The Richardson 'REST API maturity model' describes how people commonly exchange HTTP payloads, how those systems can be augmented to get closer and closer to the ideas behind REST. Most people are perfectly happy with staying at levels 0 through 2, because HATEOAS is hard in the sense that you actually have to think about stuff like link relations, hypermedia mediatypes, and the semantic web. This is something most people on the 'normal web' don't even do properly, even though the 'normal web' is quite obviously hypermedia, and quite obviously an embodiment of REST. But despite its lack of deployment, Level 3 describes an API that has truly and unambiguously earned the right to call itself a 'REST API'.
In REST, these 'link relations' are extremely important, because they represent well-specified ways of relating one resource to another. They are as critical to REST as mediatypes (MIME types) are. In the graph theory sense, link relations are the label on an edge, URIs are the nodes, and mediatypes are formats that can be used as concrete representations of the nodes. The IANA maintains a Link Relations registry [1].
But since the 'normal web' is used by humans, we don't need link relations to traverse a website; we can instead use context clues, familiarity, or other techniques to achieve what we want on a website; the entire discipline dedicated to coaxing humans into interacting with websites is called 'UI/UX'.
Meanwhile, Kinlan envisions... well, to be frank I'm not entirely sure what Kinlan envisions because after mentioning 'headless web' he writes a lot about service workers, notification hooks, then points to ways we already use the web in ways that aren't out of our own volition, like in WebViews and IFrames (which he writes about in detail here [2]). I think his point is largely that every app maker wants to control the viewport around which content is surfaced, so they can add their own control and bring their own community, auth, ads [3], etc, and we're finally at a point to where the web platform has matured enough to make these websites or 'web includes' access capabilities offered by the device or the device's platform runtime, such as notifications, sensor data, and the like.
In a sense, they're not contradictory at all. The Richardson model clearly strives towards the semantic web, while Kinlan's writings predict that the semantic web has essentially failed, leaving people to write custom scripts and parsers and crawlers and bots to mash web content, some of it 'Web Platform-enabled', into other places.
>I think his point is largely that every app maker wants to control the viewport around which content is surfaced, so they can add their own control and bring their own community, auth, ads [3], etc, and we're finally at a point to where the web platform has matured enough to make these websites or 'web includes' access capabilities offered by the device or the device's platform runtime, such as notifications, sensor data, and the like.
I wish I had actually said that in the post. It's quite succinct.
In REST, these 'link relations' are extremely important, because they represent well-specified ways of relating one resource to another. They are as critical to REST as mediatypes (MIME types) are. In the graph theory sense, link relations are the label on an edge, URIs are the nodes, and mediatypes are formats that can be used as concrete representations of the nodes. The IANA maintains a Link Relations registry [1].
But since the 'normal web' is used by humans, we don't need link relations to traverse a website; we can instead use context clues, familiarity, or other techniques to achieve what we want on a website; the entire discipline dedicated to coaxing humans into interacting with websites is called 'UI/UX'.
Meanwhile, Kinlan envisions... well, to be frank I'm not entirely sure what Kinlan envisions because after mentioning 'headless web' he writes a lot about service workers, notification hooks, then points to ways we already use the web in ways that aren't out of our own volition, like in WebViews and IFrames (which he writes about in detail here [2]). I think his point is largely that every app maker wants to control the viewport around which content is surfaced, so they can add their own control and bring their own community, auth, ads [3], etc, and we're finally at a point to where the web platform has matured enough to make these websites or 'web includes' access capabilities offered by the device or the device's platform runtime, such as notifications, sensor data, and the like.
In a sense, they're not contradictory at all. The Richardson model clearly strives towards the semantic web, while Kinlan's writings predict that the semantic web has essentially failed, leaving people to write custom scripts and parsers and crawlers and bots to mash web content, some of it 'Web Platform-enabled', into other places.
[1] http://www.iana.org/assignments/link-relations/link-relation...
[2] https://paul.kinlan.me/slice-the-web
[3] https://paul.kinlan.me/rise-of-the-meta-platforms/