« first day (1499 days earlier)      last day (3169 days later) » 

mikeazo
mikeazo
12:15 PM
We should lose the "Beta" label soon!
60
Q: Design-Independent Graduation is on for early September!

AnaThis is a follow-up to an earlier discussion Feedback Requested: Design-Independent Graduation. Like the title implies, design-independent graduation will go into effect and become a regular practice the second week of September! This means all sites which are cleared for graduation going forwa...

discussion featured election beta-sites graduating-sites
And have migration paths setup!
 
yyyyyyy
yyyyyyy
12:25 PM
@mikeazo: Migration is great. Finally we can stop writing comments about how they should've posted on Stack Overflow and just move the question there...
oh, and @mikeazo: In fact, you beat me by 39 seconds, assuming the mouse-over is correct. :-)
 
 
7 hours later…
SEJPM
SEJPM
7:41 PM
I just saw Schneier's post on this.
It's indeed pretty funny
 
SEJPM
SEJPM
8:17 PM
We already had some snake-oil here
Maybe we (Crypto.SE community) can build something up from that
maybe containing some fancy number theory
"Here's a standardized modulus. The authors assure not to own the factorization as of time of this writing."
and of course longer keys are always more convincing and fancy diagrams are always helpful
"AES-256 has been known to be weak against adversaries with capability to do 2^256 evaluations of AES, because of this we propose to use 8x AES-256 providing 8x more security!"
 
mikeazo
mikeazo
@yyyyyyy Gotta say I'm a little bummed you got +15 and I only got +6 crypto.stackexchange.com/questions/27579/…
I didn't think my answer was that bad, so maybe by beating me by 40 seconds...
 
SEJPM
SEJPM
@mikeazo +15 vs +7 now
 
mikeazo
mikeazo
@SEJPM "Distinguishing the C code from randomly generated code should be hard." I like it.
@SEJPM :)
 
SEJPM
SEJPM
one could also include some fancy assumptions: "Assuming "Never Say Anything" holds, finding the keys is provably hard."
short term for the problem would be NSA of course, just like DLP for discrete log prob
 
mikeazo
mikeazo
No Such Algorithm
 
SEJPM
SEJPM
8:30 PM
@mikeazo even better :)
 
mikeazo
mikeazo
Cipher-Cipher-Chaining-Padded mode or CCCP mode. We can do it per character, XOR the output of the previous encrypted character with the current encrypted character c_i = E(p_i)\xor c_{i-1}. Of course the first encrypted character is xored with a random, 128-bit One Time Pad.
The Padding in CCCP is just to throw people off :)
 
SEJPM
SEJPM
@mikeazo Obviously superior to ECB as patterns are obviously hidden by the xor
Perfectly secure under the NSA assumption
even better: each block is secure assuming E is a PRP
if each block is secure -> all blocks need to be secure!
@mikeazo well.... we would need padding in case p_i isn't a 16-byte block...
 
 
3 hours later…
yyyyyyy
yyyyyyy
11:32 PM
@mikeazo Yeah, I'm highly confused as to why that answer is upvoted so much. I've definitely written better things with almost no response, and suddenly people care about this?!
 

« first day (1499 days earlier)      last day (3169 days later) »