And how much cheaper things get, 6 months ago a peice of kit like this would have cost you nearly, erm, $5000 [http://slashdot.org/article.pl?sid=01/07/19/15542 16&mode=thread]. Whoops
I'm right behind this guy. Terrorists aren't slashdotters, they're too busy planning terrorism than researching relative crypto standards. Would the average person on the street differenciate between 56 bit DES and 128 bit AES? Would they fsck!
SO accourding to your almighty plan, my algorythm of C(m)=P(m)^K(m%2) would be secure if it had a longer key? Surely what you do with the key is important? An algorythm that simplifies the process 16 fold will need a longer key. I'm so sick of 'I want 2^512 bit cryto because it's a big number'. Hasn't AMD CPU markings taught us anything? It's whats inside that counts, not the number on the outside.
You complete cretin. How can you generalise all crypto accourding to the keylength? Doesn't the algorythm come into it at all? A 32 bit key could be safe, and a 2^32 bit key unsafe, depending on the algorythm. Besides, I don't see you trying to assure the security of a nation.
Firstly, to the cretins with the '40 bits was designed to be hacked' conspiricy theory - yes it was. It used to be balancing act between resources available to legitimate agencies (NSA), and available to ammatures (The general public). It was designed to be hacked, thats why it was designed by them, modded down from a 128 bit version supplied to them.
Secondly, to the '40 bit crypto is nothing' cretinds : What 40 bit crypto? Are you honestly telling me that you beleive key length is the only thing to affect crypto? Doesn't the algorythm count for anything any more? Which takes long 56 bit DES or 56 bit XOR?
I don't think it's too much of a leap of faith to beleive the NSA know more than a bunch of/.ers since they were 40 years ahead of the public at one point, and if you beleive 56 bit DES is as easy as cycling through every key like 56 bit XOR, your just prooving that point. Look at the internal structure of DES, it's a design miracle.
I think you've misunderstood 'odd numbered' kernels. Linux kernels are numbered x.yy.zz, where X is pretty much likely to stay 2 (for the foreseeable future), yy is the major version number, and zz is the minor version number. A development kernel has an odd y, ie 2.1,2.3,2.5, while stable kernels have even y, ie 2.0,2.2,2.4, and in each major version, the minor version increases from 0 upwards, usually with increasing features and stability. Therefor the most stable is even y, and high z.
Who's modding this thread? Why is everyone a troll for having a laugh? Anyway, everyone knows the fastest way to upgrade linux is to alter the version verision number in the source.
Odd numbers are development kernels for new features. I would stick to high revision even number kernels for mission critical stuff, such as 2.0.4x and 2.2.2x.
Mj
Wouldn't it have been better to set the length at the product of two primes? That way there'd only be one possible way of interpretting it, and any species without factorisation isn't worth contacting :
Theres no way http://www.zeosync.com/index.htm was written by anyone who cares about/saving/ space:p
But on a serious note - random data doesn't necessarly mean unique data - all data is psuedo random, and could occour, and if you watch any random data stream long enough, and with a big enough past-data buffer, patterns will emmerge.
Mj
Slashdot Mirror
twice as less?
And how much cheaper things get, 6 months ago a peice of kit like this would have cost you nearly, erm, $5000 [http://slashdot.org/article.pl?sid=01/07/19/15542 16&mode=thread]. Whoops
How is this not off topic?
I guess everyone's uber-cheap PC cracking solution forgot to include the multi billion dollar military campaign needed to aquire the plaintext
I'm right behind this guy. Terrorists aren't slashdotters, they're too busy planning terrorism than researching relative crypto standards. Would the average person on the street differenciate between 56 bit DES and 128 bit AES? Would they fsck!
SO accourding to your almighty plan, my algorythm of C(m)=P(m)^K(m%2) would be secure if it had a longer key? Surely what you do with the key is important? An algorythm that simplifies the process 16 fold will need a longer key. I'm so sick of 'I want 2^512 bit cryto because it's a big number'. Hasn't AMD CPU markings taught us anything? It's whats inside that counts, not the number on the outside.
You complete cretin. How can you generalise all crypto accourding to the keylength? Doesn't the algorythm come into it at all? A 32 bit key could be safe, and a 2^32 bit key unsafe, depending on the algorythm. Besides, I don't see you trying to assure the security of a nation.
And appearantly the majority of /.ers too.
Secondly, to the '40 bit crypto is nothing' cretinds : What 40 bit crypto? Are you honestly telling me that you beleive key length is the only thing to affect crypto? Doesn't the algorythm count for anything any more? Which takes long 56 bit DES or 56 bit XOR?
I don't think it's too much of a leap of faith to beleive the NSA know more than a bunch of /.ers since they were 40 years ahead of the public at one point, and if you beleive 56 bit DES is as easy as cycling through every key like 56 bit XOR, your just prooving that point. Look at the internal structure of DES, it's a design miracle.
I think you've misunderstood 'odd numbered' kernels. Linux kernels are numbered x.yy.zz, where X is pretty much likely to stay 2 (for the foreseeable future), yy is the major version number, and zz is the minor version number. A development kernel has an odd y, ie 2.1,2.3,2.5, while stable kernels have even y, ie 2.0,2.2,2.4, and in each major version, the minor version increases from 0 upwards, usually with increasing features and stability. Therefor the most stable is even y, and high z.
It's sad when people can't have a laugh, especially when it's concerning a /hobby/ which was started for /fun/ :p
Because it's dripping blood from whatever it just cut due to it's sharpness, it looks like it's bleeding.
Who's modding this thread? Why is everyone a troll for having a laugh? Anyway, everyone knows the fastest way to upgrade linux is to alter the version verision number in the source.
Odd numbers are development kernels for new features. I would stick to high revision even number kernels for mission critical stuff, such as 2.0.4x and 2.2.2x. Mj
No 2.6.x yet? :p (only kidding - well done everyone)
Wouldn't it have been better to set the length at the product of two primes? That way there'd only be one possible way of interpretting it, and any species without factorisation isn't worth contacting :
Theres no way http://www.zeosync.com/index.htm was written by anyone who cares about /saving/ space :p
But on a serious note - random data doesn't necessarly mean unique data - all data is psuedo random, and could occour, and if you watch any random data stream long enough, and with a big enough past-data buffer, patterns will emmerge.
Mj