I need some help regarding one of the steps in Pollard's p-1.
What I understood: We are assuming that there exists $p$, a prime factor of $n$, such that $p-1$ is $B$-powersmooth for some $B$ (not $B$-smooth!) and that we're trying to build a number $M$ (to be used as an exponent) such that $(p-1)|M$. Most commonly this is done in two ways:
- $M = B!$
- $M = \prod {q^a} $, for all primes $q <= B$ and $a$ sufficiently large.
Let's focus on the second. My confusion comes from the value of a. Using the main assumption we made ($p-1$ is $B$-powersmooth) it seems clear that $a = \lfloor log_q{B} \rfloor$. However, some sources (including Wikipedia) use $a = \lfloor log_q{n} \rfloor$. Since we can assume $n >> B$ this works too but makes $a$ unnecessarily large.
At first I thought it's just an error on Wikipedia and I made a correction, but I found several revisions in page history that change this both ways ($B$ to $n$ and back), with no real discussion, which makes me unsure. Can someone explain why $n$ makes more sense than $B$ here, or confirm that $n$ was just an error?