Most of the new draft of the Concurrent Programs paper has to do with trying to specify problems and solutions in synchronization via an atomic “compare and swap” operation. Even these operations are surprisingly complicated once put under the microscope
