Merge pull request #49 from ctkelley/main

skip CI

docs/src/Details/N2Work.md

@@ -6,7 +6,7 @@ in the triangular solvers are done in-place. We terminate on small residuals.
 
 The premise behind IR is that reducing the $O(N^3)$ cost of the
 factorization will make the solve faster because everything else
-is $O(N^2)$ work. It's worth looking to this.
+is $O(N^2)$ work. It's worth looking into this.
 
 We will use the old-fashioned definition of a FLOP as an add, a multiply
 and a bit of address computation. So we have $N^2$ flops for any of