Architecture War Stories

Stefan is a great entertainer, but when it comes to distributed system architecture, I am not quite sure I would follow him beyond the few jokes he cracks.

For instance, if he had spent a bit of time drawing the REST "meta"model, he could have easily predicted that the world would remain action oriented (Web APIs) instead of becoming Resource Oriented. I am glad some people get to charge $$$ to convert a perfectly understandable action verb into an HTTP verb + a (weird) noun (such as login into POST + session), but beyond that REST-for-distributed-systems is pretty much the ultimate architectural failure of the past decade, just as laughable as some of the examples he provides.

Actually, the best way to understand what you are doing is to draw its metamodel. For instance, when we are writing any kind of software, we somehow encode states, types, actions and relationships (STAR), yet most people, especially Stefan would dismiss that idea. People like him try very hard to reify these "meta" concepts into a single idea. And, without much "meta" thinking you are easily lead to believe that everything could be a function (FP), a resource (REST), a process (pi-calculus), a class (OO), ... Heck, even in SQL a relationship is reified as an attribute of a type.

I would actually argue that all of Stefan's stories would have had a better outcome if people had drawn a metamodel of their approach.