Scratch

ohaiderstudios wrote:

I don't feel like having this die just yet, so I shall bump it.
I'm doing some research on S-Boxes right now, they're very interesting, I might try to port one to Scratch.

Ohaider
Chief Theoretical Cryptographer

Mmk, thanks.

roijac wrote:

try to login as roijac WITHOUT editing the project 

Request that you remove this condition. I'm not sure if it's possible to figure out without at least doing that; if not re-implementing your algorithm in another, faster language for brute-forcing (like Python or C  )

Magnie wrote:

blob8108 wrote:

roijac wrote:

try to login as roijac WITHOUT editing the project 

Request that you remove this condition. I'm not sure if it's possible to figure out without at least doing that; if not re-implementing your algorithm in another, faster language for brute-forcing (like Python or C  )

Try logging in with 'asdfasdf'. I was able to. 

Secret to breaking hashes: input long strings. (That's not true for most common hashes, but that's the basic concept)

Actually, it would appear that roijac's hash has a rather high number of collisions... 

Magnie wrote:

blob8108 wrote:

Magnie wrote:

Try logging in with 'asdfasdf'. I was able to. 

Secret to breaking hashes: input long strings. (That's not true for most common hashes, but that's the basic concept)

Actually, it would appear that roijac's hash has a rather high number of collisions... 

He's updated the project and it 'asdfasdf' doesn't collide anymore. 

I noticed!    There are still a few, though, I think... 

SciTecCf wrote:

Maybe, for really simple encryption, you can use my Encoder and Decoder, although when decoded it's all capitals...

Interesting. It only seems to be a substitution cipher with a twist. All you need to do is shift each letter back by "loopcounter" (or shift it back by the number of letters to the left of it) then do a frequency check and you'll be able to figure out the string. Also, a key should be part of the encryption so it can't be decrypted by anyone without the key since currently, the key is the same for everyone.

So basically you want to be able to send a message that will come as random letters and a way to decode it? Does it need to be done in numbers? (sorry, I have no idea what "hashing" is, but I can easily do a coder/decoder.)

Have you seen my project, CipherSaber? It uses RC4 with a salted password. It doesn't qualify since the code is spread out over a whole mess of sprites, but if anybody wants to raid it for ideas/code, feel free. It shows one way to do xor among other things.

Fast Text Encoding might be of interest too.

P.S. If anybody can implement public key encryption in Scratch I will be in awe of their awesomeness forever.

The purpose of a hash is to compare two things. If you hash something twice, you get the same answer. So one common use is instead of storing passwords, you hash the passwords and then store the result. And when someone enters their password, you hash it and see if it matches. That way you don't actually have to keep a list of passwords around, where they could easily be stolen.

I have done some analysis of the time constraints of roijac's and my hash functions:

Platform: Windows 7, Intel i5 CPU, 6 GB RAM

DOWNLOADED NO TURBO
My Hash: 2 minutes
Roijac's Hash: 15 seconds

DOWNLOADED TURBO
My Hash: 3 seconds
Roijac's Hash: <1 second

As you can see, Roijac's hash beats mine hands down when it comes to speed.

Last edited by ohaiderstudios (2012-05-22 18:30:32)

I have nothing to add really...

Just got back from vacation, not feeling too interested.

I'm working on a hash. It uses a bunch of Algebraic formulas.
It is based off of a double input 3DES, but workes with Dec., not Bin.