The author is declaring one approach to REST as the "one true way." Actually, I think the author is correct, ie pointing to a HATEOAS style REST approach. But the REST standard is interpreted. The article calls all of the RESTful APIs that are fragile in one way or another "so-called-REST." I kind of agree with that assessment of quality, but at the same time, they aren't "so-called-REST." They are actual REST. We can all just wish those were "so-called-REST" and that there was some additional mechanism to enforce "real" REST. As it stands, accusations of "that's not really REST" amount to being "No True Scotsman" arguments.