And, how do you intend to 'capture the exact GPS coordinates of caves', too?
Satellite pictures and radar. Maybe fly a little recon drone nearby. Totally doable.
How do you intend to tell such a cave with fighters in it from, say, a similar cave sheltering a herder? Or a family hiding from the American Devils?
I guess that depends on whether the planners care that there's a difference. Not that I support it, but one school of thought says "Kill them all. God will know his own."
[From a comment above]
A little troop of snipers can hold a valley against a company.
Unless that company has a radar that can track bullets, precisely locate the gun that fired them, and launch a little wire-guided missile that blows up a few meters above the shooter. The missiles would be pretty cheap in mass production, maybe $50k apiece. I don't know if that kind of system exists today, but it's just a matter of gluing together existing technology.
Or how about a laser system that illuminates everything within rifle range with a very bright, very sharp pulse. It detects anything that reflects: rifle scopes, binoculars, human eyes, horse eyes, etc. It then sends a focused high-energy pulse towards each reflection, burning out every retina and camera that was down-range. Unless the guerrillas can see through pinhole/slit glasses at night, they are pretty much fucked.
Those who think that technology is meaningless in a guerrilla war in the mountains obviously haven't been paying attention to modern technology. A guerrilla war against a large, *committed* first world nation is pretty much unwinnable, especially if the first world nation isn't worried about collateral damage.
That's an absolutely *massive* liability there, especially since it can be proved that the ISP was aware of the infected machines and did pretty much nothing to eliminate the problem.
The legal liability is minimal, since transport sellers have no statutory responsibility to be friendly. The world is *full* of cracked boxes that ISPs ignore because the compromisers aren't spewing lots of packets and the box owners keep paying. The real constraints for a massive worm like Nimda are providing an acceptable quality of service to customers, and minimizing upstream bandwidth costs. There simply isn't any money in having one third of your customers spray random packets as fast as they can.
Your incredulity shows your almost unspeakable insularity. Your use of a Unix-based, system-specific command to provide evidence against me is solid proof of that insularity.
I also design logic for FPGAs and design systems that incorporate microcontrollers. I am fully aware that there is a place for simple chips that pack a lot of bang for the buck, and that the software/firmware/logic for them is narrowly tailored for specific jobs.
So it is valid under some circumstances to say "an MMU is not good". However, Chuck Moore makes all sorts of statements like that one *as general laws*, which really squicks me. E.g., this page which says "The word */ multiplies by a ratio, with a double-length intermediate product. It eliminates the need for floating-point." Eliminates. Not "eliminates fp when you have a known and small dynamic range", but "eliminates". This has got to be one of the stupidest things I've ever seen.
Another example: This is Chuck "Yes, that's all it takes." Moore's IDE driver. Puh-leeze. It's the "Hello World" of IDE drivers. It does *nothing*. No DMA. No locking for running on SMP machines. No autodetection and adaptation to drive capabilities. No workarounds for chipset bugs. No blacklisting of drives that are buggy. If you want to play in the real world, you need code like this (the Linux IDE driver).
Look at this page where he says "With the huge RAM of modern computers, an operating system is no longer necessary, if it ever was." Idiocy, written by a person who has never designed a system of significant complexity. RAM has *nothing* to do with whether an operating system is needed: it has everything to do with complexity. E.g., the only sane way to use floppy, IDE, SCSI, flash, CD-ROM, and network-mounted drives in the same system is to have a generic drive-access layer. Then you'll add a generic removeable drive layer to support things like CD-ROMs and flash drives that might suddenly disappear. If you want to support tons of hardware and software, you have to have an operating system.
This page is *full* of idiocy. "Don't try for platform portability. Most platform differences concern hardware interfaces. These are intrinsically different. Any attempt to make
them appear the same achieves the lowest common denominator. That is, ignores the features that made the hardware attractive in the first
place." Except, of course, for the usual case where the hardware has an upgrade that is easy to turn on and use, and which hardware is available is not known until runtime.
When you have multiple processors for that cheap, it's far more secure to just set up a seperate processor.
Let's see, how many processes do I have running?
$psax|wc-l
82
So I'm supposed to have 82 processors on account of having 82 processes that might become runnable at any moment? What about shell servers with hundreds of simultaneous users? Lack of memory protection is just ludicrous, it's so not-real-world it isn't even funny.
And as far as being more secure, that's ludicrous too. A functioning MMU is just as secure.
Good point. Where I am in the midwest there tends not to be enough economy of scale to justify steam, and I tend to forget about it. Although anybody that uses forced-air heating, which I imagine includes a lot of finance and logistics companies with fancy thermostats, would be still be screwed.
Killing the power would have massive effects, but it wouldn't cause a lot of people to freeze to death, if that's what you were thinking.
Maybe not on a blow-up-the-World-Trade-Center scale, but still, killing a hundred or so old and poor people using a leaf blower and a bag of industrial lubricant is still a pretty good kill ratio, especially if the terrorist claimed responsibility and worked the public fear angle for maximum effect.
The real question is, "why aren't Japanese people trying to kill us? Why aren't German people trying to kill us? What did we do differently?
Anyone who can't answer that question should be removed from US policy-making. Just look at the Chinese! We blew up their embassy, and they nearly killed everyone on one of our military airplanes and afterwards hassled us mightily, and they're Most Favored Nation with no signs of that changing! On the one hand, we have the strong-willed nuclear-armed intellectual descendants of Chairman Mao who are on their way to becoming allies, and on the other hand we have (and this is only a guess based on the scanty information available) poorly-armed desert punks who would've nuked Manhattan if they could.
Unless we want repeat performances, we'd better figure out how to avoid instigating this kind of warfare in the future. The risk mitigation techniques needed are vastly different than for conventional warfare, and I fear that the old guard will get hundreds of thousands of Americans killed before they figure this (much like they did when they ignored air power in a previous war). And least the Enemy hit the Pentagon. Thank God for that. Even the most fossilized Cold Warrior has trouble ignoring an airplane hitting *his* building.
Very true, and, while I sincerely hope the U.S. identifies those responsible and narrowly targets any retaliation, many mid-East civilians also celebrated the attack...
As did many Americans celebrate the death of tens of thousands of Iraqi soldiers in their bunkers, after which they put "Nuke Iraq" bumper stickers on their cars (remember those?). Judge not lest you be judged.
Also worth noting that many mid-East leaders of all stripes and colours are denouncing this terrorist act. Don't paint all the mid-East with one brush. This terrorist attack was the action of a very, very small radical group that is roundly despised by many mid-East civilians.
Is it even certain yet that the attackers are Middle Eastern? After the Murrah building bombing in Oklahoma City there was considerable harrasment of Middle Eastern immigrants in the area, it being a forgone conclusion that those goddamn raghead extremist fureners did it. Which looked pretty stupid when it turned out to have been an American, and a former US soldier at that!
People, remember that there are plenty of violent, organized fools right here in the US: anarchists, cults, people who think the South Shall Rise Again, and so forth. It does look suspiciously like a foreign attack, but don't jump to conclusions. There will be a hell of a lot of evidence, from surveillance tapes at the airports where they embarked, to radio reports from the pilots, perhaps even cell-phone reports from airline passengers. FBI field agents were no doubt on their way to the airports within a couple of hours of the attacks. If these terrorists are as incompentent as terrorists usually are, there will be tons of good, solid evidence against them, and in the meantime it would be a Bad Thing to jump to any political conclusions.
As for the Palestinians and Iraqis celebrating: big deal. If California slid off into the sea, they'd celebrate. Condeming the US is a hobby/spectator sport for them.
After the Murrah Building was destroyed in Oklahoma City, and the wreckage cleared away, it was decided to leave the lot permanently vacant. While I understand the motivation for this decision, I respectfully disagree. There are those (few in number, I hope) who will always see the empty spot where the Murrah building once was as a monument to McVeigh, not to his victims. He left his mark on the city. There is no way that a rebuilt Murrah Building could be so interpreted.
You are not alone in feeling this way. There are many here in Oklahoma -- including me -- who also thought that the best monument to the victims would be a new federal building. Unfortunately the busybodies and bureaucrats have a way of dominating such things.
Yes. The United States has been pushing people around and generally being a nuisance to the Middle East for nigh on 50 years now. The US is very big, and very strong, and they are very small and very weak. We have B-52 Stratofortresses and they have rifles and Molotov cocktails.
In a power disparity like that, there's only one way for the little guy to fight: infrequent lightning attacks. These guys don't have the resources to carpet bomb Manhattan with daisy cutters: that would take tens of thousand of people and a very high industrial level. They *do* have the resources to send a dozen men to hijack a handful of planes. A dozen men, I might add, that are brainwashed, do not know where they were trained, and cannot identify any of the few dozen (perhaps up to few hundred) other people in the organization.
While eliminating various terrorist targets (HOPEFULLY the ones responsible) won't ensure our safety, it will make it clear to everyone that there is a price to pay for inflicting violence on our country.
And who were we to blow up to avenge the act of Timothy McVeigh? Detroit? Drop a laser-guided bomb on the 'traitors' who rented him the truck?
That's the problem: guerilla warfare presents a diffuse target, and mixes it in with non-combatants and even friendlies. Remember your uncles and fathers in Vietnam: they *lost*. For all the might and resolve and industrial capacity and high-level of technology they *lost*.
The only way to win a guerrilla war against a diffuse enemy is not to fight. Do nothing beyond diplomacy, and try to keep the saber rattling quiet. They'll either become peaceful like Vietnam, or even more warlike like Iraq. In the former case, everybody wins. In the latter case, you will be fighting an open war against a massed enemy, and that is the kind of war the US can win.
More importantly this is a wakeup call of how lax the so-called security in this country is. While I have no desire to have my privacy invaded, etc., it's frightening how easy it is for a group of people to commit such large-scale violence in such a small amount of time.
Two points: 1) That's the essence of liberty: people can walk around and do whatever they want. The only 'solution' is a police state a la the Third Reich. 2) This is a very, very, very, very small-scale amount of violence. Compared to any fire bombing of a city during WWII, it's tiny. Compared to the number of people who die horribly *every year* during the monsoon season on the Indian subcontinent, it's tiny. Compared to the Battle of Gettysberg, it's tiny. Compared to the great flu epidemic early in the 20th centuring, it's absolutely nothing. Compared to a few months of the German death camps, it's nothing.
Sure, for the people involved it's the worst thing that could possibly happen, but the *nation* can take it. The *nation* can afford to sit back and ask "What strategic changes do we make to keep this from becoming a routine thing?"
Because we cannot let it become a routine thing. There are five billion potential enemies out there, many of them who, through personal experience, equate the US with death and suffering. It only takes a few dozen such people to kill hundreds of thousands of Americans. Make no mistake: we got off *easy* this time. Those planes could easily have been loaded with anthrax, or radioactive dust. Or they could have used something much more certain than planes, like poisoning the water supply, or blowing graphite dust into every major electrical substation in the NYC area in the dead of winter.
Unless it also... fiddles with my hosts.allow file, I'm not particularly concerned.
Whoa, cowboy!/etc/hosts.allow only affects friendly programs that bother to parse it (e.g., inetd, or programs that use tcpwrappers). An unfriendly program is free to ignore it.
However, your advice to use kernel firewalling is sound. 'Defense in depth' is the only way to go.
It's somewhat unconventional, but I think of the frame buffer as a two-dimensional signal. Similarly, I think of an MRI image as a three-dimensional signal. If you look at the nuts and bolts of graphics processing, it's pretty much like any other signal processing: data ordered along axes, acted on by vector and matrix operations.
My desktop doesnt hold too many fixed pixel width apps visible at 640x480. By switching to 1280x1024, it contains many more of those applications visible at any given time.
You're the one who missed it: Berlin doesn't work in either units of pixels or centimeters. It works in abstract length units which default to being centimeters. If the window manager is sanely designed, it will let you change the scale factor however you want. If you want a Netscape window that is scaled down to the size of a postage stamp with 3-pixel-tall fonts, just shrink it. Want another Netscape on the same screen to be normal size? No problem.
This is unspeakably nice. Neither X nor Win32 handle scaling properly. (And don't say Windows does it right because you can choose font sizes. If you choose a good font size for a 1600x1200 display, it will warn you that nothing will work and that you are completely fucked.)
Thanks for playing though.
And here are some lovely parting gifts for you too.;-)
Secondly, the idea of running EVERYTHING through OpenGL is particularly bothering. Most video hardware has some very specific optimisations for 2D work and by going through a specific 3D interface you are tossing all those performance advantages out the window. Sure, ok they want to create the ability to play with Windows in 3D - my question would have to be "Why".
The 3D stuff is used because it lets you do neat 2D stuff really fast, and not to make animated 3D windows fly around in space. To support games, the 3D hardware can scale and otherwise transform bitmaps to random locations and orientations on the screen. This is used, e.g., to apply a brick texture to the side of a building in a game. You can draw the application window into a 'texture' and let the video card draw it anywhere on the screen at any size. (With rotation and perspective, if you're *really* feeling silly.)
You can also align the texture buffer on page boundaries, and map the window's "frame buffer" directly into the app's address space. This lets programs have near-direct access to the frame buffer without any danger of blowing away the system.
Performance = Mhz * Instructions_per_Clock.... No one in their right mind would say a 1.4Ghz P4 is a better performer than a 1.4Ghz Athlon.
Don't be so cocksure about that. The P4 may have low instructions/clock efficiency for Word or Excel, but who cares? It runs them plenty fast for most people. There comes a point where the CPU is fast enough for randomly branching code, and it doesn't need to do that instruction mix any faster (modulo software bloat).
Where people want speed is signal processing: audio codecs for music and telephony, graphics, video compression and decompression, photorealistic games, and so forth. These programs run like absolute dogs on all existing processors (or at least there aren't any programmers sitting around wondering what to do with all the excess signal processing power). Fortunately, they also have very few branches in the core algorithms, which means very few pipeline refills, which means that the P4's deep pipeline is not much of a penalty. The thing that counts is how fast the pipeline is clocked.
So why does the P4 have a deep pipeline? Basic eletrical engineering: break the operations down into smaller stages with shorter propagation delays, and you can clock the pipeline faster. And speed is what counts for signal processing. The Athlon may be more efficient for some instruction mixes, but it does so at the expense of having complex logic with large propagation delays, which is harder to scale to higher clock speeds.
The P4's busses and cache fetching support this theory. After all, if you can do more signal calculations, you need more data to calculate on. Well, they made major efforts to build faster, wider busses so that the CPU can talk to the outside world faster. And the cache prefetches great whacking chunks of data to try to keep the bus as busy as possible. People ridiculed the P4 because the agressive prefetching hurts office applications, but they were missing the point, which is that agressive fetching is the only way to feed the core's voracious appetite for data when processing signals. Rambus, although it seems to be failing, was another part of their plan to stream vast quantities of data into and out of the CPU. Not random access, but streaming.
IMHO, the fast external bus and the ability to scale to higher clock speeds smells like signal processing.
What is the difference between a static strap and touching the grounded case of the unit you're about to work on? I can't see any, and from personal experience I'm far more apt to touch the case than I am to make sure I have the strap on and that the strap is grounded.
The strap provides good, continuous protection. The problem with holding your hand on the case is that 1) your hand -- especially the pads of your fingers and the palm -- is not a good conductor, and 2) it takes only a few milliseconds to build up enough charge to zap a component (just lifting your shoe off the carpet can build up a destructive charge). I consider a grounded strap essential when you've got your hands in the guts of a piece of equipment.
That said, I usually don't bother when I'm playing with cheap computers at home. I'm lazy and I can afford to replace the whole machine if it fries.
I agree with you that static protection is necessary -- don't get me wrong. Where I was digging in, however, was with the choice of static protection you had offered. Now of course, the ionizing air filters and so on weren't always around and neither were the shoe straps / antistatic shoes, but I was more or less taking the side of the techs who found the straps a pain in the ass.
Ionizers aren't as good as you think. Oh, when you have people wandering around with semi-enclosed, semi-hardened modules, they can help some, but they are no substitute for direct grounding.
I feel I have a better solution (touching the case) and it seems to work for me.
If you hooked an electrometer up to yourself while doing this, I think you'd be surprised at just how fast a few hundred volts can build up.
I can imagine... but really why weren't these super-duper-hyper-expensive cards built with some form of ESD protection (hell for that price it could have been DAMN GOOD protection) on the external ports?
That doesn't help when you're holding the board by its edges, and maybe accidentally touching chips in the middle of the board.
Pure economics. $0.50 on a strap is $0.50 less in profit per mobo.
It's basic economics. Most mobos are installed professionally by people who have nice $20 straps just laying around. Paying millions extra for the few people who don't already have straps is unjustifiable.
NB - I will never again design another single-sourced Maxim chip into one of my designs ever again. And if the alternate source is Linear Technology then it might as well be single-sourced. I am not alone in thinking like this!
But 16 week lead times and minimum quantities of 5000 are fun!;-)
I'm pretty sure that this has nothing to do with the Digital Millenium Copyright Act. In this case, the FBI seemed to be quite devious, not stupid. What does this have to do with Copyright violation?
He obtained copies of pages, which pages were subject to copyright, and which obtainment was without the authority of the copyright holder, and this was done by means of an device that circumvented access controls.
What OS in it's right mind allows code in the Stack Segment to be executed? If it's stack, it's obviously not a valid instruction, and should have been trapped.
Executable stack can be used for trampolines and thunking. True, it *is* something of a kluge, but it is legitimate.
Tantalum isn't "the crucial ingredient in capacitors", and the electronics industry doesn't "depend on this stuff" at all.
Tell that to all the people who were screaming -- me included -- when the tantalum cap supply was constrained. Tantalum caps may not be the only type, but they *are* important.
Most caps are made with ceramic materials (e.g. clays) or paper soaked in an electrolytic solution, but there are many other dielectrics available, like polypropylene, mica, etc., each with their own favourable characteristics.
The density (capacity per unit volume) of the ceramics and plastic film caps is too low for power supply filtering. Aluminum caps have good density but most have fairly high series resistance. Tantalum caps are have a great combination of high density and low equivalent series resistance.
But they're expensive, and polarised - you have to plug them in the right way, or they literally blow up.
Yeah, baby! The classic flash/bang of a dying tantalum.
They also have a high failure rate in use (especially infant mortality). Somebody once told me that Motorola didn't allow tantalum caps in pagers simply because too many of them spontaneously die. Ceramics cost more, but the savings on warranty returns and poor customer experience paid for it.
For the things that tantalums are most often used for (power-supply filtering), a kind of capacitor called multilayer ceramic actually works better.
It's only just now that ceramics with a competitive density are available, and they are still rather costly and availability still isn't good. Even then, the same advances that improve ceramics also work for tantalums, which have been getting steadily better too.
Now word is getting out that the new breed of multilayer ceramic chip caps can do just as well, people aren't using tantalums nearly as much as they were. I think this is the real reason for the tantalum ore crash.
The move to alternatives is part of the supply improvement, but there's also the fact that cell phones are slumping and manufacturers increased capacity.
Interesting. Can one of you Ruby or Python elmers tell me if its possible to, say, have the user input an arbitrary snippet of code at runtime
return (foo + bar)
and have the body of the multiplication operator be replaced with this, making it now an addition operator?
In Python, yes. At any point where you could define a class's * operator, you could define it by compiling an arbitrary string on the fly.
Slashdot Mirror
Or how about a laser system that illuminates everything within rifle range with a very bright, very sharp pulse. It detects anything that reflects: rifle scopes, binoculars, human eyes, horse eyes, etc. It then sends a focused high-energy pulse towards each reflection, burning out every retina and camera that was down-range. Unless the guerrillas can see through pinhole/slit glasses at night, they are pretty much fucked.
Those who think that technology is meaningless in a guerrilla war in the mountains obviously haven't been paying attention to modern technology. A guerrilla war against a large, *committed* first world nation is pretty much unwinnable, especially if the first world nation isn't worried about collateral damage.
So it is valid under some circumstances to say "an MMU is not good". However, Chuck Moore makes all sorts of statements like that one *as general laws*, which really squicks me. E.g., this page which says "The word */ multiplies by a ratio, with a double-length intermediate product. It eliminates the need for floating-point." Eliminates. Not "eliminates fp when you have a known and small dynamic range", but "eliminates". This has got to be one of the stupidest things I've ever seen. Another example: This is Chuck "Yes, that's all it takes." Moore's IDE driver. Puh-leeze. It's the "Hello World" of IDE drivers. It does *nothing*. No DMA. No locking for running on SMP machines. No autodetection and adaptation to drive capabilities. No workarounds for chipset bugs. No blacklisting of drives that are buggy. If you want to play in the real world, you need code like this (the Linux IDE driver).
Look at this page where he says "With the huge RAM of modern computers, an operating system is no longer necessary, if it ever was." Idiocy, written by a person who has never designed a system of significant complexity. RAM has *nothing* to do with whether an operating system is needed: it has everything to do with complexity. E.g., the only sane way to use floppy, IDE, SCSI, flash, CD-ROM, and network-mounted drives in the same system is to have a generic drive-access layer. Then you'll add a generic removeable drive layer to support things like CD-ROMs and flash drives that might suddenly disappear. If you want to support tons of hardware and software, you have to have an operating system.
This page is *full* of idiocy. "Don't try for platform portability. Most platform differences concern hardware interfaces. These are intrinsically different. Any attempt to make them appear the same achieves the lowest common denominator. That is, ignores the features that made the hardware attractive in the first place." Except, of course, for the usual case where the hardware has an upgrade that is easy to turn on and use, and which hardware is available is not known until runtime.
So I'm supposed to have 82 processors on account of having 82 processes that might become runnable at any moment? What about shell servers with hundreds of simultaneous users? Lack of memory protection is just ludicrous, it's so not-real-world it isn't even funny.
And as far as being more secure, that's ludicrous too. A functioning MMU is just as secure.
Unless we want repeat performances, we'd better figure out how to avoid instigating this kind of warfare in the future. The risk mitigation techniques needed are vastly different than for conventional warfare, and I fear that the old guard will get hundreds of thousands of Americans killed before they figure this (much like they did when they ignored air power in a previous war). And least the Enemy hit the Pentagon. Thank God for that. Even the most fossilized Cold Warrior has trouble ignoring an airplane hitting *his* building.
As did many Americans celebrate the death of tens of thousands of Iraqi soldiers in their bunkers, after which they put "Nuke Iraq" bumper stickers on their cars (remember those?). Judge not lest you be judged.
Is it even certain yet that the attackers are Middle Eastern? After the Murrah building bombing in Oklahoma City there was considerable harrasment of Middle Eastern immigrants in the area, it being a forgone conclusion that those goddamn raghead extremist fureners did it. Which looked pretty stupid when it turned out to have been an American, and a former US soldier at that!
People, remember that there are plenty of violent, organized fools right here in the US: anarchists, cults, people who think the South Shall Rise Again, and so forth. It does look suspiciously like a foreign attack, but don't jump to conclusions. There will be a hell of a lot of evidence, from surveillance tapes at the airports where they embarked, to radio reports from the pilots, perhaps even cell-phone reports from airline passengers. FBI field agents were no doubt on their way to the airports within a couple of hours of the attacks. If these terrorists are as incompentent as terrorists usually are, there will be tons of good, solid evidence against them, and in the meantime it would be a Bad Thing to jump to any political conclusions.
As for the Palestinians and Iraqis celebrating: big deal. If California slid off into the sea, they'd celebrate. Condeming the US is a hobby/spectator sport for them.
You are not alone in feeling this way. There are many here in Oklahoma -- including me -- who also thought that the best monument to the victims would be a new federal building. Unfortunately the busybodies and bureaucrats have a way of dominating such things.
Yes. The United States has been pushing people around and generally being a nuisance to the Middle East for nigh on 50 years now. The US is very big, and very strong, and they are very small and very weak. We have B-52 Stratofortresses and they have rifles and Molotov cocktails.
In a power disparity like that, there's only one way for the little guy to fight: infrequent lightning attacks. These guys don't have the resources to carpet bomb Manhattan with daisy cutters: that would take tens of thousand of people and a very high industrial level. They *do* have the resources to send a dozen men to hijack a handful of planes. A dozen men, I might add, that are brainwashed, do not know where they were trained, and cannot identify any of the few dozen (perhaps up to few hundred) other people in the organization.
And who were we to blow up to avenge the act of Timothy McVeigh? Detroit? Drop a laser-guided bomb on the 'traitors' who rented him the truck?
That's the problem: guerilla warfare presents a diffuse target, and mixes it in with non-combatants and even friendlies. Remember your uncles and fathers in Vietnam: they *lost*. For all the might and resolve and industrial capacity and high-level of technology they *lost*.
The only way to win a guerrilla war against a diffuse enemy is not to fight. Do nothing beyond diplomacy, and try to keep the saber rattling quiet. They'll either become peaceful like Vietnam, or even more warlike like Iraq. In the former case, everybody wins. In the latter case, you will be fighting an open war against a massed enemy, and that is the kind of war the US can win.
Two points: 1) That's the essence of liberty: people can walk around and do whatever they want. The only 'solution' is a police state a la the Third Reich. 2) This is a very, very, very, very small-scale amount of violence. Compared to any fire bombing of a city during WWII, it's tiny. Compared to the number of people who die horribly *every year* during the monsoon season on the Indian subcontinent, it's tiny. Compared to the Battle of Gettysberg, it's tiny. Compared to the great flu epidemic early in the 20th centuring, it's absolutely nothing. Compared to a few months of the German death camps, it's nothing.
Sure, for the people involved it's the worst thing that could possibly happen, but the *nation* can take it. The *nation* can afford to sit back and ask "What strategic changes do we make to keep this from becoming a routine thing?"
Because we cannot let it become a routine thing. There are five billion potential enemies out there, many of them who, through personal experience, equate the US with death and suffering. It only takes a few dozen such people to kill hundreds of thousands of Americans. Make no mistake: we got off *easy* this time. Those planes could easily have been loaded with anthrax, or radioactive dust. Or they could have used something much more certain than planes, like poisoning the water supply, or blowing graphite dust into every major electrical substation in the NYC area in the dead of winter.
You hire people based on how good their Slashdot posts are? I've heard of weird interview techniques, but this takes the cake!
Whoa, cowboy!
However, your advice to use kernel firewalling is sound. 'Defense in depth' is the only way to go.
It's somewhat unconventional, but I think of the frame buffer as a two-dimensional signal. Similarly, I think of an MRI image as a three-dimensional signal. If you look at the nuts and bolts of graphics processing, it's pretty much like any other signal processing: data ordered along axes, acted on by vector and matrix operations.
You're the one who missed it: Berlin doesn't work in either units of pixels or centimeters. It works in abstract length units which default to being centimeters. If the window manager is sanely designed, it will let you change the scale factor however you want. If you want a Netscape window that is scaled down to the size of a postage stamp with 3-pixel-tall fonts, just shrink it. Want another Netscape on the same screen to be normal size? No problem.
This is unspeakably nice. Neither X nor Win32 handle scaling properly. (And don't say Windows does it right because you can choose font sizes. If you choose a good font size for a 1600x1200 display, it will warn you that nothing will work and that you are completely fucked.)
And here are some lovely parting gifts for you too.
The 3D stuff is used because it lets you do neat 2D stuff really fast, and not to make animated 3D windows fly around in space. To support games, the 3D hardware can scale and otherwise transform bitmaps to random locations and orientations on the screen. This is used, e.g., to apply a brick texture to the side of a building in a game. You can draw the application window into a 'texture' and let the video card draw it anywhere on the screen at any size. (With rotation and perspective, if you're *really* feeling silly.)
You can also align the texture buffer on page boundaries, and map the window's "frame buffer" directly into the app's address space. This lets programs have near-direct access to the frame buffer without any danger of blowing away the system.
Don't be so cocksure about that. The P4 may have low instructions/clock efficiency for Word or Excel, but who cares? It runs them plenty fast for most people. There comes a point where the CPU is fast enough for randomly branching code, and it doesn't need to do that instruction mix any faster (modulo software bloat).
Where people want speed is signal processing: audio codecs for music and telephony, graphics, video compression and decompression, photorealistic games, and so forth. These programs run like absolute dogs on all existing processors (or at least there aren't any programmers sitting around wondering what to do with all the excess signal processing power). Fortunately, they also have very few branches in the core algorithms, which means very few pipeline refills, which means that the P4's deep pipeline is not much of a penalty. The thing that counts is how fast the pipeline is clocked.
So why does the P4 have a deep pipeline? Basic eletrical engineering: break the operations down into smaller stages with shorter propagation delays, and you can clock the pipeline faster. And speed is what counts for signal processing. The Athlon may be more efficient for some instruction mixes, but it does so at the expense of having complex logic with large propagation delays, which is harder to scale to higher clock speeds.
The P4's busses and cache fetching support this theory. After all, if you can do more signal calculations, you need more data to calculate on. Well, they made major efforts to build faster, wider busses so that the CPU can talk to the outside world faster. And the cache prefetches great whacking chunks of data to try to keep the bus as busy as possible. People ridiculed the P4 because the agressive prefetching hurts office applications, but they were missing the point, which is that agressive fetching is the only way to feed the core's voracious appetite for data when processing signals. Rambus, although it seems to be failing, was another part of their plan to stream vast quantities of data into and out of the CPU. Not random access, but streaming.
IMHO, the fast external bus and the ability to scale to higher clock speeds smells like signal processing.
The strap provides good, continuous protection. The problem with holding your hand on the case is that 1) your hand -- especially the pads of your fingers and the palm -- is not a good conductor, and 2) it takes only a few milliseconds to build up enough charge to zap a component (just lifting your shoe off the carpet can build up a destructive charge). I consider a grounded strap essential when you've got your hands in the guts of a piece of equipment.
That said, I usually don't bother when I'm playing with cheap computers at home. I'm lazy and I can afford to replace the whole machine if it fries.
Ionizers aren't as good as you think. Oh, when you have people wandering around with semi-enclosed, semi-hardened modules, they can help some, but they are no substitute for direct grounding.
If you hooked an electrometer up to yourself while doing this, I think you'd be surprised at just how fast a few hundred volts can build up.
That doesn't help when you're holding the board by its edges, and maybe accidentally touching chips in the middle of the board.
It's basic economics. Most mobos are installed professionally by people who have nice $20 straps just laying around. Paying millions extra for the few people who don't already have straps is unjustifiable.
But 16 week lead times and minimum quantities of 5000 are fun!
Dammit. That didn't do what I hope it'd do. Sorry for the noise.
..._
He obtained copies of pages, which pages were subject to copyright, and which obtainment was without the authority of the copyright holder, and this was done by means of an device that circumvented access controls.
Looks as good as any other DMCA case to me.
They also have a high failure rate in use (especially infant mortality). Somebody once told me that Motorola didn't allow tantalum caps in pagers simply because too many of them spontaneously die. Ceramics cost more, but the savings on warranty returns and poor customer experience paid for it.
It's only just now that ceramics with a competitive density are available, and they are still rather costly and availability still isn't good. Even then, the same advances that improve ceramics also work for tantalums, which have been getting steadily better too. The move to alternatives is part of the supply improvement, but there's also the fact that cell phones are slumping and manufacturers increased capacity.