The Solitaire Encryption Algorithm (1999)

schneier.com

39 points·nealrs··8 comments

It's worth noting that despite Schneier's best efforts, the Solitaire algorithm may not be as cryptographically strong as he hoped:

http://www.ciphergoth.org/crypto/solitaire/

(there's also the potential problem that the NSA's supercomputers have grown a lot more powerful in the decade-and-a-half since Cryptonomicon was published)

Thanks for mentioning this. I was wondering if this was "too good to be true" in the technical sense :)

Yes, but apparently it's good enough for short messages of a page or so. Its weakness only matters when an adversary has megabytes to work with.

It is also reasonable to use abbreviations. This not only impacts letter frequency, it also speeds encryption and reduces the efficacy of a Dictionary style attack.

That shows some weakness, but of the sort "less bits of entropy than ideal", not "orders of magnitude more crackable"

If it's more than 6.65 bits of entropy, those are the same thing. There remains the question of whether it is sufficient orders of magnitude more crackable...

Is anyone aware of any newer hand ciphers that have been produced since 1999, especially those subjected to continuous, aggressive high-speed attack throughout their design?

There's Handycipher[0] from 2014. I remember learning about it from Bruce's blog[1], but haven't a clue how secure it is.

[0] http://eprint.iacr.org/2014/257.pdf

[1] https://www.schneier.com/blog/archives/2014/04/a_new_pencil-...

Reply on news.ycombinator.com