Edward Felten, concluding a critique of the Posner/Becker reaction to the Grokster ruling, argues that it may be time for judges to start taking computer science classes:

As we have seen so many times, bad computer science leads to bad law. Posner seems to miss this, but Becker's stance shows appropriate caution.

One criticism of law and economics is that it works well in a seminar room but may lead to dangerous overconfidence if applied to a hard case by an overworked, generalist judge. One solution is to teach judges more economics, and economic seminars for judges have proliferated. Perhaps the time has come to run seminars in computer science for judges.

On the other hand, as a reader over @ Freedom to Tinker points out, a little education can be a dangerous thing:

The idea of trying to make judges into experts in computer science terrifies me. Trying to make them into experts in economics has resulted in some good work, but also in half-baked opinions based on terribly narrow interpretation of economic theories that might or might not model actual human behavior, with unexamined premises littering the landscape. Even expert economists get large-scale decisions terribly wrong on a regular basis — just ask Jeffrey Sachs or Joe Stiglitz.

The same goes for computer scientists — the past 40 years will give you a good idea of how accurate, en masse, CS folks have been at predicting the impact or even the broad evolution of their field. (Some predictions spot on, others not even close)

If there were some kind of seminar in computer science for judges, so that
they could avoid looking completely at sea, what principles would you want to teach them?

Perhaps you could post a reading list: top five (ten?) articles that lawyers, judges and law professors need to read to understand the technical limits that you describe?