- #1
jrtayloriv
- 5
- 0
I have been reading a bit about cryptography recently, and there are a few things that I haven't understood so far:
#1) I read that since compression removes redundant information, it is a randomizing function. Does that mean that I
would have more high quality random data afterwards if I used 'dd' to dump a few hundred megabytes of random data into a
file, and then compressed the file w/ a good compression algorithm? Or did I misunderstand what they meant by compression? Does this only work for certain types of compression? Which is the best one to use?
#2) I read that XORing data that is not completely random, will actually decrease the randomness even further. Why is
this? For example, I have a built in hardware random number generator in my laptop, but I don't trust the multibillion dollar corporation that manufactured it. In case they have engineered it to have some type of "backdoor", I was thinking about taking the output of it (perhaps every 1 KB), and XORing each 32-bit word of it with a number (which would change from block to block) from /dev/random.
If there is a backdoor built in (i.e. the data isn't completely random), would XORing these numbers together cause this pattern to be more pronounced? That is, would I make it even worse by doing this?
Thanks,
jrtayloriv
#1) I read that since compression removes redundant information, it is a randomizing function. Does that mean that I
would have more high quality random data afterwards if I used 'dd' to dump a few hundred megabytes of random data into a
file, and then compressed the file w/ a good compression algorithm? Or did I misunderstand what they meant by compression? Does this only work for certain types of compression? Which is the best one to use?
#2) I read that XORing data that is not completely random, will actually decrease the randomness even further. Why is
this? For example, I have a built in hardware random number generator in my laptop, but I don't trust the multibillion dollar corporation that manufactured it. In case they have engineered it to have some type of "backdoor", I was thinking about taking the output of it (perhaps every 1 KB), and XORing each 32-bit word of it with a number (which would change from block to block) from /dev/random.
If there is a backdoor built in (i.e. the data isn't completely random), would XORing these numbers together cause this pattern to be more pronounced? That is, would I make it even worse by doing this?
Thanks,
jrtayloriv