Oho, in this paper: Kinjal Basu, Art B. Owen: Low discrepancy constructions in the triangle http://arxiv.org/abs/1403.2649 it is shown in theorem 5 that in any integer lattice rotated by an angle with quadratic-irrational tangent, the number N of lattice points in a square of side L, obeys |N - L^2| = O( logN ). Their theorem is based on thm1 in WWL Chen and G Travaglini: Discrepancy with respect to convex polygons, J of Complexity 23,4-6 (2007) 662-672. https://rutherglen.science.mq.edu.au/wchen/researchfolder/prj14.pdf Excellent. This should be highly useful for Wilson's problem. It also demonstrates that my F(x) problem about "digital lines", has logarithmic error bound |F(X)-X| = O(logX) proving the conjecture I'd made, in the case of lines with quadratic-irrational slope. FInally, I remark about the "hierarchical solution method" I'd proposed for Wilson problem, as I said, it did not work in Linfinity norm, only in Lp norms with p finite. But it occurs to me that a modification of the method looks suitable for Linfinity norm. Namely, at some large scale L, you do not just stupidly draw matchup lines long distance from A to B. Instead, you find the least cost "alternating path" from A to B (meaning containing edges alternately not in, and in, the set of matchup line segments drawn so far) and switch it, i.e. make the line segments previously used, be not used, and vice versa. ("Cost" is the length of the longest edge thus introduced.) This approach plausibly will solve Wilson's problem with finite c, or very slow-growing c. -- Warren D. Smith http://RangeVoting.org <-- add your endorsement (by clicking "endorse" as 1st step)