Do you seriously think that earthquakes wouldnt cause of cracks in ice?
As far as weathering effects in the Antarctic, precipation averages about 16.5cm annual, making it the largest desert in the world.
The surface area of Antarctica is 14,200,000/km^2, so even though it is a desert, the amount of ice which would accumulate if there were no loss is staggering. This particular loss isn't very significant, but makes for a great story due to its surface area being equatable to something (ex: The state of RI) which is considered "large."
What is important is the volume of ice lost by this event, not the area of ice lost. All of these news agencies have latched onto the area of ice lost, not its volume, and that is in part due to the scientists themselves wrapping this up as dramatic.
Dont forget that once you form a corporation, personal responsibility is very very limited. If this company isnt making much money, then it doesn't have much to lose.
He should contract a law firm. This is too big to cheap out on, and a diverse set of skills will be needed to maximize the severe punishment that can potentialy be levied against these guys.
You don't want the countersuit to be a single-shot weapon. You want it to be an agonizing series of laser guided charges that simply wont go away no matter how this company tries to weasel out of the situation they have put themselves in.
..and of course you realize that the game engine this is based upon was originally a proprietary one, right?
They have risen to the level of using a proprietary engine, QUAKE ONE, which was open sourced NINE YEARS AGO.
This doesnt say anything good about this case of open source gaming at all, especialy since QUAKE THREE was ALSO open sourced, and is now playable in the browser.
Now, the poser that YOU are attacking was commenting on THE ARTICLE, not the developers of this "open source project"
ILP is huge for both x86 and RISC based machines. There is no good reason NOT to execute the next instruction if its independent of the one currently being executed. It helps to have a compiler which is aware of ILP so that it minimizes sequential dependency chains within the machine code.
OoO on the other hand isnt that big of a deal, because as long as you have a compiler aware of ILP, the CPU really doesnt have to manage this task 'cept in cases of very mismatched instruction latencies. Most instructions have a 3 cycle latency on a modern processor (1 cycle startup, 1 cycle execute, 1 cycle retire).. some isntructions deviate considerably (division instructions, for instance)
Larrabee itself is sort of a hybrid of a pure SIMD and pure MIMD approach, with 16-floats per register (512 bit) and 8 to 16 cores, it is pretty much a balance between SIMD and MIMD.
Note that nobody is making pure MIMD processors that even comes close to competing with the raw processing ability of SIMD, where the only non-SIMD setups that can compete are large computer clusters costing way more than Larrabee will.
This is the same road the video card makers are taking, but they lack some of the advantages that Intel has. Nobody has better fab's than Intel, and only Intel or AMD can eventualy integrate this thing on-die with a regular x86-based desktop CPU, which is why nVidia is in very serious trouble.
Aside from all the other reasons given in reply to you...
Why invest time in a feature that only a small portion of users would ever even notice or care about, where even the lack of this feature isnt even a problem?
I suspect that you never even noticed that this feature wasnt there.
..and some of them dont even do it for gaming, but instead for GPGPU
This is a real market, and as it matures the average joe will find that it offers things that they want as well.
The fact is that as long as even a small market exists, that market can expand under its own momentum to fill roles that cannot be anticipated.
I certainly wasn't thinking that there was a market for hardware accelerated graphics 20 years ago, yet I'm sure to make sure thats in the system I build today.
I certainly wasn't thinking about multi-core computers 20 years ago, yet I wouldn't buy anything less than a quad core today.
I certainly wasn't thinking about going from 20-bit to 32 -bit to 64-bit addressing 20 years ago.. I was happy with 640K and some bios above that, yet today if I build a system its going to have at least 8 gigs of memory.
I couldnt even dream of filling the 40meg hard drive I got with my first 386, yet now I am wondering if I should clean up my 500GB drive or simply buy a new 1TB drive to slot right next to it.
Yeah.. people weren't asking for those kinds of products either.. now we want them because those pesky unpredictable uses that come up that are actualy attractive to us.
If by pointing to Mann's reconstruction methods you mean to imply Mann, et al's hockey stick was debunked you are simply wrong...
I said exactly what I meant to say, and you are now trying to argue against something I didnt say, which is fine as long as you don't attribute your straw man arguement it to me.
Now.. don't attribute your strawman arguement to me. OK?
Furthermore the minor problems they did point out were adressed by Mann in a later publication in Science which you can look up yourself, this is how science works, no?
I am arguing that the veracity of the current peer review process in this field is so lacking that you do not get to appeal to its authority, that these climate experts have been known for a fact (which you admit) to use faulty statistical methods which slip right by the peer review process that you appealed to.
You don't get to use the "published in Nature" arguement as valid for their statistical value, since as I pointed out, experts in statistics do not do any reviewing of these papers prior to them being published.
This is quite simple.
Accept it, reject it.. I dont really care.. but do not reply with strawman arguements that you attribute to me as if you have some sort of refutation for my actual argument, when you apparently and obviously do not.
The same NASA that went 7 years without ever noticing a problem with their methodology that was detectable with an open source statistics package?
The discovery of the real problem was made almost immediately after NASA GISS finally revealed their methods for public scrutiny.
The same qualities that makes open source good are the same reasons that all of these scientists should open up their work. We re talking about publicly funded science here.. its not supposed to be secret.
They dont get a free pass just because they are NASA, especialy because they've fucked it up before.
One of the problems is that the peer reviewer is supposed to be an expert in the papers field (ex: climate science), rather than the methodology used (ex: statistics)
A popular example is Mann's flawed implementation of Principle Component Analysis, peer reviewed and then published by one of the very same journals that you are trying to use for your arguement-from-authority fallacy.
Thats from one of the journals you cited, so you trust that it is an accurate summary of what he did, right?
I certainly do not think that an expert in EARTH SCIENCES should be doing that stuff without supervision from an expert in what he is actualy doing.
..and as it turns out, he screwed it all up fairly badly but got published anyways.
As far as that ice data... here we have an error margin thats over 50% of the magnitude of the estimated anomaly, and thats assuming they did things right to begin with.
I still don't see evidence that a statistics expert was involved. If you have some, please enlighten.
The researchers found Antarctica's ice sheet decreased by 152 (plus or minus 80) cubic kilometers of...
An error margin greater than 50%? Presuming that this is based on a typical 3 standard deviations...
...the chebyshev limit says there is still a whopping 11% chance that the actual value is outside the range...
I don't see any statistics experts mentioned in that link, so I gotta assume that we cannot expect a normally distributed error, that in fact they have no idea what the distribution might be.
Going further, skype is proving that the data service AT&T is selling works out to a better deal than the voice service they are selling when used for the same purpose, which leads right to..
AT&T (and the rest) have either been way over-charging for their voice service for a very long time, or that they are way undercharging (taking a loss) on their data service.
Personally I think its quite clear that they are overcharging. Evidence of this are the very inexpensive prepaid voice-only services such as Tracfone, which seems to make a profit even though they arent raking in subscription money like the big boys are yet are using the very same cell towers (as if the tower owners would be taking a loss on the usage.)
So which field exactly is it that you must master in order to be a creadible expert on climate?
You claim its not physics, and it sure as hell isnt statistics (since NO climate scientist that I am aware of is an expert statistician).. so what is it?
The biggest attention getting papers are all using highly advanced (ie: hard to apply properly) statistical methods, which SHOULD be performed by experts in STATISTICS.
So here we are, letting novice statisticians convince us all that global warming is (A) real (B) man made, and (C) a threat...
...based on their practice in a field that they are not experts in.
Didn't they teach you in school that elements can be converted to other elements?
Carbon, for example, can be created from the triple collision of helium nuclei. Carbon can also be destroyed, converting it to nitrogen.
I suggest that you brush up on the concept of fusion, which will illustrate why you are terribly terribly ignorant and stuck on a concept (that "matter cannot be created or destroyed") that simply cannot be applied the way you are trying to apply it.
The real problem with his view is the completely fuzzy dividing line between "data" and "software"
He makes a strong arguement in the case of binaries.. open source means that you can rebuild those binaries, making changes if you so desire..
..but at what point does a scriptish language become a platform for programs?
Is CSS considered software? How about HTML?
The fact is that JavaScript is essentialy "plain view" open source even though it may be proprietary. Yeah, that source is not necessarily GPL or whatever other licensing flavor tweaks your twiddle..
It seems to me that Richards ethic is going too far because there is no simple dividing line.. surely he does not propose that all data must be "free", right? That he isnt suggesting that a pure GNU system would not ever interact with a non-pure data, such as an encryption key, right?
How far down the slope will his reasoning eventualy slip?
Slashdot Mirror
Have you tried other enemas?
Shhh.. dont tell anyone.
It is the most widely used CPU in the world, but has never had any significant market share in the general purpose computing segments.
Its a processor for embedded devices, pure and simple.
Honestly, the average consumer isnt changing the OS that comes with it.
Be rational. I know that its tricky with the religion and all.. but you will be better off for it.
Do you seriously think that earthquakes wouldnt cause of cracks in ice?
As far as weathering effects in the Antarctic, precipation averages about 16.5cm annual, making it the largest desert in the world.
The surface area of Antarctica is 14,200,000/km^2, so even though it is a desert, the amount of ice which would accumulate if there were no loss is staggering. This particular loss isn't very significant, but makes for a great story due to its surface area being equatable to something (ex: The state of RI) which is considered "large."
What is important is the volume of ice lost by this event, not the area of ice lost. All of these news agencies have latched onto the area of ice lost, not its volume, and that is in part due to the scientists themselves wrapping this up as dramatic.
Dont forget that once you form a corporation, personal responsibility is very very limited. If this company isnt making much money, then it doesn't have much to lose.
I don't think he should "get a lawyer"
He should contract a law firm. This is too big to cheap out on, and a diverse set of skills will be needed to maximize the severe punishment that can potentialy be levied against these guys.
You don't want the countersuit to be a single-shot weapon. You want it to be an agonizing series of laser guided charges that simply wont go away no matter how this company tries to weasel out of the situation they have put themselves in.
..and of course you realize that the game engine this is based upon was originally a proprietary one, right?
They have risen to the level of using a proprietary engine, QUAKE ONE, which was open sourced NINE YEARS AGO.
This doesnt say anything good about this case of open source gaming at all, especialy since QUAKE THREE was ALSO open sourced, and is now playable in the browser.
Now, the poser that YOU are attacking was commenting on THE ARTICLE, not the developers of this "open source project"
Give up your religion, man.
Let me tell you this one more time.
I meant exactly what I said. Things that I didn't say (all of your "I assume.." and "If you mean.." and whatnot) are not what I meant.
Its simple. Feel free to define your own complicated scanerios in your own words and then shoot them down, just don't attribute them to me. OK?
ILP is huge for both x86 and RISC based machines. There is no good reason NOT to execute the next instruction if its independent of the one currently being executed. It helps to have a compiler which is aware of ILP so that it minimizes sequential dependency chains within the machine code.
.. some isntructions deviate considerably (division instructions, for instance)
OoO on the other hand isnt that big of a deal, because as long as you have a compiler aware of ILP, the CPU really doesnt have to manage this task 'cept in cases of very mismatched instruction latencies. Most instructions have a 3 cycle latency on a modern processor (1 cycle startup, 1 cycle execute, 1 cycle retire)
Larrabee itself is sort of a hybrid of a pure SIMD and pure MIMD approach, with 16-floats per register (512 bit) and 8 to 16 cores, it is pretty much a balance between SIMD and MIMD.
Note that nobody is making pure MIMD processors that even comes close to competing with the raw processing ability of SIMD, where the only non-SIMD setups that can compete are large computer clusters costing way more than Larrabee will.
This is the same road the video card makers are taking, but they lack some of the advantages that Intel has. Nobody has better fab's than Intel, and only Intel or AMD can eventualy integrate this thing on-die with a regular x86-based desktop CPU, which is why nVidia is in very serious trouble.
Aside from all the other reasons given in reply to you...
Why invest time in a feature that only a small portion of users would ever even notice or care about, where even the lack of this feature isnt even a problem?
I suspect that you never even noticed that this feature wasnt there.
Some people buy 300watt video cards..
..and some of them dont even do it for gaming, but instead for GPGPU
This is a real market, and as it matures the average joe will find that it offers things that they want as well.
The fact is that as long as even a small market exists, that market can expand under its own momentum to fill roles that cannot be anticipated.
I certainly wasn't thinking that there was a market for hardware accelerated graphics 20 years ago, yet I'm sure to make sure thats in the system I build today.
I certainly wasn't thinking about multi-core computers 20 years ago, yet I wouldn't buy anything less than a quad core today.
I certainly wasn't thinking about going from 20-bit to 32 -bit to 64-bit addressing 20 years ago.. I was happy with 640K and some bios above that, yet today if I build a system its going to have at least 8 gigs of memory.
I couldnt even dream of filling the 40meg hard drive I got with my first 386, yet now I am wondering if I should clean up my 500GB drive or simply buy a new 1TB drive to slot right next to it.
Yeah.. people weren't asking for those kinds of products either.. now we want them because those pesky unpredictable uses that come up that are actualy attractive to us.
If by pointing to Mann's reconstruction methods you mean to imply Mann, et al's hockey stick was debunked you are simply wrong...
I said exactly what I meant to say, and you are now trying to argue against something I didnt say, which is fine as long as you don't attribute your straw man arguement it to me.
Now.. don't attribute your strawman arguement to me. OK?
Furthermore the minor problems they did point out were adressed by Mann in a later publication in Science which you can look up yourself, this is how science works, no?
I am arguing that the veracity of the current peer review process in this field is so lacking that you do not get to appeal to its authority, that these climate experts have been known for a fact (which you admit) to use faulty statistical methods which slip right by the peer review process that you appealed to.
You don't get to use the "published in Nature" arguement as valid for their statistical value, since as I pointed out, experts in statistics do not do any reviewing of these papers prior to them being published.
This is quite simple.
Accept it, reject it.. I dont really care.. but do not reply with strawman arguements that you attribute to me as if you have some sort of refutation for my actual argument, when you apparently and obviously do not.
NASA?
The same NASA that went 7 years without ever noticing a problem with their methodology that was detectable with an open source statistics package?
The discovery of the real problem was made almost immediately after NASA GISS finally revealed their methods for public scrutiny.
The same qualities that makes open source good are the same reasons that all of these scientists should open up their work. We re talking about publicly funded science here.. its not supposed to be secret.
They dont get a free pass just because they are NASA, especialy because they've fucked it up before.
One of the problems is that the peer reviewer is supposed to be an expert in the papers field (ex: climate science), rather than the methodology used (ex: statistics)
..and as it turns out, he screwed it all up fairly badly but got published anyways.
A popular example is Mann's flawed implementation of Principle Component Analysis, peer reviewed and then published by one of the very same journals that you are trying to use for your arguement-from-authority fallacy.
Lets examine what Mann was doing:
AlgorihmDescription.txt
Thats from one of the journals you cited, so you trust that it is an accurate summary of what he did, right?
I certainly do not think that an expert in EARTH SCIENCES should be doing that stuff without supervision from an expert in what he is actualy doing.
As far as that ice data... here we have an error margin thats over 50% of the magnitude of the estimated anomaly, and thats assuming they did things right to begin with.
I still don't see evidence that a statistics expert was involved. If you have some, please enlighten.
The researchers found Antarctica's ice sheet decreased by 152 (plus or minus 80) cubic kilometers of...
An error margin greater than 50%? Presuming that this is based on a typical 3 standard deviations...
...the chebyshev limit says there is still a whopping 11% chance that the actual value is outside the range...
I don't see any statistics experts mentioned in that link, so I gotta assume that we cannot expect a normally distributed error, that in fact they have no idea what the distribution might be.
Going further, skype is proving that the data service AT&T is selling works out to a better deal than the voice service they are selling when used for the same purpose, which leads right to..
AT&T (and the rest) have either been way over-charging for their voice service for a very long time, or that they are way undercharging (taking a loss) on their data service.
Personally I think its quite clear that they are overcharging. Evidence of this are the very inexpensive prepaid voice-only services such as Tracfone, which seems to make a profit even though they arent raking in subscription money like the big boys are yet are using the very same cell towers (as if the tower owners would be taking a loss on the usage.)
Seems to me that a random walk (err, random 'lets go and visit') would explain the balancing, even if you started with a 40:10 situation.
In the 50:0 situation, there is nobody to visit on the other side.
Thanks for reminding me how undescribably bad that movie was.
So which field exactly is it that you must master in order to be a creadible expert on climate?
.. so what is it?
You claim its not physics, and it sure as hell isnt statistics (since NO climate scientist that I am aware of is an expert statistician)
Here is an observation for you.
...based on their practice in a field that they are not experts in.
...you were saying?
The biggest attention getting papers are all using highly advanced (ie: hard to apply properly) statistical methods, which SHOULD be performed by experts in STATISTICS.
So here we are, letting novice statisticians convince us all that global warming is (A) real (B) man made, and (C) a threat...
I hinted at your declaration of facts which arent true.
You are now putting words in my mouth as a response to that...
I guess you did learn that you were wrong. Next you can learn how to gracefully admit when you were wrong.
Oh my...
Didn't they teach you in school that elements can be converted to other elements?
Carbon, for example, can be created from the triple collision of helium nuclei. Carbon can also be destroyed, converting it to nitrogen.
I suggest that you brush up on the concept of fusion, which will illustrate why you are terribly terribly ignorant and stuck on a concept (that "matter cannot be created or destroyed") that simply cannot be applied the way you are trying to apply it.
Re: Not all costs to the environment can be fixed by throwing more money at them.
Yeah? Name one.
I, as a crazy hobby programmer, would use it for a 1.5Tbit lookup table.
Thats right, a 1.5Tbit lookup table.
Just think of the possibilities!
The real problem with his view is the completely fuzzy dividing line between "data" and "software"
..but at what point does a scriptish language become a platform for programs?
He makes a strong arguement in the case of binaries.. open source means that you can rebuild those binaries, making changes if you so desire..
Is CSS considered software? How about HTML?
The fact is that JavaScript is essentialy "plain view" open source even though it may be proprietary. Yeah, that source is not necessarily GPL or whatever other licensing flavor tweaks your twiddle..
It seems to me that Richards ethic is going too far because there is no simple dividing line.. surely he does not propose that all data must be "free", right? That he isnt suggesting that a pure GNU system would not ever interact with a non-pure data, such as an encryption key, right?
How far down the slope will his reasoning eventualy slip?