This document failed to convince me that new systems are not "correct".
I think the author just doesn't like that all languages cannot specify what calls to external services look like in a similar way to the old COM IDLs (likely how he/she first learned about IPC).
WSDLs take that role when using many webservices and other documentation fills that role fine. It is not like IDLs added good compile time checks (at least not all the time anyway), I had plenty of fun expecting the compiler to stop me from jamming strings into ints and such. Now tools and languages are better everywhere, for example look at how external calls work in languages like Rust and the future only looks better.
Thinking about it more the author did want RPC to die. Perhaps I only disagree with the implication that RESTful services are not correct. For the first half of the article that seemed the point, then a spontaneous bought of happiness at the end that events are happening the way they are. Now I think it was just a fluff piece with too much jargon.
I think the author just doesn't like that all languages cannot specify what calls to external services look like in a similar way to the old COM IDLs (likely how he/she first learned about IPC).
WSDLs take that role when using many webservices and other documentation fills that role fine. It is not like IDLs added good compile time checks (at least not all the time anyway), I had plenty of fun expecting the compiler to stop me from jamming strings into ints and such. Now tools and languages are better everywhere, for example look at how external calls work in languages like Rust and the future only looks better.