I'm no good at math, as readers have constantly had to point out. But among the many suggestions I've received on how to ameliorate the gerry-mandering problen, the tantalizing possibility of redistricting by a neutral computer program seems to me to be among the most promising. Here's a web-page that presents various options that would take the process of redistricting out of the hands of partisans and into a completely neutral computer algorithm. The criteria used can be debated on a state by state level. But geographical contiguity, equality in population, and compactness are the top three. Avoiding gratuitous racial discrimination should also play a role. But the point is: you fight over the criteria and the weight given to them; then you let the computer do the rest with census data.

To see what might happen, here's a map of North Carolina's districts:


Now, here's a new electoral map drawn by a computer, with a particular set of algorithms plugged in:


I know which one I prefer. There are still some vagaries. Given certain criteria, computers can come up with a variety of solutions, and there's a danger that one political party in power might keep running the computer to get their preferred result. One way of dealing with this might be to require a super-majority in state legislatures to approve the new computer-driven districts to ensure that both parties have some buy-in leverage. Another way is to do what Iowa does and hand over redistricting power to a neutral body, separate from but answerable to the legislature.

We've come to accept the census-mandated, computer-driven reapportionment every ten years or so. Why not let computers do the other heavy-lifting to define the boundaries of the actual districts? This is the twenty-first century. There's no reason we shouldn't use computer power to improve democracy.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.