Or possibly not. I'm reminded of Sarah Flannery's idea that matrix multiplication might be a good one-way function for crypto. Turned out that in that context,

Or possibly not. I'm reminded of Sarah Flannery's idea that matrix multiplication might be a good one-way function for crypto. Turned out that in that context, there was a cheap inversion that wasn't immediately obvious. Her memoir/math book about this (and learning mathematical cryptography) co-written with her father is a good introduction to a lot of mathematical cryptography ideas: http://www.amazon.com/exec/obidos/ASIN/1565123778/donhosek http://www.amazon.com/exec/obidos/ASIN/1565123778/donhosek