This is not easy to do. Ernie Cohen had internalized this skill, and he could quickly zero in on the monotonic or locally stable predicates to exploit in his protocol explorations. Some people live in this way of thinking so completely that they may not even know how to articulate it, like fish in water. (Case in point: the phrase "global shared memory fiction" was pointed out to me by Marc Brooker. Sometimes it takes a slight outsider to name what insiders take for granted.) The practical takeaway for TLA+ modeling is this: when you write your guards, ask yourself whether the information the guard relies on could become stale, and if so, whether the guard is still safe to act on. If you can make your guards depend on monotonic or locally stable predicates, your protocol will be more robust, more concurrent, and closer to a correct distributed implementation.
Автор: Елена Торубарова (Руководитель российской редакции)，详情可参考搜狗输入法下载
。关于这个话题，Discord老号,海外聊天老号,Discord养号提供了深入分析
重庆医师下班途中跪地查看CT片获赞
Tragic number: N/A，更多细节参见有道翻译