I'm working on a problem from Stein and Shakarchi's Real Analysis, and it asks to construct a measurable subset $E\subset [0,1]$ such that for any non-empty sub-interval $I$ in $[0,1]$, both $E\cap I$ and $E^c \cap I$ have positive measure. The problem gives the hint "to consider a Cantor-like set of positive measure, and add in each of the intervals that are omitted in the first step of its construction, another Cantor-like set. Continue this procedure indefinitely."
I don't quite understand the hint. Am I just taking the collection of open subsets of $[0,1]$ removed from the Cantor-like set? How could these be another Cantor-like set, since they're open?