Did Ken Perlin not come through the 8-bit era? "Everybody knows" you can fit the entire English alphabet comfortably into a 3 pixel wide by 5 pixel high monospaced grid, it's been done hundreds of times. (Proportional can be even smaller, of course.)
It's rather shocking that one company can actually *source* 75 million gallons of chicken fat per year in one country. How many billion chickens have to be slaughtered to make 75 million gallons of chicken fat?
Breaking News: 'Polling Predictions' Could Skew Polling Results. Don't forget the Heisenberg Principle here, you can't observe something without changing it. I wish they would let elections run their course without publishing any stats.
Fortress allows you to code in UTF-8. However it has a multi-char ASCII equivalent for every Unicode mathematical symbol that you can use, so there is a bijective map between the Unicode and ASCII versions of the source, and you can view/edit in either. That is the only acceptable way to advocate using Unicode anywhere in programming source other than string constants. Programming languages that use ASCII have done well over those that don't, for the same reason that Unicode has done well over binary formats.
Very informative. Why does fsync not synchronously flush out *only* the data dirtied by the current process, rather than all buffers on all file systems dirtied by any process? It seems like a bad idea for a non-root thread to have so much power over how smoothly the rest of the system runs.
Though you're the CPU scheduler maintainer not the IO scheduler maintainer, can you shed light on two other situations please?
(1) As soon as RAM is exhausted and the kernel starts swapping out to disk, the desktop experience is severely impacted (and immediately so). However I would have thought that most if not all code that is actually being used by the core desktop (working memory and code for X, gtk etc.) would be marked as recently used, i.e. the pages being swapped out should be at the end of some LRU list? In other words, is the slowness due to the fact that *other* non-core-desktop-related pages are being swapped in and out, and the IO due to swapping is causing all processes to slow down, or is it because the pages containing running desktop code is constantly being swapped in and out? Is it possible to create a better desktop experience by mapping certain memory areas "core to user experience" and forbidding the VM from ousting them, so that for example an individual app might slow down if it suddenly takes up a lot of RAM and causes swapping, but interacting with the rest of the desktop is still smooth?
(2) When doing a "dd if=img.iso of=/dev/sdb" to copy an.iso file to a USB stick, once the buffers are full the entire desktop comes grinding to a halt for several seconds, then frees up for a second or two once it is no longer blocking on write, then locks up again for several seconds, and so on. Why would a blocked write lock up the whole desktop?
He should have thrown it over the White House fence and watched as both White House security forces as well as the FBI forces that have been tracking him spring out from all their hiding places -- and gotten it all on camera.
"Was that a bomb? On the ground, now, hands where I can see them!"
"What? Just returning Federal property to its owner!"
Craig Mundie, Chief Research and Strategy Officer for Microsoft, gave a demo today at the MIT CSAIL Dertouzos Lecture, and faked most of the demo too.
During the only part he didn't fake, he showed the debug output of the controller on one screen (skeletal model superimposed over body silhouette), and the movement was *extremely* noisy and jumpy, easily fooled unless all your limbs are visible in the silhouette. Games are going to have to do a ton of event noise smoothing to get anything useful.
It also couldn't distinguish between when the body was facing forward and backwards. (Mundie did say though that it could track 4 people but had problems with occlusions.)
And Mundie kept referring to how they're at the limit of what's possible with today's computing power. Except he had no concept of what is possible given certain CPU speeds. For example, in one demo Kinect motion was decoded on an X-Box and then fed to a PC where his motion allowed him to rotate a 3D model with his hands. The model was an exploded view of a car's parts, consisting of (in my estimate) maybe 100k polygons rendered at 60FPS. He said this visualization taxed to the limit the system it was running on -- an 8-core Xeon system with the fastest nVidia card on the market. Rly??
...and head-to-tail collisions would be a real problem. And if you are moving in a chain of schweebs, there will be the inevitable lazy guy somewhere in the chain not pedaling or pretending to pedal, so someone else will do all the work. And if the schweeb capsules are publicly shared not privately owned, they'll get really, really gross and sweaty. And if everybody is commuting in one direction and they are a shared public resource, all the schweebs will end up at one end of the line...
However I'm guessing they'll eventually design passing lanes to partially solve the slow/lazy rider problems, and you'll be able to steer into passing lanes (if you can even see where you're going from that body position...) The really smart (and IMO only smart) aspects of schweeb are (1) it should reduce or eliminate biking deaths, because it lifts you off the road surface away from cars (and you can no longer fall off your bike; hopefully your schweeb won't fall off the track either) -- and (2) it potentially dramatically reduces the number of intersections by lifting the route graph off the 2D plane into 3D space. This solves the planar graph embedding problem (which forbids K_5 and K_{3,3} subgraphs) and allows a much denser point-to-point connectivity graph without having to wait at the schweeb equivalent of a red light. (This is the real reason we all need flying cars, incidentally.)
Personally I think that what killed it was that I should have transparently been able to see my gmail inside wave. Requiring a separate window guarantees that I wouldn't use it regularly.
Huh? Having its own inbox has yet to kill facebook, though I wish it would.
Wave more than justified its existence, and dropping it is not a net loss: GWT is a gazillion times faster and more featureful today because Wave caused it to crack at the seams, repeatedly, in every way imaginable. See the Google I/O talks from 2009 for more info.
Also I suspect that killing off wave had something to do with acquiring the Etherpad team. Etherpad is Wave done right. (Except for the hand-crafted Javascript part...)
It's been said, but duh. The contacts are hardly the problem. All that's needed is one of these per battery: http://en.wikipedia.org/wiki/Diode_bridge The worst part is that people will start putting batteries into other products the wrong way around and they'll think they're broken.
Are they *trying* to punish companies for doing the write thing and being honest? The only thing the uproar over Google's honest and unnecessary disclosure (they could have just deleted the data) will do is to further reinforce the Pavlovian conditioning of corporate dishonesty.
He's probably talking about the upcoming dual-core Snapdragon processor, underclocked from 1.5GHz to 1GHz to save on battery. I really hope the industry doesn't start multiplying GHz by core counts.
Of course recovering the oil for sale is their top priority, they are businessmen. However BP profiting from the well at this point would be extremely negatively received by the public.
They already sawed off a big chunk of the drilling pipe, and they could do so again if needed. And I thought they already tried pumping kill mud into the hole?
We still need to plug the hole... and BP seem completely incapable of doing so themselves.
This video on YouTube shows a pretty promising solution: http://www.youtube.com/watch?v=CDAkIU6zYkY
Slashdot Mirror
Yeah but it's a royal pain to write anything nontrivial in straight OpenGL.
In case anyone missed it: reintroduction program
Did Ken Perlin not come through the 8-bit era? "Everybody knows" you can fit the entire English alphabet comfortably into a 3 pixel wide by 5 pixel high monospaced grid, it's been done hundreds of times. (Proportional can be even smaller, of course.)
It's rather shocking that one company can actually *source* 75 million gallons of chicken fat per year in one country. How many billion chickens have to be slaughtered to make 75 million gallons of chicken fat?
...if so, then banks have a very, very big problem.
Breaking News: 'Polling Predictions' Could Skew Polling Results. Don't forget the Heisenberg Principle here, you can't observe something without changing it. I wish they would let elections run their course without publishing any stats.
Sorry, meant to say "for the same reason XML [not Unicode] has done well over binary formats".
Fortress allows you to code in UTF-8. However it has a multi-char ASCII equivalent for every Unicode mathematical symbol that you can use, so there is a bijective map between the Unicode and ASCII versions of the source, and you can view/edit in either. That is the only acceptable way to advocate using Unicode anywhere in programming source other than string constants. Programming languages that use ASCII have done well over those that don't, for the same reason that Unicode has done well over binary formats.
Let's go back to APL!
Very informative. Why does fsync not synchronously flush out *only* the data dirtied by the current process, rather than all buffers on all file systems dirtied by any process? It seems like a bad idea for a non-root thread to have so much power over how smoothly the rest of the system runs.
Thanks Ingo for the detailed explanation.
.iso file to a USB stick, once the buffers are full the entire desktop comes grinding to a halt for several seconds, then frees up for a second or two once it is no longer blocking on write, then locks up again for several seconds, and so on. Why would a blocked write lock up the whole desktop?
Though you're the CPU scheduler maintainer not the IO scheduler maintainer, can you shed light on two other situations please?
(1) As soon as RAM is exhausted and the kernel starts swapping out to disk, the desktop experience is severely impacted (and immediately so). However I would have thought that most if not all code that is actually being used by the core desktop (working memory and code for X, gtk etc.) would be marked as recently used, i.e. the pages being swapped out should be at the end of some LRU list? In other words, is the slowness due to the fact that *other* non-core-desktop-related pages are being swapped in and out, and the IO due to swapping is causing all processes to slow down, or is it because the pages containing running desktop code is constantly being swapped in and out? Is it possible to create a better desktop experience by mapping certain memory areas "core to user experience" and forbidding the VM from ousting them, so that for example an individual app might slow down if it suddenly takes up a lot of RAM and causes swapping, but interacting with the rest of the desktop is still smooth?
(2) When doing a "dd if=img.iso of=/dev/sdb" to copy an
He should have thrown it over the White House fence and watched as both White House security forces as well as the FBI forces that have been tracking him spring out from all their hiding places -- and gotten it all on camera.
"Was that a bomb? On the ground, now, hands where I can see them!"
"What? Just returning Federal property to its owner!"
Craig Mundie, Chief Research and Strategy Officer for Microsoft, gave a demo today at the MIT CSAIL Dertouzos Lecture, and faked most of the demo too.
During the only part he didn't fake, he showed the debug output of the controller on one screen (skeletal model superimposed over body silhouette), and the movement was *extremely* noisy and jumpy, easily fooled unless all your limbs are visible in the silhouette. Games are going to have to do a ton of event noise smoothing to get anything useful.
It also couldn't distinguish between when the body was facing forward and backwards. (Mundie did say though that it could track 4 people but had problems with occlusions.)
And Mundie kept referring to how they're at the limit of what's possible with today's computing power. Except he had no concept of what is possible given certain CPU speeds. For example, in one demo Kinect motion was decoded on an X-Box and then fed to a PC where his motion allowed him to rotate a 3D model with his hands. The model was an exploded view of a car's parts, consisting of (in my estimate) maybe 100k polygons rendered at 60FPS. He said this visualization taxed to the limit the system it was running on -- an 8-core Xeon system with the fastest nVidia card on the market. Rly??
...and head-to-tail collisions would be a real problem. And if you are moving in a chain of schweebs, there will be the inevitable lazy guy somewhere in the chain not pedaling or pretending to pedal, so someone else will do all the work. And if the schweeb capsules are publicly shared not privately owned, they'll get really, really gross and sweaty. And if everybody is commuting in one direction and they are a shared public resource, all the schweebs will end up at one end of the line...
However I'm guessing they'll eventually design passing lanes to partially solve the slow/lazy rider problems, and you'll be able to steer into passing lanes (if you can even see where you're going from that body position...) The really smart (and IMO only smart) aspects of schweeb are (1) it should reduce or eliminate biking deaths, because it lifts you off the road surface away from cars (and you can no longer fall off your bike; hopefully your schweeb won't fall off the track either) -- and (2) it potentially dramatically reduces the number of intersections by lifting the route graph off the 2D plane into 3D space. This solves the planar graph embedding problem (which forbids K_5 and K_{3,3} subgraphs) and allows a much denser point-to-point connectivity graph without having to wait at the schweeb equivalent of a red light. (This is the real reason we all need flying cars, incidentally.)
Why using your email password on any other account is particularly bad, and what to do about it if your password is stolen
And we're supposed to believe that Google accidentally recording a few unencrypted Wifi data packets in their Street View rig is bad?
Personally I think that what killed it was that I should have transparently been able to see my gmail inside wave. Requiring a separate window guarantees that I wouldn't use it regularly.
Huh? Having its own inbox has yet to kill facebook, though I wish it would.
Wave more than justified its existence, and dropping it is not a net loss: GWT is a gazillion times faster and more featureful today because Wave caused it to crack at the seams, repeatedly, in every way imaginable. See the Google I/O talks from 2009 for more info. Also I suspect that killing off wave had something to do with acquiring the Etherpad team. Etherpad is Wave done right. (Except for the hand-crafted Javascript part...)
It's been said, but duh. The contacts are hardly the problem. All that's needed is one of these per battery: http://en.wikipedia.org/wiki/Diode_bridge The worst part is that people will start putting batteries into other products the wrong way around and they'll think they're broken.
Are they *trying* to punish companies for doing the write thing and being honest? The only thing the uproar over Google's honest and unnecessary disclosure (they could have just deleted the data) will do is to further reinforce the Pavlovian conditioning of corporate dishonesty.
He's probably talking about the upcoming dual-core Snapdragon processor, underclocked from 1.5GHz to 1GHz to save on battery. I really hope the industry doesn't start multiplying GHz by core counts.
end of message
Of course recovering the oil for sale is their top priority, they are businessmen. However BP profiting from the well at this point would be extremely negatively received by the public.
They already sawed off a big chunk of the drilling pipe, and they could do so again if needed. And I thought they already tried pumping kill mud into the hole?
We still need to plug the hole... and BP seem completely incapable of doing so themselves. This video on YouTube shows a pretty promising solution: http://www.youtube.com/watch?v=CDAkIU6zYkY