What modern OS's can these things run? e.g. Solaris 9 or 10 or any newer Linux? (RH6 seems a little out of date and I'd be worried about security if using it for an internet server)
A survey (sorry, don't remember where, IIRC it was posted on NTBugTrack) said there's about 90% success rate, or 10% of computers have problems after installation.
Corporate respondents reported about 6% failure rate, and small bus & induhviduals were close to 12%.
FWIW, libertarianism has been co-opted by the far right. See Ayn Rand/Rand Institute, etc. (basically, cooperations are free to screw anyone over w/o gov't supervision; e.g. liberty for cooperations)
I don't think the current political def fits many slashdotters.
It's not to take load off the CPU, it's because the GPU is MUCH faster for this kind of thing.
According to the linked article on Tom's Hardware, the GPU's can do 40GFlops (40 billion floating-point operations per second) versus 6GFlops for a fast CPU!
Replace that 2 with another N if there the qbits can be in more than 2 possible positions. e.g. if they can be in 4 positions, and only one will work, then it's O(4^n). I think.
So, where do you store the pads if you want to keep the encrypted messages??
Thats why I said OTP isn't really secure. (In practice. I know it's secure in theory.)
What do you do if you want to keep your laptop hard drive encrypted with OTP? You'd have to carry around the OTP on a DVD, and then what do you do when you update a file on the HD? You need a new OTP for the whole drive. Or you need a OTP for each file on the HD each time the file changes!
And if you get arrested, the spooks will find the DVD with the pad(s).
Also note that Quantum-generated OTPs won't solve any of these problems. It only would solve the key-exchange problem.
OTP is a LOT less practical for general use than people here seem to think.
So quantum computing should be able to do the "large nubmer factoring" exercise necessary to crack the key...
That's the idea, and as of last year, people have been able to build 4-qbit quantum computers that can instantly factor a 4-bit number.
Obviously that's not very useful as I can factor a 4-bit number in my head in about a second, but if larger q computers can be stabilized, then factoring larger numbers instantly will be possible.
Even if all else fails we will always have the one time pad which is completely unbreakable (when proper pad discipline is observed) albeit somewhat cumbersome in practice.
OTP is EXTREMELY cumbersome...
Not only do you have the key exchange problem, but if you want to store the encrypted text, you also have to keep a copy of the pad somewhere, for EACH document.
Therefore, OTP isn't really unbreakable either.
With key-based crypto, you can keep a password-protected copy of the key(s) on a usb token or whatever, and you only need one key for infinite documents.
OTP could partially be helped by using commonly available documents, images, binaries, whatever as the pad, but then you increase the chances of someone else finding the pad, and you still have to store an index of which document uses which pad somewhere.
Slashdot Mirror
What modern OS's can these things run?
e.g. Solaris 9 or 10 or any newer Linux?
(RH6 seems a little out of date and I'd be worried about security if using it for an internet server)
Thanks
BTW, one of the reasons for worse luck outside corps is probably more spyware/malware infection.
:P
Lots of people have said if you're infected, SP2 will probably kill the machine.
Sounds like a feature to me though
That's one of the new security features in SP2, not a bug.
A survey (sorry, don't remember where, IIRC it was posted on NTBugTrack) said there's about 90% success rate, or 10% of computers have problems after installation.
Corporate respondents reported about 6% failure rate, and small bus & induhviduals were close to 12%.
Not only are LED lights expensive, they use much more electricity than equivalent florescent or compact florescent lighting.
(they are more power-efficient than incandescent though)
My niece was doing an experiment like that in school this year; her plants seemed to like rock music better than classical.
It probably just needs more silkworms.
Hate to burst your bubble, but there's a Category 5 hurricane (Javier) coming towards us.
Well, it was a Cat 5 last night anyways. It's slowing down now. Still will probably be VERY wet and somewhat windy if/when it gets here.
Cray's new product IS a Linux cluster!
Only advantage it (currently) has is a custom HyperTransportInfiniband bridge.
And yes, Cray is trying to claim it's different than a "Linux cluster".
I've tried sniffing on my Motorola cable modem several years ago, and only saw DHCP broadcasts.
Maybe there's a way to put it in promiscous mode, but it's not the default.
(Yes, I did put my NIC in promiscous mode)
There was a destructive internet worm recently.
It attacked PC's via a hole in BlackICE firewall.
After reproducing for a little while, it began randomly overwriting sectors on the HD. Eventually your OS (and probably a lot of data) would be fubar.
URL: http://www.f-secure.com/v-descs/witty.shtml
I guess no one has noticed, but coral doesn't seem to cache the images!
FWIW, libertarianism has been co-opted by the far right. See Ayn Rand/Rand Institute, etc.
(basically, cooperations are free to screw anyone over w/o gov't supervision; e.g. liberty for cooperations)
I don't think the current political def fits many slashdotters.
the cost of professional studio DSP solutions which can run into the high five-figure range
How is that less than a consumer 3d card???
And no, this isn't for gamers. RTFA. It's about doing audio stuff like measuring and analyzing the acoustic properties of concert halls, etc.
This sounds very much like the reactors the Chinese are trying to deploy, covered in Wired this month.
Price
Well its certainly more interesting than having to do them by hand!
It's not to take load off the CPU, it's because the GPU is MUCH faster for this kind of thing.
According to the linked article on Tom's Hardware, the GPU's can do 40GFlops (40 billion floating-point operations per second) versus 6GFlops for a fast CPU!
Replace that 2 with another N if there the qbits can be in more than 2 possible positions.
e.g. if they can be in 4 positions, and only one will work, then it's O(4^n). I think.
So, where do you store the pads if you want to keep the encrypted messages??
Thats why I said OTP isn't really secure. (In practice. I know it's secure in theory.)
What do you do if you want to keep your laptop hard drive encrypted with OTP?
You'd have to carry around the OTP on a DVD, and then what do you do when you update a file on the HD? You need a new OTP for the whole drive. Or you need a OTP for each file on the HD each time the file changes!
And if you get arrested, the spooks will find the DVD with the pad(s).
Also note that Quantum-generated OTPs won't solve any of these problems. It only would solve the key-exchange problem.
OTP is a LOT less practical for general use than people here seem to think.
I'm too tired to think right now, but IIRC the problem is that the atoms or molecules are hard to keep in the correct position(s).
So I'd guess it'd be an O(2^n) problem, but like I said, I can't think right now.
Ask Zaphod Beeblebrox!
4-qbit q computers have already been built (as of 2002).
e akers.html#walterdaugherity
/., IBM built a 5-qbit computer in 2000.
(They can factor a 4-bit number.)
Not very useful, but a great demonstration.
These claims were presented by a Physics prof at DefCon 10.
http://www.defcon.org/html/defcon-10/defcon-10-sp
And, According to
That's the idea, and as of last year, people have been able to build 4-qbit quantum computers that can instantly factor a 4-bit number.
Obviously that's not very useful as I can factor a 4-bit number in my head in about a second, but if larger q computers can be stabilized, then factoring larger numbers instantly will be possible.
OTP is EXTREMELY cumbersome...
Not only do you have the key exchange problem, but if you want to store the encrypted text, you also have to keep a copy of the pad somewhere, for EACH document.
Therefore, OTP isn't really unbreakable either.
With key-based crypto, you can keep a password-protected copy of the key(s) on a usb token or whatever, and you only need one key for infinite documents.
OTP could partially be helped by using commonly available documents, images, binaries, whatever as the pad, but then you increase the chances of someone else finding the pad, and you still have to store an index of which document uses which pad somewhere.