This paper re-examines the problem of distributed consensus with the aim of improving performance and understanding.

I’ve been dividing the continent again. Back in the summer of 2000, on a coast-to-coast car trip, I began wondering about the Great Divide, the line that traces the spine of North America, separating the Atlantic and Pacific watersheds. How would you identify that line out on the landscape, or on a topographic map? Eventually I wrote some programs to explore the problem, and then an American Scientist column. The column became a chapter in my book Group Theory in the Bedroom.