Apparently, you are too stupid to understand that electromagnetic radiation and nuclear radiation are not the same and do not have the same properties. So please stop telling other people they are wrong.
If there was a dictionary of fingerprint patterns, we wouldn't be using fingerprints to identify people. Hopefully, there are a lot more parameters than in a password. Try to reconstruct the pattern on my finger from a string like v9xc87z6b5684852b79v8x68bxvb78xv. That would be impossible, because that string would not have enough bits to represent the whole pattern.
Your post just proved that you don't know what you are talking about. I have read the article you are referring to. The only conclusion one can draw from it is that many biometric systems have security problems. It is almost completely irrelevant to my point, except that part about glass doors (which is the method they used).
The article did not demonstrate that data could be extracted from an existing system and the thumb reconstructed from that data. The above article tested mostly low-security and consumer-grade systems, and admitted that the method did not work with the better systems out there.
How the hell can you "spoof other systems", short of making a duplicate thumb? What keeps those perps from just getting your thumbprint off a glass door you touched? Besides, what makes you think you can duplicate the thumb from the biometric system's data file? If whoever made that system had a shred of intelligence, they would use a one-way hash for thumbprints and match based on that.
It's not like they teach soldering in EE programs. Granted, you might find someone experienced, but be careful. Many electrical engineering _students_ these days won't even know how to solder, much less do it well. Watch out.
McVoy IS a tantrum-throwing child. If you knew anything about the whole BitKeeper situation, you would have noticed that. Yeah, he let some projects use his software. WOW. Let's see, it cost him exactly $0.00 and he got lots of free publicity from it. He has all the compassion of a loan shark.
First to file does NOT mean you can take something that someone else invented and patent it. It means that when two people invent the same thing simultaneously, the first one to file the patent gets it. Of course, if the first person publishes the invention, it will become prior art, and therefore ineligible for patent protection.
The first to invent thing was always kooky because it required inventors to keep very good records of when they invented what. When you get a good idea, do you run to a notary public to get it notarized? That's what "first to invent" requires as documentation. Every other country in the world uses "first to file" as the standard for patents.
You are right about the efficiency problems. However, you are mistaken on your definition of a full-wave rectifier. You are probably thinking of a voltage doubler (which are commonly used in SMPSs). A full-wave rectifier is just 4 diodes, and it will produce 120V * sqrt(2) which is approximately 170Vpeak. The waveform from it looks like abs(sin(120*pi*t)). It can then be filtered, and the voltage will be ~170VDC. A voltage doubler charges one capacitor with the positive voltage swing and another with the negative swing, so you double the voltage. Here's a link to a voltage doubler circuit. This is a full-wave rectifier.
I wasn't aware voltage error was that high! Presumably that's for unregulated power supplies?
Most power bricks (the ones that plug directly into an outlet and have a transformer inside) are completely unregulated. The voltage that's printed on them is strictly nominal, and the internal resistance of the transformer is so high that the voltage drops considerably with load. For instance, a randomly picked adapter out of my junkbox says 3V on it. It actually puts out 3.88V no load, which is a 30% error. Generally, these are meant to be used with a regulator inside the device.
A switching power supply that powers my speakers (this one is regulated) is rated for 18V and actually puts out 18.7V no load. This is only a 3.8% error, which is to be expected given the component tolerances in it. But specifying 14.1V for a device implies a 0.3% tolerance, which is far too much to expect from an inexpensive power brick.
The reason a UPS will have a huge transformer is because it has to put out a 50 or 60Hz sine wave, and not DC. Transformers running at 60Hz need far more inductance than ones running at, say, 500kHz (a common frequency for switching power supplies). More inductance means lots of metal and lots of turns of thick wire. At high frequencies, you can get away with small transformers using just a few turns of wire. It's also easier to rectify and filter 500KHz square waves than 60Hz sine waves (although the interference problems are very challenging). A computer power supply is a switching design, which is why it can put out 300W at multiple voltages and not weigh a ton.
Technically, that makes the 240V system a two-phase system. However, in that particular case, it was most likely due to a swapped hot and neutral at the outlet.
I agree, having it run off of 12VDC is a stupid idea. It will not work simply due to the very large required currents.
Besides, you wouldn't generate 170VDC, but 300VDC - you full wave rectify the signal and dump the +170VDC into one cap, and the -170VDC into another.
A full-wave rectifier would produce 170V, not 300V. You can only get 300V if you use a voltage doubler circuit.
Only for a fixed output voltage design running off a high voltage main bar.
Fixed or variable doesn't make much of a difference (it's just a matter of changing the PWM duty cycle), and neither does using a buck/boost topology. The main losses are in the switching elements and the transformers, and are mainly determined by how much you are willing to pay. Of course, running it off of a very low voltage like 12V won't be very efficient due to the high currents involved.
But there are plenty of peripherals that need voltages like 14.1v.
Generally, a voltage requirement like 14.1V means the peripheral will accept anything from 12V to 16V but they don't want you using a third-party adapter. Even switching adapters are usually off by half a volt or more, and regular transformer bricks are +/- 30% or more.
Also, you don't need heavy transformers to put out high currents. Read about switching power supplies sometime. With those, the transformer size depends mainly on switching frequency.
You do bring up a valid point. I think the main reason nobody makes such a device is because it's impossible to make idiot-proof. If someone doesn't set the voltage or polarity properly, they will be very pissed when they fry the device. When you have multiple outputs, this is extremely easy to do.
That is really a very wide voltage swing. If the main internal power bus is less than 26V, it means each output module must be a buck/boost converter (which is more complex and expensive than the other solutions, as well as less efficient).
The main power bus will be about 170VDC, since the power brick will presumably be AC powered.
However, for 2V a 50W supply is 25A - that takes BIG wire to carry, and special connectors.
Generally, devices don't require much more than 10A. However, they often do require up to 50W. Anyway, this should not be a problem.
See the power requirements - you are going to have a problem without forced air cooling, and if you want that to be quiet you have to have a fairly large device so that you can move lots of air slowly.
80% efficiency is unacceptable here. With decent quality components, you should be able to get up to about 95% efficiency out of a switching power supply, and perhaps even more. Besides, average power should be a lot less than 50W, more like 12W.
You also run into the possiblity of ground loops - you have one ground path from the computer to the gizmo, and one ground path from the gizmo to the power supply.
A shared ground is absolutely unacceptable in this application. This won't just cause noise, it will cause lots of smoke. All devices powered by a DC adapter assume they are running off of a transformer-type power brick, so they connect the grounds as they see fit. If the grounds are not isolated and you connect one device to another, chances are good their ground arrangements won't be compatible and you will short out something and fry one or both devices. I've personally seen two PlayStations get fried because they were connected to different phases of an AC line of an apparently miswired house and were connected together with a link cable. Apparently, the designers didn't think to isolate the console from ground in the power supply (although it could have been grounded through something else, like the TV).
A good solution would be to have an isolated switching supply (one that uses a transformer) for each output port. This will be almost impossible to do cheaply, though. I doubt such a device could be made for under $150, even with mass production in China.
OK, you may be right. I generally get Crucial, since they manufacture their own memory chips (so there's little chance of getting counterfeits), and they have very reasonable prices. Haven't had a single problem yet. Nevertheless, it's a good idea to avoid testing the ESD protection circuitry on RAM sticks you buy. I've seen way too many people install memory in a careless manner and then wonder why it has errors.
The kingston stick probably died because you mishandled it. I am willing to bet money you weren't wearing an antistatic wrist strap when you opened that package.
If I type in rm -rf / instead of something else, I would notice when it takes a few seconds. It will try to delete/bin,/boot,/dev, and/etc before it gets to/home. I will most likely notice it before it gets there. As a user, it will very quickly figure out it doesn't have write permissions to anything in / and will start deleting the home directory.
rm -Rf / as nonroot will make you give a sigh of relief. As root will be your nightmare.
Yeah, let's see, what's better. As root, the command will delete a few files I can easily reinstall from CD (I tend to notice when an rm command takes more than half a second). As a user, this command will delete all of my documents, my code, my email archive, my MP3 collection, and so on. So, tell me, what's worse?
Do you mean to suggest that home broadband didn't exist AT ALL in 1997?
As you say, consumers didn't have access to it. About the most high-speed connection one could get would have been 56k, and even those were just appearing (and cost hundreds of dollars). Two-way cable and DSL simply weren't available to 99.9% of consumers. Besides which, the DVD standard was developed around 1994-1996. By 1997, everything was finished.
think there is good reason to believe that DVD encryption did not just appear out of thin air but was devised out of legitimate piracy concerns (considering the long history of software and just-developing MP3 piracy).
Considering that the MP3 format was created specifically for DVDs, I think it is highly improbable that the developers thought MP3s could be used for music piracy.
With the inevitable explosion of home broadband looming, I think it is quite obvious why encryption was implemented on DVDs.
Home broadband availability has nothing to do with it, since it wasn't available to most consumers until at least 1999 or 2000 and did not become mainstream until about 2001. Besides, CSS does nothing to close the analog hole. The main reason the system was created was to prevent perfect disc-to-disc copies.
The result? Far tougher encryption is being promoted for the next generation of DVDs. Thanks again, pirates!
The only reason they didn't use tougher encryption to start out was because of US export controls that used to restrict everything stronger than 40 bits. The only reason the rules specified 40 bit encryption was because it was not secure. The rules were changed somewhere in 1997 or 1998. You can thank the EFF and other cyber-rights groups for that one, dumbass.
If you smoke or have a lot of dust in your house, it is hardly a surprise. You cannot clean the read head properly without disassembling it. The thing that gets dirty is usually not the lens, but rather the mirror and other optics inside. It's rather difficult to get to them, although they possible to clean as a last resort. See repairfaq.org for more information.
An Oracle write is about 1/3rd the speed of the linux filesystem (or roughly 1/4 of a pure raw write).
Have you seen the size of a server required to even _run_ Oracle? It's a huge resource hog, and you want to build something like it into the operating system. Not to mention that you admit that the most optimized DB is still 3 times slower than a normal filesystem. Also, part of the reason it's so fast is because of aggressive caching. Sure, it runs fine on a dedicated server with 4 gigs of RAM.
The fileystem can easily tell what's junk from what the user wants to keep based on what directories its stored in.
If you are only using it for documents directories, you might as well just use CVS. It would work just as well and doesn't require overhead from the OS.
You, sir, are an idiot. DVDs have had identical copy protection schemes long before broadband even became available. The original CSS had this exact scheme, in fact, although the keys were so short (40 bits) that it was easy to brute-force.
Ok, explain to me how you write garbage all over the BIOS, or mark blocks as bad on your hard-disk, etc as a normal user?
You can't fuck up the bios from the OS to the best of my knowledge -- even as root. If you can, that should be disabled pronto because there is no legitimate reason to write to the bios from within Linux. In any case, there have been very few viruses that managed to mess with the BIOS -- it's pretty hard.
As far as bad blocks all over the hard drive: did you miss the bit about reimaging it? Bad blocks are a filesystem construct, not a hardware feature.
Root not only has the potential to render the O/S un-usable - it has hardware level access as well.
There's not that much hardware-level stuff you can do from the OS. Plus, writing a virus that can do something destructive to the hardware would take more intelligence than virus writers have.
Not quite so easy to re-flash your bios when your PC won't boot from any devices now, is it?
Most motherboards these days have a backup bootloader BIOS that is in permanent ROM for this exact purpose. Just yesterday, I had a BIOS flash go bad because the floppy disk decided to quit. No problem -- the bootloader automatically started and booted off the floppy.
Anyway, software should not be able to completely destroy any hardware. This would simply be bad hardware design. What if a driver goes nuts and decides to reflash the BIOS?
You are proposing a ridiculous, resource-hogging system. Currently, a file write results in just that -- a file write. With the versioned system, it's more like read the old file, perform the delta computation, write a log file, write the delta, update the tables, and so on. This will slow down the computer a few dozen times -- think about how long calculating the delta for, say, a multi-gigabyte video file would take. Not to mention, how the hell would the filesystem tell apart valuable files and temporary junk that the user doesn't want to keep? This problem should be solved mostly in userspace, not on the filesystem level.
Slashdot Mirror
Apparently, you are too stupid to understand that electromagnetic radiation and nuclear radiation are not the same and do not have the same properties. So please stop telling other people they are wrong.
If there was a dictionary of fingerprint patterns, we wouldn't be using fingerprints to identify people. Hopefully, there are a lot more parameters than in a password. Try to reconstruct the pattern on my finger from a string like v9xc87z6b5684852b79v8x68bxvb78xv. That would be impossible, because that string would not have enough bits to represent the whole pattern.
Have you fucktards completely missed my point about one-way hashes?
Your post just proved that you don't know what you are talking about. I have read the article you are referring to. The only conclusion one can draw from it is that many biometric systems have security problems. It is almost completely irrelevant to my point, except that part about glass doors (which is the method they used).
The article did not demonstrate that data could be extracted from an existing system and the thumb reconstructed from that data. The above article tested mostly low-security and consumer-grade systems, and admitted that the method did not work with the better systems out there.
How the hell can you "spoof other systems", short of making a duplicate thumb? What keeps those perps from just getting your thumbprint off a glass door you touched? Besides, what makes you think you can duplicate the thumb from the biometric system's data file? If whoever made that system had a shred of intelligence, they would use a one-way hash for thumbprints and match based on that.
Think before you post next time.
It's not like they teach soldering in EE programs. Granted, you might find someone experienced, but be careful. Many electrical engineering _students_ these days won't even know how to solder, much less do it well. Watch out.
McVoy IS a tantrum-throwing child. If you knew anything about the whole BitKeeper situation, you would have noticed that. Yeah, he let some projects use his software. WOW. Let's see, it cost him exactly $0.00 and he got lots of free publicity from it. He has all the compassion of a loan shark.
First to file does NOT mean you can take something that someone else invented and patent it. It means that when two people invent the same thing simultaneously, the first one to file the patent gets it. Of course, if the first person publishes the invention, it will become prior art, and therefore ineligible for patent protection.
The first to invent thing was always kooky because it required inventors to keep very good records of when they invented what. When you get a good idea, do you run to a notary public to get it notarized? That's what "first to invent" requires as documentation. Every other country in the world uses "first to file" as the standard for patents.
You are right about the efficiency problems. However, you are mistaken on your definition of a full-wave rectifier. You are probably thinking of a voltage doubler (which are commonly used in SMPSs). A full-wave rectifier is just 4 diodes, and it will produce 120V * sqrt(2) which is approximately 170Vpeak. The waveform from it looks like abs(sin(120*pi*t)). It can then be filtered, and the voltage will be ~170VDC. A voltage doubler charges one capacitor with the positive voltage swing and another with the negative swing, so you double the voltage. Here's a link to a voltage doubler circuit. This is a full-wave rectifier.
I wasn't aware voltage error was that high! Presumably that's for unregulated power supplies?
Most power bricks (the ones that plug directly into an outlet and have a transformer inside) are completely unregulated. The voltage that's printed on them is strictly nominal, and the internal resistance of the transformer is so high that the voltage drops considerably with load. For instance, a randomly picked adapter out of my junkbox says 3V on it. It actually puts out 3.88V no load, which is a 30% error. Generally, these are meant to be used with a regulator inside the device.
A switching power supply that powers my speakers (this one is regulated) is rated for 18V and actually puts out 18.7V no load. This is only a 3.8% error, which is to be expected given the component tolerances in it. But specifying 14.1V for a device implies a 0.3% tolerance, which is far too much to expect from an inexpensive power brick.
The reason a UPS will have a huge transformer is because it has to put out a 50 or 60Hz sine wave, and not DC. Transformers running at 60Hz need far more inductance than ones running at, say, 500kHz (a common frequency for switching power supplies). More inductance means lots of metal and lots of turns of thick wire. At high frequencies, you can get away with small transformers using just a few turns of wire. It's also easier to rectify and filter 500KHz square waves than 60Hz sine waves (although the interference problems are very challenging). A computer power supply is a switching design, which is why it can put out 300W at multiple voltages and not weigh a ton.
Technically, that makes the 240V system a two-phase system. However, in that particular case, it was most likely due to a swapped hot and neutral at the outlet.
I agree, having it run off of 12VDC is a stupid idea. It will not work simply due to the very large required currents.
Besides, you wouldn't generate 170VDC, but 300VDC - you full wave rectify the signal and dump the +170VDC into one cap, and the -170VDC into another.
A full-wave rectifier would produce 170V, not 300V. You can only get 300V if you use a voltage doubler circuit.
Only for a fixed output voltage design running off a high voltage main bar.
Fixed or variable doesn't make much of a difference (it's just a matter of changing the PWM duty cycle), and neither does using a buck/boost topology. The main losses are in the switching elements and the transformers, and are mainly determined by how much you are willing to pay. Of course, running it off of a very low voltage like 12V won't be very efficient due to the high currents involved.
But there are plenty of peripherals that need voltages like 14.1v.
Generally, a voltage requirement like 14.1V means the peripheral will accept anything from 12V to 16V but they don't want you using a third-party adapter. Even switching adapters are usually off by half a volt or more, and regular transformer bricks are +/- 30% or more.
Also, you don't need heavy transformers to put out high currents. Read about switching power supplies sometime. With those, the transformer size depends mainly on switching frequency.
You do bring up a valid point. I think the main reason nobody makes such a device is because it's impossible to make idiot-proof. If someone doesn't set the voltage or polarity properly, they will be very pissed when they fry the device. When you have multiple outputs, this is extremely easy to do.
That is really a very wide voltage swing. If the main internal power bus is less than 26V, it means each output module must be a buck/boost converter (which is more complex and expensive than the other solutions, as well as less efficient).
The main power bus will be about 170VDC, since the power brick will presumably be AC powered.
However, for 2V a 50W supply is 25A - that takes BIG wire to carry, and special connectors.
Generally, devices don't require much more than 10A. However, they often do require up to 50W. Anyway, this should not be a problem.
See the power requirements - you are going to have a problem without forced air cooling, and if you want that to be quiet you have to have a fairly large device so that you can move lots of air slowly.
80% efficiency is unacceptable here. With decent quality components, you should be able to get up to about 95% efficiency out of a switching power supply, and perhaps even more. Besides, average power should be a lot less than 50W, more like 12W.
You also run into the possiblity of ground loops - you have one ground path from the computer to the gizmo, and one ground path from the gizmo to the power supply.
A shared ground is absolutely unacceptable in this application. This won't just cause noise, it will cause lots of smoke. All devices powered by a DC adapter assume they are running off of a transformer-type power brick, so they connect the grounds as they see fit. If the grounds are not isolated and you connect one device to another, chances are good their ground arrangements won't be compatible and you will short out something and fry one or both devices. I've personally seen two PlayStations get fried because they were connected to different phases of an AC line of an apparently miswired house and were connected together with a link cable. Apparently, the designers didn't think to isolate the console from ground in the power supply (although it could have been grounded through something else, like the TV).
A good solution would be to have an isolated switching supply (one that uses a transformer) for each output port. This will be almost impossible to do cheaply, though. I doubt such a device could be made for under $150, even with mass production in China.
OK, you may be right. I generally get Crucial, since they manufacture their own memory chips (so there's little chance of getting counterfeits), and they have very reasonable prices. Haven't had a single problem yet. Nevertheless, it's a good idea to avoid testing the ESD protection circuitry on RAM sticks you buy. I've seen way too many people install memory in a careless manner and then wonder why it has errors.
The kingston stick probably died because you mishandled it. I am willing to bet money you weren't wearing an antistatic wrist strap when you opened that package.
If I type in rm -rf / instead of something else, I would notice when it takes a few seconds. It will try to delete /bin, /boot, /dev, and /etc before it gets to /home. I will most likely notice it before it gets there. As a user, it will very quickly figure out it doesn't have write permissions to anything in / and will start deleting the home directory.
RTFA, dumbass. Robertson was talking about single-user systems.
rm -Rf / as nonroot will make you give a sigh of relief. As root will be your nightmare.
Yeah, let's see, what's better. As root, the command will delete a few files I can easily reinstall from CD (I tend to notice when an rm command takes more than half a second). As a user, this command will delete all of my documents, my code, my email archive, my MP3 collection, and so on. So, tell me, what's worse?
Do you mean to suggest that home broadband didn't exist AT ALL in 1997?
As you say, consumers didn't have access to it. About the most high-speed connection one could get would have been 56k, and even those were just appearing (and cost hundreds of dollars). Two-way cable and DSL simply weren't available to 99.9% of consumers. Besides which, the DVD standard was developed around 1994-1996. By 1997, everything was finished.
think there is good reason to believe that DVD encryption did not just appear out of thin air but was devised out of legitimate piracy concerns (considering the long history of software and just-developing MP3 piracy).
Considering that the MP3 format was created specifically for DVDs, I think it is highly improbable that the developers thought MP3s could be used for music piracy.
With the inevitable explosion of home broadband looming, I think it is quite obvious why encryption was implemented on DVDs.
Home broadband availability has nothing to do with it, since it wasn't available to most consumers until at least 1999 or 2000 and did not become mainstream until about 2001. Besides, CSS does nothing to close the analog hole. The main reason the system was created was to prevent perfect disc-to-disc copies.
The result? Far tougher encryption is being promoted for the next generation of DVDs. Thanks again, pirates!
The only reason they didn't use tougher encryption to start out was because of US export controls that used to restrict everything stronger than 40 bits. The only reason the rules specified 40 bit encryption was because it was not secure. The rules were changed somewhere in 1997 or 1998. You can thank the EFF and other cyber-rights groups for that one, dumbass.
If you smoke or have a lot of dust in your house, it is hardly a surprise. You cannot clean the read head properly without disassembling it. The thing that gets dirty is usually not the lens, but rather the mirror and other optics inside. It's rather difficult to get to them, although they possible to clean as a last resort. See repairfaq.org for more information.
An Oracle write is about 1/3rd the speed of the linux filesystem (or roughly 1/4 of a pure raw write).
Have you seen the size of a server required to even _run_ Oracle? It's a huge resource hog, and you want to build something like it into the operating system. Not to mention that you admit that the most optimized DB is still 3 times slower than a normal filesystem. Also, part of the reason it's so fast is because of aggressive caching. Sure, it runs fine on a dedicated server with 4 gigs of RAM.
The fileystem can easily tell what's junk from what the user wants to keep based on what directories its stored in.
If you are only using it for documents directories, you might as well just use CVS. It would work just as well and doesn't require overhead from the OS.
You, sir, are an idiot. DVDs have had identical copy protection schemes long before broadband even became available. The original CSS had this exact scheme, in fact, although the keys were so short (40 bits) that it was easy to brute-force.
Ok, explain to me how you write garbage all over the BIOS, or mark blocks as bad on your hard-disk, etc as a normal user?
You can't fuck up the bios from the OS to the best of my knowledge -- even as root. If you can, that should be disabled pronto because there is no legitimate reason to write to the bios from within Linux. In any case, there have been very few viruses that managed to mess with the BIOS -- it's pretty hard.
As far as bad blocks all over the hard drive: did you miss the bit about reimaging it? Bad blocks are a filesystem construct, not a hardware feature.
Root not only has the potential to render the O/S un-usable - it has hardware level access as well.
There's not that much hardware-level stuff you can do from the OS. Plus, writing a virus that can do something destructive to the hardware would take more intelligence than virus writers have.
Not quite so easy to re-flash your bios when your PC won't boot from any devices now, is it?
Most motherboards these days have a backup bootloader BIOS that is in permanent ROM for this exact purpose. Just yesterday, I had a BIOS flash go bad because the floppy disk decided to quit. No problem -- the bootloader automatically started and booted off the floppy.
Anyway, software should not be able to completely destroy any hardware. This would simply be bad hardware design. What if a driver goes nuts and decides to reflash the BIOS?
You are proposing a ridiculous, resource-hogging system. Currently, a file write results in just that -- a file write. With the versioned system, it's more like read the old file, perform the delta computation, write a log file, write the delta, update the tables, and so on. This will slow down the computer a few dozen times -- think about how long calculating the delta for, say, a multi-gigabyte video file would take. Not to mention, how the hell would the filesystem tell apart valuable files and temporary junk that the user doesn't want to keep? This problem should be solved mostly in userspace, not on the filesystem level.