The main reason for needing flashblock in FF is that the flash plugin tends to lock up the whole browser on a regular basis, so you want to run the plugin only when needed to minimize the probability of that happening. Chrome runs the plugin in its own process, so the probability of Flash locking up the browser is zero to begin with. Ever since I started using Chrome, I have an icon in my freaking task bar that runs ps axu | grep libflashplayer | grep $LOGNAME | grep -v grep | awk '{print $2}' | xargs kill. Apart from Chrome itself, that icon is the single greatest productivity booster I've installed in years.
So, they say that the system should be deployed at a single location and should be able to reach any point on the surface of the earth within less than an hour. That means that the average velocity of that cruise missile or whatever it is must be AT LEAST 20,000 km/h = 5.55 km/s, i.e. 70% of orbital speed. And they want to achieve this inside the atmosphere. What kind of engine would be able to do that? Nuclear ramjets? Or what? Even just the energy consumption due to atmospheric drag would be enormous.
Furthermore, the proposed example mission scenario (taking out Bin Laden) seems unrealistic. Bin Laden's location is known to a few 100 km or so. If the Pentagon learns that Bin Laden hides in cave XY in northern Pakistan, why would they start a cruise missile at Vandenberg AFB rather than from an aircraft carrier in the Persian gulf or some army base in Afghanistan?
(repost -- the first time I posted this, the article was visible only for me when I was logged in)
Immediate mode requires at least as many (usually more 3 times more) calls as you have verticies in your model, during which the GPU is wasting time, and the driver is doing complex things to pack data into buffers in graphics memory.
Meanwhile, vertex arrays require a single upload of a constant array to graphics memory, which happens quickly as a single memcpy, and then frees the graphics card to get on with it.
Unless I'm very much mistaken, vertex arrays (in client memory) is what the driver does internally to implement immediate mode anyway. I.e. all your glVertex etc. calls are collected in an array, which is sent to the GPU on glEnd(). So, if you just do all that yourself to avoid using immediate mode, you won't gain any performance. What IS faster is using vertex buffer objects, because they reside in graphics memory and thus DON'T "require a single upload of a constant array to graphics memory" -- which might otherwise become a major bottleneck if you're rendering the same geometry over and over again.
Immediate mode requires at least as many (usually more 3 times more) calls as you have verticies in your model, during which the GPU is wasting time, and the driver is doing complex things to pack data into buffers in graphics memory.
Meanwhile, vertex arrays require a single upload of a constant array to graphics memory, which happens quickly as a single memcpy, and then frees the graphics card to get on with it.
Unless I'm very much mistaken, vertex arrays (in client memory) is what the driver does internally to implement immediate mode anyway. I.e. all your glVertex etc. calls are collected in an array, which is sent to the GPU on glEnd(). So, if you just do all that yourself to avoid using immediate mode, you won't gain any performance. What IS faster is using vertex buffer objects, because they reside in graphics memory and thus DON'T "require a single upload of a constant array to graphics memory" -- which might otherwise become a major bottleneck if you're rendering the same geometry over and over again.
That animation isn't based on a measurement though, it's based on a computer simulation with debatable accuracy (whose initial conditions were based on some measurements, again with debatable accuracy).
Riiiiight, so the "rules of engagement" say that it's completely OK to shell all cars on Baghdad's streets except those that are "marked". Give me a break.
It is not uncommon for the enemy to drive up in vans and jump out.
Huh? The guy "jumped out" because he was going to help a wounded person. He didn't do anything else. If five guys with weapons had jumped out of that van, then go ahead and engage them, but that guy did not do any of those things. If you think that it's still OK to open fire at the van on the grounds that vans have been used in the past for transporting terrorists, then by that logic you have to fire at all vans in Baghdad, or to make things a little easier for you, just carpet-bomb the whole place to get rid of all terrorists in there at once.
If it's mmap()ed in from a file, another process can mmap() the same file and it will only physically use up the memory once (in the buffer cache in the kernel), and it will even only use up parts of the mmap()ed area on an on-demand basis. If it's an anonymous mapping (i.e. a memory allocation not backed by a file), it'll still only start physically mapping the pages that are being used, and allow processes to reserve a total amount of memory that's higher than the physical memory size by overcommitting memory and/or resorting to swap space when necessary.
I think he was talking about namespace aliasing, so you could e.g. use both org.foo.Table and org.bar.Table in the same source file without having to write out the fully-qualified name of either of them.
They've never done it for under $1000 though, which this guy did.
You have to add labor costs to that, though. The guy probably invested many hours of work into this without getting paid. If some NASA employee did the same thing, he/she would receive a paycheck that would amount to much more than $1000.
As a non-US citizen, I wonder what's gonna happen now to all those heart-wrenching US medical dramas if every kindhearted-yet-ill-and-uninsured Ghetto kid that needs a new aortic valve can just get one without any tough ethical/legal/financial dilemma. That's gonna be horrible!
There is a prohibition of mutilating a corpse, but saving a life trumps that.
So, corneal or kidney transplants from a corpse would still be prohibited because the receiver's life doesn't depend on it? That still doesn't sound very logical to me.
The last thing I heard was that these Toyota problems happen in the US only, or at least have devastating consequences in the US only. There are reports of more than 50 deaths that are supposed to have been caused by these "uncontrollable acceleration" issues in the US, while the combined death toll of these problems in all other countries is zero. Now, I'm generally not a supporter of conspiracy theories or speculative explanations that suggest things like some mild mass psychosis (culturally induced, irrational fear of evil engines going berserk, maybe?), but if the above numbers are correct, they're far beyond the realm of mere statistical fluctuations.
It's Dijkstra we're talking about. And Dijkstra does have quite a few genuinely useful and lasting achievements speaking for him (semaphores and such), while Wirth was indeed, more or less, a quiche-eater.
Slashdot Mirror
http://developer.apple.com/safaridemos/ doesn't work, too. shows the same "you have to get safari".
Not in recent Chrome builds. The demos all work here without problems (Chrome 5.0.375.55 beta / Linux).
How about putting an existing warhead into a pressure vessel?
Thanks, I actually didn't know about that. Very nice.
The main reason for needing flashblock in FF is that the flash plugin tends to lock up the whole browser on a regular basis, so you want to run the plugin only when needed to minimize the probability of that happening. Chrome runs the plugin in its own process, so the probability of Flash locking up the browser is zero to begin with. Ever since I started using Chrome, I have an icon in my freaking task bar that runs ps axu | grep libflashplayer | grep $LOGNAME | grep -v grep | awk '{print $2}' | xargs kill. Apart from Chrome itself, that icon is the single greatest productivity booster I've installed in years.
Furthermore, the proposed example mission scenario (taking out Bin Laden) seems unrealistic. Bin Laden's location is known to a few 100 km or so. If the Pentagon learns that Bin Laden hides in cave XY in northern Pakistan, why would they start a cruise missile at Vandenberg AFB rather than from an aircraft carrier in the Persian gulf or some army base in Afghanistan?
90 hours/week... This guy is insane and should get a better job.
Maybe he already has an incredible job, which is the reason he works 90 hours/week in the first place.
Immediate mode requires at least as many (usually more 3 times more) calls as you have verticies in your model, during which the GPU is wasting time, and the driver is doing complex things to pack data into buffers in graphics memory.
Meanwhile, vertex arrays require a single upload of a constant array to graphics memory, which happens quickly as a single memcpy, and then frees the graphics card to get on with it.
Unless I'm very much mistaken, vertex arrays (in client memory) is what the driver does internally to implement immediate mode anyway. I.e. all your glVertex etc. calls are collected in an array, which is sent to the GPU on glEnd(). So, if you just do all that yourself to avoid using immediate mode, you won't gain any performance. What IS faster is using vertex buffer objects, because they reside in graphics memory and thus DON'T "require a single upload of a constant array to graphics memory" -- which might otherwise become a major bottleneck if you're rendering the same geometry over and over again.
Immediate mode requires at least as many (usually more 3 times more) calls as you have verticies in your model, during which the GPU is wasting time, and the driver is doing complex things to pack data into buffers in graphics memory.
Meanwhile, vertex arrays require a single upload of a constant array to graphics memory, which happens quickly as a single memcpy, and then frees the graphics card to get on with it.
Unless I'm very much mistaken, vertex arrays (in client memory) is what the driver does internally to implement immediate mode anyway. I.e. all your glVertex etc. calls are collected in an array, which is sent to the GPU on glEnd(). So, if you just do all that yourself to avoid using immediate mode, you won't gain any performance. What IS faster is using vertex buffer objects, because they reside in graphics memory and thus DON'T "require a single upload of a constant array to graphics memory" -- which might otherwise become a major bottleneck if you're rendering the same geometry over and over again.
That animation isn't based on a measurement though, it's based on a computer simulation with debatable accuracy (whose initial conditions were based on some measurements, again with debatable accuracy).
The internet has progressed nicely in the 00s.
Riiiiight, so the "rules of engagement" say that it's completely OK to shell all cars on Baghdad's streets except those that are "marked". Give me a break.
They are not collecting weapons. The gunner says at one point that they might, but they clearly don't.
Huh? The guy "jumped out" because he was going to help a wounded person. He didn't do anything else. If five guys with weapons had jumped out of that van, then go ahead and engage them, but that guy did not do any of those things. If you think that it's still OK to open fire at the van on the grounds that vans have been used in the past for transporting terrorists, then by that logic you have to fire at all vans in Baghdad, or to make things a little easier for you, just carpet-bomb the whole place to get rid of all terrorists in there at once.
If it's mmap()ed in from a file, another process can mmap() the same file and it will only physically use up the memory once (in the buffer cache in the kernel), and it will even only use up parts of the mmap()ed area on an on-demand basis. If it's an anonymous mapping (i.e. a memory allocation not backed by a file), it'll still only start physically mapping the pages that are being used, and allow processes to reserve a total amount of memory that's higher than the physical memory size by overcommitting memory and/or resorting to swap space when necessary.
I think he was talking about namespace aliasing, so you could e.g. use both org.foo.Table and org.bar.Table in the same source file without having to write out the fully-qualified name of either of them.
You have to add labor costs to that, though. The guy probably invested many hours of work into this without getting paid. If some NASA employee did the same thing, he/she would receive a paycheck that would amount to much more than $1000.
As a non-US citizen, I wonder what's gonna happen now to all those heart-wrenching US medical dramas if every kindhearted-yet-ill-and-uninsured Ghetto kid that needs a new aortic valve can just get one without any tough ethical/legal/financial dilemma. That's gonna be horrible!
The only one who believes that is Hasselhoff himself...
http://www.endlesssearch.co.uk/science_cosmicreligion.htm
So, corneal or kidney transplants from a corpse would still be prohibited because the receiver's life doesn't depend on it? That still doesn't sound very logical to me.
The last thing I heard was that these Toyota problems happen in the US only, or at least have devastating consequences in the US only. There are reports of more than 50 deaths that are supposed to have been caused by these "uncontrollable acceleration" issues in the US, while the combined death toll of these problems in all other countries is zero. Now, I'm generally not a supporter of conspiracy theories or speculative explanations that suggest things like some mild mass psychosis (culturally induced, irrational fear of evil engines going berserk, maybe?), but if the above numbers are correct, they're far beyond the realm of mere statistical fluctuations.
Drop shadows? I always thought they do that with fragment shaders.
It's Dijkstra we're talking about. And Dijkstra does have quite a few genuinely useful and lasting achievements speaking for him (semaphores and such), while Wirth was indeed, more or less, a quiche-eater.
I couldn't decide whether to mod this "funny" or "insightful".
Cheese in a vacuum should dehydrate / dry out and become pretty hard. I'm not sure what the descent engine would've done to it, though.