One time pad encryption is proven uncrackable, provided the key is only used once and is truly random. i.e. a key as long as the message being encrypted has possible solutions for all possible combinations of intelligible messages.
But PC's random number generators are pseudo random, sourced with the same seed and they produce the same series of 'random' bytes. What however if we intermix that with a external random factor, such generating blocks of random bytes, with periodic read current cpu temperature as a factor in setting the next random blocks seed ... or using that 'external measure' as some other form of juggling around the random byte series. Does that not transform pseudo to real random?
For long term privacy, a one time pad type choice seems to fit the bill. After all a few decades ago, what then seemed like difficult/expensive to crack methods are today trivially cracked.
pseudo -> real random
pseudo -> real random
[size=75]( ͡° ͜ʖ ͡°) :wq[/size]
[url=http://murga-linux.com/puppy/viewtopic.php?p=1028256#1028256][size=75]Fatdog multi-session usb[/url][/size]
[size=75][url=https://hashbang.sh]echo url|sed -e 's/^/(c/' -e 's/$/ hashbang.sh)/'|sh[/url][/size]
[url=http://murga-linux.com/puppy/viewtopic.php?p=1028256#1028256][size=75]Fatdog multi-session usb[/url][/size]
[size=75][url=https://hashbang.sh]echo url|sed -e 's/^/(c/' -e 's/$/ hashbang.sh)/'|sh[/url][/size]
In theory, yes, a one-time pad is unbreakable. However, what you propose still falls short of true randomness. As you point out, PRNGs (pseudo-random-number generators) yield the same sequence of numbers every time; the seed only alters where in the sequence one begins. Interrupting the sequence with "random" reseedings as you propose will make things more difficult for the cryptanalyst, but traces of the underlying deterministic sequence will still show through, giving the cryptanalyst a toehold to work with.
Usually, the problems with one-time pad schemes are more practical than theoretical. For example, the recipient of the encrypted message must also have a copy of the key. How is one going to guarantee the security of that copy of the key? If it can be secured physically, that begs the question of why that physical security wasn't sufficient for the original document. We could encrypt the key...but that puts us right back at square one.
In addition to the problem of securing and distributing the key, there is the fact that the key has to be at least as long as the document being encrypted. This may not be an issue in some situations, but in many cases it IS an issue, which is why encryption schemes that can use a key that is much shorter than the message being encrypted are so widely used.
I'm not claiming that one-time pads are useless. I am only suggesting that the practical issues somewhat limit the range of scenarios where they are usable. Given a good, truly random physical process such as electrical noise to generate the key, and a low volume of messages that need to be encrypted -- so that the space requirements for storing the decryption keys do not become a problem -- one-time pads are great. Especially if one isn't actually "sending" the encrypted messages but merely storing them for oneself, as this eliminates the key distribution problem.
Usually, the problems with one-time pad schemes are more practical than theoretical. For example, the recipient of the encrypted message must also have a copy of the key. How is one going to guarantee the security of that copy of the key? If it can be secured physically, that begs the question of why that physical security wasn't sufficient for the original document. We could encrypt the key...but that puts us right back at square one.
In addition to the problem of securing and distributing the key, there is the fact that the key has to be at least as long as the document being encrypted. This may not be an issue in some situations, but in many cases it IS an issue, which is why encryption schemes that can use a key that is much shorter than the message being encrypted are so widely used.
I'm not claiming that one-time pads are useless. I am only suggesting that the practical issues somewhat limit the range of scenarios where they are usable. Given a good, truly random physical process such as electrical noise to generate the key, and a low volume of messages that need to be encrypted -- so that the space requirements for storing the decryption keys do not become a problem -- one-time pads are great. Especially if one isn't actually "sending" the encrypted messages but merely storing them for oneself, as this eliminates the key distribution problem.
I have always found that atmospheric noise or radioactive decay are better 'natural' sources of randomness than electrical noise (generally man-made with buried patterns of usage/strength, phase).
JMHObservations
8Geee
JMHObservations
8Geee
Linux user #498913 "Some people need to reimagine their thinking."
"Zuckerberg: a large city inhabited by mentally challenged people."
"Zuckerberg: a large city inhabited by mentally challenged people."