News for hackers. Stuff that matters.
Re:If your cipher were secure, this image wouldn't have repe... by xtacy on Friday February 28, 2020 @07:51PM (#22447820)
It's a funny coincidence that I was just reading this book (https://toc.cryptobook.us/), Figure 4.5 on book page 102, that talks about this exact issue.

I doubt the author realises that the system must be resistant against a class of attacks (in this case, I believe a chosen plaintext attack). From reading the thread, and comparing it to the example in the book, it seems like the non-uniform patterns in the output highlight a possibility of a CPA.

And of course, the author wants a fully implemented + concrete attack instead of pointing to what's an obvious flaw to the crypto community.

Re:If your cipher were secure, this image wouldn't have repe... by jascii on Friday February 28, 2020 @07:58PM (#22447860)
At some point in the thread the author complains that it isn't fair to point out the repetitive patterns "because the clear-text is all zeroes and this would be masked by actual data"
Re:If your cipher were secure, this image wouldn't have repe... by jschwartzi on Friday February 28, 2020 @08:19PM (#22447978)
My first thought was "what if I wanted to encrypt and transmit all-zeroes as a message?"
Re:If your cipher were secure, this image wouldn't have repe... by sidewndr46 on Friday February 28, 2020 @08:21PM (#22447997)
So the author tried to encrypt the null byte, repeated thousands of times? That seems like the most pointless test I could dream up for any processing algorithm.

Computes all the primes less than 0? Already.

Achieve compression approaching 100% on a string of zeroes? Easy.

Losslessly compress an image with all pixels having opacity of 0? Easy

Re:If your cipher were secure, this image wouldn't have repe... by henrikschroder on Friday February 28, 2020 @09:17PM (#22448277)
If an all-null message results in a simple/repeating ciphertext, it means that there exists at least one message whose ciphertext leaks information, and that in turns means that there are probably more messages like it, and it's also probable that all messages leak info in varying degrees.
Re:If your cipher were secure, this image wouldn't have repe... by RL_Quine on Friday February 28, 2020 @08:37PM (#22448067)
It’s a good demonstration of failure that might otherwise be masked
Re:If your cipher were secure, this image wouldn't have repe... by nkrisc on Saturday February 29, 2020 @10:32AM (#22451018)
What good is encryption that only works for some inputs?
Re:If your cipher were secure, this image wouldn't have repe... by fanf2 on Friday February 28, 2020 @08:55PM (#22448169)
I am slightly disappointed that the crypto book did not use a picture of Tux to demonstrate ECB https://blog.filippo.io/the-ecb-penguin/