Separating a large, complicated task into multiple orthogonal pieces, and assigning each of these pieces to its own component, has been a fundamental part of software engineering for many decades.
Separation of concerns is what OO does. Separation of concerns is what UNIX does. Separation of concerns is what TCP/IP does. Separation of concerns is why global variables are considered "bad." Separation of concerns is not a new idea.
This jargon-filled, somewhat poorly worded article is merely repeating a fundamental principle that should be obvious to anyone who's ever written any program larger than a couple hundred lines.
What is meant by "[the] client is free to have full coherence of server's public domain including all its public API, domain objects and schemata"?
"CSDS regards hypermedia an important aspect of REST; it is a semantic web of interconnected resources. Client will have full coherence of the axes of such relationships and can effectively use to navigate the semantic web - since it is part of the public domain. But for it to become the engine of the application is server dominance." These sentences are a long string of big words, but I'm not sure that there's actually anything meaningful in them.
Reading this article makes me really wish I had enough karma to downvote.
Decision to separate domains of client and server is a big one and has its implications. CSDS is common sense as is HATEOAS but they are not consistent. ROCA is also common sense but does not agree with CSDS.