Scott Fluhrer edited untitled.tex  almost 9 years ago

Commit id: 7c93a0974a0fa0c634221aa37ffbc798e1c39728

deletions | additions      

       

Here, we don't get any regular pattern, and this value would, at first glance, appear random. And, in fact, if we look at the digits of $k+nr$ (for modest random $r$) expressed in base 48, we don't detect any correlation between the digits of that and the digits of $k$ expressed in base 48. This implies that blinding with this value (in base 48) is likely as effective as blinding based on a random prime in base 32.  Of course, once we consider nonpower-of-2 bases, we lose the two advantages that we formally had; let us examine how bad that makes the situation.  \begin{itemize}  \item Implementing radix of non-power-of-2  \item Generating exponents into non-power-of-2 base