I would expect vampire bat saliva to have a powerful anti-clotting agent -- and to be as inoccuous otherwise as possible.
Consider: If the victims of vampire bats tended to die, the bats would soon find themselves with fewer hosts to feed from. By minimizing the damage done during a feeding, the bat could go back to the host again and again, feeding multiple times. A colony that did less damage would have a better food base and would spread more than one which killed its victims.
I've heard from a lot of people, and my conclusion is, the best set of furniture to put in a bedroom is -- a bed.
No, don't mod this as funny. I'm serious.
The way human psychology works, you want to have a place where you do nothing else but go to sleep. Any other distractions, and you're creating a basis for insomnia. I can understand the need to save space, so perhaps add some clothes storage so you can get dressed there, but even that is a distraction. If possible, you should do that in another room. The bedroom can be just large enough for the bed, and a little stand for an alarm clock, and that's all.
Don't put a TV or stereo in the bedroom. Don't read in the bedroom. Again, you're only distracting yourself from falling asleep. I know you're using them to try to fall asleep; don't. If you want to watch TV or read, do it in another room.
I'm afraid MPEG will have to make do on half their expected revenue. (Frankly, I suspect it'll be more than half; by cutting their own prices, they'll gain more customers, and since costs for royalties are pretty much arbitrary, they won't have more in expenses to lay out.)
Microsoft can price their product however they please. When they start causing problems, by restricting the platforms their codec performs on, or restricting the performance on other platforms, or if they wait 'til MPEG is dead and then raise their rates, THEN you can slam them for monopolistic practices.
In the meantime, projects like Ogg will proceed, as will DivX, producing competitors MS may prove hard to beat. So let 'em try to take over the market...
You are correct, sir; if you obtained one key, you could write an emulator that would operate using that key. Or you could clone Palladium hardware containing that key.
It wouldn't last, though, I'm afraid...
Once someone had gotten wind you'd cloned that particular key (and there might be several ways they could find out; multiple installations of software using that key, for instance), that key would be disallowed for future software registrations. It would mean you could no longer update your OS, and new installs of software couldn't be done, either.
The X-Box is designed like that first class of Palladium chips, and security has been bypassed by placing a mod-chip in the data path of the key access. As I said before, in future implementations the crypto hardware will be inside the CPU chip, so there won't be a line to tap.
You are correct the key is stored somewhere; but it's not anywhere it can be read. It's kept where a separate crypto processor can use it to validate signatures and decrypt code, but the PC has no access to it. Reading the key would involve physically opening and tapping into the chip, which is a practical impossibility for you or me. Such an effort might be worthwhile if reading the key would result in breaking the entire Palladium system, but as I said, the system is designed to thwart class breaks.
You picture the processor decrypting code and storing it in main memory, but in fact the decrypted code is only stored locally. Again, on the first implementations, this code might be intercepted while it's on the bus between the crypto chip and the CPU, in the future it'll be impossible.
As for "tricking the OS into thinking it's trusted"... Nope. The machine won't boot without trusted code, period. The BIOS is signed and trusted. It'll shut down the crypto processor and boot a non-trusted OS (at least in current specs), but from that point no trusted software can run. It will check the signature on a trusted OS and boot that; if the OS is modified, the signature will no longer match, and the OS won't boot.
TCPA/Palladium is an extremely elegant, hardy, and EVIL system. It worries me greatly.
Palladium and open-source are pretty close to mutually exclusive. One COULD make a trusted *ix distribution, but either (1) the Palladium key would be held only by the distributor, and anyone writing patches would have to run the OS in untrusted mode, or (2) the Palladium key would be publically available -- and therefore no one would write trusted apps for it, for what would be the point? I do not know whether one could generate a working key from out of the blue, either.
> Do you really think that this won't get hacked > (somehow), within a few months or less?
The Palladium specification is actually very strong, and is designed to prevent "class breaks"; or, in other words, if you broke the key on one computer, it wouldn't affect the security on all the other machines out there. It's an open spec; if you want to examine the security, take a looksie.
The first Palladium-enabled PCs will have separate encryption processors for dealing with trusted source. Future PCs will have that encryption built right into the CPU, and yanking it won't be an option. Further, when a program or OS is "trusted", it's not only signed but can also be encrypted, to prevent reverse-engineering; physically yanking/destroying the Palladium hardware would prevent the OS and programs from being accessed entirely.
Actually no; not only would VMware have to be Palladium-enabled, but the OS it was booted under would need to be as well. Otherwise, the hardware wouldn't allow the program access to the encryption hardware.
The original Palladium spec calls for a trusted machine to only allow trusted access by trusted operating systems. This means Palladium-encrypted code won't run except under a Palladium-rated OS. If the OS isn't trusted, then no Palladium-enabled programs can run.
This will mean that WINE will be useless for many future Windows apps, especially those dealing with multimedia. It also means future versions of Windows will be written specifically to defeat applications like VMware, so as to not violate the security.
These are bad, though they don't prevent one from booting a non-Palladium-enabled OS and using alternative applications. What I keep worrying about is the TCPA *2.0* specification. The original spec allows an alternative to a "trusted" platform, but future specs may require a PC boot a Palladium-enabled OS -- or none at all.
I suspect a big reason why this technology was never fully developed back in 1984 is because the technology wasn't there.
Remember that, if you're building an integrated circuit the old-fashioned way, your feature size (for example, the width of a conductor) is at most the same as the wavelength of light used to prepare it for etching. Until recently, nothing over low-ultraviolet light could be used, and therefore the features were relatively large compared to the wavelength of visible light. An effective antenna designed to utilize visible light frequencies would require features at least an order of magnitude smaller than the frequencies it's meant to absorb.
Or in other words, they couldn't do it.
Today we're able to create integrated circuits with features less than 150 nanometers across, so these optical-frequency rectennas are now feasible. Whether they're practical is another matter; remember, you'd need to use the latest fabrication techniques to get features that small, so manufacturing would be expensive.
If the machines are at more than one location, use one flunky at each. If there are a whole lot of machines at a single location, use more than one flunky. Use multiple CDs per flunky, so he can go service more than one machine while the others are writing.
Yes, sending a CD image over the internet can burn a lot of bandwidth. Use sneakernet instead. Heck, if you're in a hurry, Fedex it. Otherwise the updates may be DAYS late -- horrors!
As for workstations not being identical -- if they aren't, and you've got thousands of them, your IS department is going to be humongous. Most places will standardize the systems rather than try and support all the possible combinations of software. At worst, you shouldn't see more than three or four different installs, at least that IS is responsible for.
I just got spammed with a notice of a "new Network Solutions". Among other things, they're offering frequent flyer miles... So I'm no longer worried they'll make a big come back.:-P
First off -- you should be running two tiers of systems; one where a default set of applications are installed, and users' installs aren't guaranteed to stick; and one where a user assumes responsibility of his own machine and has to figure out his own problems.
Now your job is greatly simplified. Use a utility that overwrites the boot partition on a machine with the image stored on a CD. (Let users store their data files in a second partition.) Update the OS to the current level, and make an image CD using it. Then get a flunky to go to each machine and re-image it. (Do this after hours when the place is empty.)
You've probably got minute calcium deposits on the surface, making it less smooth. You might want to try using a buffing tool. Think about how shiny your car could be, too!
If buffing won't do it, then yes, there may be some coating the factory applied before it sent you the unit, and you might be able to apply more. Remember, it worked for a year before it gummed up on you.
Yeah, at 100% efficiency... I think you'd be hard-pressed to reach 10% efficiency. That would still be useful, with a 4' x 8' collector producing 4 liters of fresh water per day...
But then, you'd also need to take into account that there aren't a lot of places in the US where you can get consistently sunny skies, clear for ten hours per day. A run of cloudy weather for a couple of weeks would devastate your system unless you had a sizeable reserve tank.
Further, solar evaporators need a surprising amount of maintenance. The surfaces need to be kept clean or the efficiency drops, and they have to be flushed periodically to keep bacteria from growing in the fresh water outlets.
To sum up -- it's do-able, but not quite as easy as building one and then getting free water forever.
Don't some MP3 player companies sell their "jukebox"-type units with no hard drive installed? I don't think Canada's laws impose that tariff on those units... You could then get a new or used notebook drive, and trip the MP3 fantastic.
If you want the net to be completely free, then not only can't portions of it be restricted as intellectual property, then it also shouldn't be restricted by a license like the GPL.
I would tend to oppose the use of a standard in which the specification is GPL'd, because in the case of a standard, I'd like commercial entities to be able to incorporate access to such a standard in their software; something they probably wouldn't do if they would be forced to open their source. This would effectively limit access to that standard to non-commercial programs.
If the author of a GPL'd work wishes to submit the protocol involved as a standard, then let him submit it as FREE. He can keep the code GPL'd, but the standard must be free for use -- even by commercial entities.
"Yes, Lieutenant. I've already heard your name, rank, and serial number, over and over again. Now, I'd like to show you this photo... Steady! (Hold him, please.) Our sources looked up your next of kin in your medical records... This is a recent photo of your mother and father, hm? Our operatives are quite good at photography, we train them well.
"Now where were we? Oh yes. Now, Lieutenant, I'd like you to begin talking. And please remember, your parents' lives depend on what you say. Name, rank and serial number are not acceptable."
You're getting a lot of advice to buy a notebook PC instead of a compact machine for traveling, and I concur; among other things, while carrying a compact system unit wouldn't be so hard, current LCD monitors meant for desktops really aren't meant for travel, and you'd probably wind up busting it...
If you have extra carrying capacity available, then I'd recommend you bring a full-sized keyboard and a good mouse. I'm guessing your biggest peeve with laptop PCs is they don't feel like a real machine. Add those two light and cheap peripherals and it'll feel much more solid.
You're right, I'm sorry... I should not have said the system was unencrypted, but that the system had unacceptably weak encryption.
But you say it yourself -- a higher level crypto protocol wouldn't be supported in hardware. It also can't be supported in software, because during boot there IS no software loaded. (Unless your system's BIOS supports such things, which I doubt.) So therefore, at boot time, this operation would be unacceptably insecure.
Digital cellular and PCS use a voice-grade compression, which squeezes your conversation down to about 4kbps, in each direction. Your 14.4kbps, 9.6kbps or even 4.8kbps fax connection will not handle the compression at all well, and in the case of data transmission, even at 2400bps or less you're going to see data corruption.
With analog cellular, on the other hand, you've got two actual analog radio channels dedicated to the conversation, and while noise would be a concern, a modem signal with error correction can squeeze through it.
Slashdot Mirror
And just for people who don't know, PB = petabyte, and eb = exabyte.
I would expect vampire bat saliva to have a powerful anti-clotting agent -- and to be as inoccuous otherwise as possible.
Consider: If the victims of vampire bats tended to die, the bats would soon find themselves with fewer hosts to feed from. By minimizing the damage done during a feeding, the bat could go back to the host again and again, feeding multiple times. A colony that did less damage would have a better food base and would spread more than one which killed its victims.
I've heard from a lot of people, and my conclusion is, the best set of furniture to put in a bedroom is -- a bed.
No, don't mod this as funny. I'm serious.
The way human psychology works, you want to have a place where you do nothing else but go to sleep. Any other distractions, and you're creating a basis for insomnia. I can understand the need to save space, so perhaps add some clothes storage so you can get dressed there, but even that is a distraction. If possible, you should do that in another room. The bedroom can be just large enough for the bed, and a little stand for an alarm clock, and that's all.
Don't put a TV or stereo in the bedroom. Don't read in the bedroom. Again, you're only distracting yourself from falling asleep. I know you're using them to try to fall asleep; don't. If you want to watch TV or read, do it in another room.
I'm afraid MPEG will have to make do on half their expected revenue. (Frankly, I suspect it'll be more than half; by cutting their own prices, they'll gain more customers, and since costs for royalties are pretty much arbitrary, they won't have more in expenses to lay out.)
Microsoft can price their product however they please. When they start causing problems, by restricting the platforms their codec performs on, or restricting the performance on other platforms, or if they wait 'til MPEG is dead and then raise their rates, THEN you can slam them for monopolistic practices.
In the meantime, projects like Ogg will proceed, as will DivX, producing competitors MS may prove hard to beat. So let 'em try to take over the market...
You are correct, sir; if you obtained one key, you could write an emulator that would operate using that key. Or you could clone Palladium hardware containing that key.
It wouldn't last, though, I'm afraid...
Once someone had gotten wind you'd cloned that particular key (and there might be several ways they could find out; multiple installations of software using that key, for instance), that key would be disallowed for future software registrations. It would mean you could no longer update your OS, and new installs of software couldn't be done, either.
Evil, I say. Evil.
The X-Box is designed like that first class of Palladium chips, and security has been bypassed by placing a mod-chip in the data path of the key access. As I said before, in future implementations the crypto hardware will be inside the CPU chip, so there won't be a line to tap.
You are correct the key is stored somewhere; but it's not anywhere it can be read. It's kept where a separate crypto processor can use it to validate signatures and decrypt code, but the PC has no access to it. Reading the key would involve physically opening and tapping into the chip, which is a practical impossibility for you or me. Such an effort might be worthwhile if reading the key would result in breaking the entire Palladium system, but as I said, the system is designed to thwart class breaks.
You picture the processor decrypting code and storing it in main memory, but in fact the decrypted code is only stored locally. Again, on the first implementations, this code might be intercepted while it's on the bus between the crypto chip and the CPU, in the future it'll be impossible.
As for "tricking the OS into thinking it's trusted"... Nope. The machine won't boot without trusted code, period. The BIOS is signed and trusted. It'll shut down the crypto processor and boot a non-trusted OS (at least in current specs), but from that point no trusted software can run. It will check the signature on a trusted OS and boot that; if the OS is modified, the signature will no longer match, and the OS won't boot.
TCPA/Palladium is an extremely elegant, hardy, and EVIL system. It worries me greatly.
Palladium and open-source are pretty close to mutually exclusive. One COULD make a trusted *ix distribution, but either (1) the Palladium key would be held only by the distributor, and anyone writing patches would have to run the OS in untrusted mode, or (2) the Palladium key would be publically available -- and therefore no one would write trusted apps for it, for what would be the point? I do not know whether one could generate a working key from out of the blue, either.
> Do you really think that this won't get hacked
> (somehow), within a few months or less?
The Palladium specification is actually very strong, and is designed to prevent "class breaks"; or, in other words, if you broke the key on one computer, it wouldn't affect the security on all the other machines out there. It's an open spec; if you want to examine the security, take a looksie.
The first Palladium-enabled PCs will have separate encryption processors for dealing with trusted source. Future PCs will have that encryption built right into the CPU, and yanking it won't be an option. Further, when a program or OS is "trusted", it's not only signed but can also be encrypted, to prevent reverse-engineering; physically yanking/destroying the Palladium hardware would prevent the OS and programs from being accessed entirely.
Actually no; not only would VMware have to be Palladium-enabled, but the OS it was booted under would need to be as well. Otherwise, the hardware wouldn't allow the program access to the encryption hardware.
The original Palladium spec calls for a trusted machine to only allow trusted access by trusted operating systems. This means Palladium-encrypted code won't run except under a Palladium-rated OS. If the OS isn't trusted, then no Palladium-enabled programs can run.
This will mean that WINE will be useless for many future Windows apps, especially those dealing with multimedia. It also means future versions of Windows will be written specifically to defeat applications like VMware, so as to not violate the security.
These are bad, though they don't prevent one from booting a non-Palladium-enabled OS and using alternative applications. What I keep worrying about is the TCPA *2.0* specification. The original spec allows an alternative to a "trusted" platform, but future specs may require a PC boot a Palladium-enabled OS -- or none at all.
I suspect a big reason why this technology was never fully developed back in 1984 is because the technology wasn't there.
Remember that, if you're building an integrated circuit the old-fashioned way, your feature size (for example, the width of a conductor) is at most the same as the wavelength of light used to prepare it for etching. Until recently, nothing over low-ultraviolet light could be used, and therefore the features were relatively large compared to the wavelength of visible light. An effective antenna designed to utilize visible light frequencies would require features at least an order of magnitude smaller than the frequencies it's meant to absorb.
Or in other words, they couldn't do it.
Today we're able to create integrated circuits with features less than 150 nanometers across, so these optical-frequency rectennas are now feasible. Whether they're practical is another matter; remember, you'd need to use the latest fabrication techniques to get features that small, so manufacturing would be expensive.
http://search.news.yahoo.com/search/news?p=smartru ck&c=news_photos
(working link to Smartruck II page) http://www.smartruck2.com
If the machines are at more than one location, use one flunky at each. If there are a whole lot of machines at a single location, use more than one flunky. Use multiple CDs per flunky, so he can go service more than one machine while the others are writing.
Yes, sending a CD image over the internet can burn a lot of bandwidth. Use sneakernet instead. Heck, if you're in a hurry, Fedex it. Otherwise the updates may be DAYS late -- horrors!
As for workstations not being identical -- if they aren't, and you've got thousands of them, your IS department is going to be humongous. Most places will standardize the systems rather than try and support all the possible combinations of software. At worst, you shouldn't see more than three or four different installs, at least that IS is responsible for.
His big SF series about "Honor Harrington" relied on a FTL communications system based on creating gravity waves.
I just got spammed with a notice of a "new Network Solutions". Among other things, they're offering frequent flyer miles... So I'm no longer worried they'll make a big come back. :-P
First off -- you should be running two tiers of systems; one where a default set of applications are installed, and users' installs aren't guaranteed to stick; and one where a user assumes responsibility of his own machine and has to figure out his own problems.
Now your job is greatly simplified. Use a utility that overwrites the boot partition on a machine with the image stored on a CD. (Let users store their data files in a second partition.) Update the OS to the current level, and make an image CD using it. Then get a flunky to go to each machine and re-image it. (Do this after hours when the place is empty.)
Presto. You're updated.
You've probably got minute calcium deposits on the surface, making it less smooth. You might want to try using a buffing tool. Think about how shiny your car could be, too!
If buffing won't do it, then yes, there may be some coating the factory applied before it sent you the unit, and you might be able to apply more. Remember, it worked for a year before it gummed up on you.
Yeah, at 100% efficiency... I think you'd be hard-pressed to reach 10% efficiency. That would still be useful, with a 4' x 8' collector producing 4 liters of fresh water per day...
But then, you'd also need to take into account that there aren't a lot of places in the US where you can get consistently sunny skies, clear for ten hours per day. A run of cloudy weather for a couple of weeks would devastate your system unless you had a sizeable reserve tank.
Further, solar evaporators need a surprising amount of maintenance. The surfaces need to be kept clean or the efficiency drops, and they have to be flushed periodically to keep bacteria from growing in the fresh water outlets.
To sum up -- it's do-able, but not quite as easy as building one and then getting free water forever.
Outsourcing is easy.
In-house is profitable.
You choose.
Don't some MP3 player companies sell their "jukebox"-type units with no hard drive installed? I don't think Canada's laws impose that tariff on those units... You could then get a new or used notebook drive, and trip the MP3 fantastic.
The GPL is not an unrestricted license.
I repeat:
THE GPL IS NOT AN UNRESTRICTED LICENSE.
If you want the net to be completely free, then not only can't portions of it be restricted as intellectual property, then it also shouldn't be restricted by a license like the GPL.
I would tend to oppose the use of a standard in which the specification is GPL'd, because in the case of a standard, I'd like commercial entities to be able to incorporate access to such a standard in their software; something they probably wouldn't do if they would be forced to open their source. This would effectively limit access to that standard to non-commercial programs.
If the author of a GPL'd work wishes to submit the protocol involved as a standard, then let him submit it as FREE. He can keep the code GPL'd, but the standard must be free for use -- even by commercial entities.
"Yes, Lieutenant. I've already heard your name, rank, and serial number, over and over again. Now, I'd like to show you this photo... Steady! (Hold him, please.) Our sources looked up your next of kin in your medical records... This is a recent photo of your mother and father, hm? Our operatives are quite good at photography, we train them well.
"Now where were we? Oh yes. Now, Lieutenant, I'd like you to begin talking. And please remember, your parents' lives depend on what you say. Name, rank and serial number are not acceptable."
You're getting a lot of advice to buy a notebook PC instead of a compact machine for traveling, and I concur; among other things, while carrying a compact system unit wouldn't be so hard, current LCD monitors meant for desktops really aren't meant for travel, and you'd probably wind up busting it...
If you have extra carrying capacity available, then I'd recommend you bring a full-sized keyboard and a good mouse. I'm guessing your biggest peeve with laptop PCs is they don't feel like a real machine. Add those two light and cheap peripherals and it'll feel much more solid.
You're right, I'm sorry... I should not have said the system was unencrypted, but that the system had unacceptably weak encryption.
But you say it yourself -- a higher level crypto protocol wouldn't be supported in hardware. It also can't be supported in software, because during boot there IS no software loaded. (Unless your system's BIOS supports such things, which I doubt.) So therefore, at boot time, this operation would be unacceptably insecure.
Digital cellular and PCS use a voice-grade compression, which squeezes your conversation down to about 4kbps, in each direction. Your 14.4kbps, 9.6kbps or even 4.8kbps fax connection will not handle the compression at all well, and in the case of data transmission, even at 2400bps or less you're going to see data corruption.
With analog cellular, on the other hand, you've got two actual analog radio channels dedicated to the conversation, and while noise would be a concern, a modem signal with error correction can squeeze through it.