My point is that "procedure calls" have always covered up failure modes and making them "remote" papers over all of the problems of distributed computing. The

My point is that "procedure calls" have always covered up failure modes and making them "remote" papers over all of the problems of distributed computing. The other problem with "procedure calls" is they are imperative and grow without any constraints on their implementation without very careful design and review. The functionality of the "procedure" is unbound and has unknown dependencies and side effects.