encoding information, and they can be calculated. The early work on this topic formed the basis for the modern digital world.
It is annoylogy, you insensitive clod :)
No, it just seams to bes that way.
Care to back that up with a legitimate example? I don't recall many people bashing MS for revoking the pirated Vista activation code.
While my Athlon T-bird 900 could probably have cooked eggs (and it eventually *did* cook itself), it pales in comparison to Intel's Pentium4 NukeBurst architecture. I'm glad that both companies eventually got the message. Now we cook mainly with our graphics cards.
His research is using radiation exposure 4000-100000 times greater than what you'll get from an access point. Now, by all means be concerned about cell phones if you want, and follow the outcomes of "accelerated" studies using large doses. However, there is no substitute for long-term studies with meaningful doses. Most of this research is 10-15 years old, and no such long-term, normal-dosage study has emerged since then to demonstrate risk from low-power devices.
Drinking 1000 glasses of water will kill you quite easily; Does that mean drinking water demonstrates an adverse effect?
A cell phone is operating up to 1 watt a few inches from much of your brain, while a WiFi base station is under 100mw eight feet away. The difference in power reaching your brain is ~10000x. So, while I don't think its too much of a knee-jerk reaction to worry some about cell phones and to pursue research on it, worrying about something four orders of magnitude less powerful is ridiculous at this point. If it turns out cell phones kill people, this could be revisited. Doing anything before then is a misuse of resources that could be better spent on almost any kind of safety improvement.
So, what you're saying is basically: If you drink 1000 glasses of water in one day you will die, therefore each glass of water is dangerous. You're comparing apples and oranges, or rather pebbles and boulders. A microwave is so much more powerful than WiFi equipment that this is a conservative comparison (10,000x - 100,000x). I do think it is perfectly fine to be concerned about something like a cell phone, which is operating 2 inches from your brain. Any RF equipment not directly on your body will be minuscule by comparison however, thanks to our friendly inverse square law. Uninformed people fear the cell towers and WiFi base stations because they don't understand this. All they have to do to "protect" their child is to make sure their own child doesn't have a transmitter on them.
Well, for cooling it really depends on the case. For one of my Lian-Li's, taking off the side panel will mess with the airflow, but then I bought those cases for their well designed airflow. I've owned cheap cases in the past where they seemed to think that air could come from nowhere: "We'll pair a rear exhaust fan with a front fan, but to keep it quiet, we'll seal off the front entirely, except for a tiny bottom slit to suck up dust."
Are you trying to say that Apple didn't invent the non-beige computer case? Clearly the SGI indigo or onyx computers would only came in beige, right? What's next, are you going to tell me Apple didn't create the first mp3 player with their ultra-revolutionary ipod?
Niagra 1's FP indeed sucks, but I thought they were fixing that for Niagra 2. I heard about an FPU-per-core design being done (maybe that's a later follow-on?).
Google are experts at scaling out, not up. Running applications to serve 10000 users at the same time is different from executing one massive program. The main way in which Google and HPC (High Performance Computing, i.e. Supercomputers) overlap is in enormous file systems.
I disagree entirely. Consumers are willing to change on a dime. As recent history has shown, delays in Windows Vista have driven huge increases in people switching to Mac or Linux.
oh wait...
First off, don't treat me like an idiot. After 10 years of college and graduate school, I know perfectly well how to use google and wikipedia.
most of the controversy that I'm aware of right now does not involve taking stem cells from aborted fetuses, or aborted embryos. It involves taking embryos that were already created and frozen in a lab for in vitro fertilization, but never used.
The only reason the "current controversey" is about embryos is that there has already been a fight over other sources. When stem-cell research first started, they were using every available source, and some on the fringe even talked about deliberately getting women pregnant to generate cells. That sort of thing is what started the controversey, they pushed fast and hard on the boundary, until people realized what was going on, and some finally pushed back. Which brings us to where we are today, where the decision is over harvesting embryos. Don't think for a minute that the medical research community did this on their own; It was only after significant uproar that they relented.
So, given where we are, is it wrong to not want to pay for that sort of research? I'm not completely against privately funded research into such things, but I find it distasteful enough that I would not want to help support it. It's much like buying products not tested on animals - you aren't trying to put the animal testers in jail, but you try not to reward it by paying them. I realize that many in Congress, and the White House do not see this middle ground, and are less reasonable. Don't take that to mean everyone exists on that side only, or completely the opposite.
Furthermore, most in vitro clinics destroy the unwanted embryos after the couple has successfully conceived.
Do the couples get asked what to do with their embryos? I bet many clinics are not upfront about it.
Right now, these embryos are just being destroyed, but instead, they could be use to cure people!
If I'm guilty of misinformation, then you are just as guilty. Many of the current stem-cell treatments can use adult stem cells, which don't involve embryo/fetus harvesting (you can get adult cells from umbilical cords, for example). Researchers cry that we are targeting stem-cell research as a whole, when really its just embryonic stem-cell research which is targeted. They also cry that we are "banning research" while in many cases, we're just not publicly funding it. It seems to me that promising research should be able to justify funding from drug companies, which are constantly awash in funds while our government is broke. We've also got C. Reeves and M.J. Fox asking for embryo research, even though the most successful treatment to date for spinal injuries used adult stem cells. If that isn't trying to appeal to raw emotions, I don't know what is. It's also being deliberately misleading, since they don't mention the non-controversial approach even exists while they push for embryonic stem-cell research.
Is it really that bad to expect that when developing targeted treatments, the partially differentiated adult stem-cells are going to work adequately? The embryonic cells can do almost anything, which makes them more dangerous in that sense than a more targeted treatment. It's true that the embryonic cells are better for basic research, but is it too much to ask to either (1) do that with your own money, or (2) do that in another country.
I don't see the problem either. Laws are meant to be interpreted by a judge. They can appeal it if they don't like it.
The exact medium is now has become irrelevant, considering that automatic gateways exist between email, text messaging, IMs, and html (webmail,web im). Online is online, and in particular, IMs with delayed delivery (ICQ,AIM), and websites with accounts and "personal messages", can function exactly like traditional email. I don't see why someone should get off the hook for any non-RFC-conforming email message: "The message was missing required headers, therefore it was not email, and you must acquit."
Much of the progress in modern social government is due to judges willing to think beyond what legislatures are willing to consider. I am not a huge fan of the legal system, but congress/legislatures are much worse. The judges are almost always better informed when they make decisions compared to legislators when they pass laws.
So how long until we have some person who is opposed to harvesting aborted fetuses wielding this nugget of information in his crusade against taxpayer funding of stem-cell research?
There, I fixed it for you.
Why don't we start harvesting organs from prisoners against their will, and carry out various risky medical research on the long-term prison population? At least then we would be consistent. It's sad when the average person can watch a movie such as The Island, and yet not see any parallels to the choices we face in current society.
Would it be too much to ask to wait for the first design review?