Somewhere it was recommended to perhaps do Baillie-PSW after Miller-Rabin. That as a belt-and-suspenders approach.
But as I read it, Baillie-PSW seems merely a pairing of Lucas to Miller-Rabin.
Which makes the first paragraph above to seem semi-redundant.
Say I have Miller-Rabin already coded (in Forth). Ought I proceed to code Baillie-PSW? Or ought I instead code Lucas to follow Miller-Rabin?
Or am I missing a subtle nuance somewhere?