I wrote a program that lives on my usb drive. It takes a string of text, turns each character into an integer, applies a seeecreet set of arithmetic on the integer, converts it back into an character that it writes into a new "encrypted" text file. Seeing as all the arithmetic is performed in RAM and the only trace left behind would be the "encrypted" text file. and the program and its guts are not stored on the computers memory at any time. How hard would it be to crack the message? Now lets assume you are smarter than me and do it in a better way. Really, Is it possible to easily crack an "encryption" that is just a long random key stored on an external memory device that performs some transformative operation on a file?
depends enitrely on the 'arithmetic'. if it's complex, doesn't have same outputs for same characters, uses modulus etc it might be OK
>>53665299
>the program and its guts are not stored on the computers memory at any time
Wait, you have a USB drive. USB drive itself cannot run a program because it's just a drive. So you have to run the program on the computer. When you run such a program, it's loaded into the computer's memory.
>>53665405
He could use one of those tiny portable mp3's that have flash and tiny amount of ram capable of running a complex kernel and w/e programs he wants.
>>53665405
Loaded. But it doesn't store anything locally. Wouldn't the memory used by your tiny program be flushed out in a matter of seconds after it was run? How would anyone find it?
>>53665433
If someone has total access to the computer the algorithm is being processed on then youre fucked, it should be piss easy to crack. If what you mean is what are the chances of someone trying to reverse the decryption through bruteforce and assuming the algorithm is strong enough and rare it should be close to impossible.
>>53665299
If you encode the entire string like that, patterns can be discovered and it's easily cracked.
There are plenty of great encryption methods and algorithms for you to use in your programs, use those. No need to re-invent the wheel. Besides, it's gonna be as effective as ROT13, twice.
>look ma, I rolled my own encryption!
It's shit.
>>53665299
Don't roll your own crypto. Don't rely on secrecy to provide security.
Who are you actually encrypting this against? What's to stop someone grabbing your USB? If they get your key and your algorithm it's all over.
Is it a straight 1:1 conversion? What is your plan for preventing repeating patterns and other common attacks? Do you have an IV-like element?
If you use a purely random key that is never reused in whole or in part then delete it (i.e. a one-time pad), it is theoretically impossible to break. But from what I've been told this is impractical for large scale usage and data at rest. It is extremely unlikely you will ever be able to get that level of security using a USB. Writing good cryptograhpy is ridiculously hard even for experts, so it's very likely you will make a mistake in your algorithm that will bring the whole thing crashing down. (See the post above this one)
