Academically it might be correct, but shipping real features will in most cases be more important than hitting some text book definition of correctness.
Academically it might be correct, but shipping real features will in most cases be more important than hitting some text book definition of correctness.
Sure, you’re right: pragmatics, in practice, are more important than theory.
But you’re assuming that there is a real contradiction between shipping features and RESTful design. I believe that RESTful design can in many cases actually increase feature delivery speed through its decoupling of clients and servers and more deeply due to its operational model.
its decoupling of clients and servers.
Notice that both of those are plural words. When you have many clients and many servers implementing a protocol a formal agreement of protocol is required. REST (which I will not claim to understand well) makes a formal agreement much easier, but you still need some agreement. However when there is just one server and just one client (I'll count all web browsers as one since the browser protocols are well defined enough) you can go faster by just implementing both sides and testing they work for a long time.