Ok, I finally figured your code out (and yes, this time I did). I had tried very hard to figure out what was wrong with it, but finally I just put it aside and worked on something else for awhile. Now that I'm done with that, I looked at it with "fresh eyes" and realized what was wrong. Very simple question, yes I made a thread about it, but here it goes. In the state, (a 4*4 block of bytes to anyone else out there), the encryption routine performs shifting right to left. Your encryption/decryption table lookups were all fine, but I noticed that they were exact opposite of each other. The encryption routine accessed them (in a visual, 4*4 box kind of a way) in a diagonal down to the right, which surprised me because I thought the shifting went from right to left. Your decryption routine of course did the opposite, complementing each other and providing the balance. I can see both sides of the problem, where what I think is supposed to happen accesses them "after shifting" and where your side acts like the data has not been shifted before grabbing the bytes. I got a little mixed up, could you clear up this little question of if you pretend like it's before or after shifting that you access them?

PS: (I'll probably be able to delete this thread if u delete yours and then I'll delete my post) Sry bout that last bit of confusion (or should I say byte). Ok, sry for that too. Thanks again if u or anyone helps me!

what's this? looks like response but it's in separate thread. should i move it to some existing thread?

It's a reply from a much earlier thread, but that one was just a side-question itself. It's just a specific question that can be deleted if it's answered, which it probably will be. Don't have any other way of contacting him (he' s probably happy, I'm sure he's tired of me now), so I know he'll read this and then it can be deleted if he answers or not.

Who is 'he'? Can I assume you mean me? It seems I was the only person to give you any Rijndael code so perhaps you do mean me?

If you are talking about the code that I posted in another thread then I can answer that I never shifted the data in memory. It is by far more efficient to work on the data in place and and shift it in the registers either when loading or when storing. Thus the sequence is something like: grab the data in shifted pattern - apply encoding - store the data in normal pattern.

Thanks, that occured to me shortly after posting it. I just decided on three seperate exported functions for each cipher, and it's coming together now. Thanks for helpin' again, I don't think I'll need to read up on anything more or freak out over things not working again (at least I hope not)