Take this quote: “A REST API should be entered with no prior knowledge beyond the initial URI (bookmark) and set of standardized media types that are appropriate for the intended audience (i.e., expected to be understood by any client that might use the API). From that point on, all application state transitions must be driven by client selection of server-provided choices that are present in the received representations or implied by the user’s manipulation of those representations.”

If the client application only understands media types and isn’t supposed to know anything about the interrelationships of the data or possible actions on it, and there is no user that could select from the choices provided by the server, then it’s not clear how the client can do anything purposeful.

Surely, an automated client, or rather its developer, needs a model (a schema) of what is possible to do with the API. Roy Fieldings doesn’t address that aspect at all. At best, his REST API would provide a way for the client to map its model to the actual server calls to make, based on configuration information provided by the server as “hypertext”. But the point of such an indirection is unclear, because the configuration information itself would have to follow a schema known and understood by the client, so again wouldn’t be RESTful in Roy Fielding’s sense.

People are trying to fill in the blanks of what Roy Fielding might have meant, but in the end it just doesn’t make a lot of sense for what REST APIs are used in practice.