I was asked the O() of binary search. I said "log n". They asked "what base?"
... obviously they wanted 2, but O() doesn't work that way - changing base is a constant factor. Sometimes there's a tension between figuring out someone's understanding of the algorithm and someone's understanding of the notation (and math behind it). Of course, I gave them both answers...
> changing base is a constant factor.
Only if the base is constant. For example B-trees have O(log_B n)-time operations, and you can't ignore the B because it is a parameter (hence non-constant).
Re:Big-O Algorithm Complexity Cheat Sheet
by crossroads091 on Saturday May 04, 2013 @04:54PM
(#5656281)
What was the second answer?
2 and the "no need for base in big-O" makes 2 :).
It is good to know what you should know about though.