After today’s disastrous decision, here are some things that line-drawers could do in the future:

• Instruct a computer algorithm to generate huge numbers of maps that comply with all nonpartisan criteria and produce as large and durable an advantage as possible for the line-drawing party. Then pick an actual plan from this array of potential gerrymanders. This plan will be close to impregnable if it’s challenged on nonpartisan grounds. But it will still massively benefit the line-drawing party, probably more than any human-drawn map could.
• Revise districts after each election to optimize their performance in the next election. Any districts slipping away from the line-drawing party could have some copartisans added to them. Any districts becoming overly safe could have some copartisans subtracted. Decennial redistricting, in other words, could become a thing of the past. Redistricting every two years is so much more effective.
• Design noncontiguous districts in order to avoid the constraints of political geography. A state with many Democrats concentrated in cities (like my Illinois) could join clusters of urban Democrats with slightly smaller clusters of rural Republicans hundreds of miles away. These clusters wouldn’t have to be connected since no federal law, and no other law in many states, mandates contiguity. A state could even adopt entirely nongeographic districts, e.g., by assigning a representative (and sufficiently numerous) sample of the state’s population to each district.