[PaulHoule]

That’s a backwards way to think about it.

JSON-LD lets you paint some extra structure onto a JSON document by adding a little metadata. That is, add some semantic smarts to existing JSON document.

What people miss is that the semantic web was never an attempt to fit everybody into a straightjacket but rather an attempt to mash together all the data in the world into a huge Katamari ball.

[zcw100]

That so beautifully captures the attitude of the semantic web community, "What people miss...". Didn't miss anything. It's this attitude, that the problem isn't with what's been done, it's that other people have failed to recognize how brilliant it all is.

[PaulHoule]

It hasn't been communicated well. The work hasn't been planned well. SHACL should have been introduced before OWL. RDFS doesn't really work as a data integration language because you can't write a rule like

   ?x :lengthInInches ?t -> ?x :lengthInCentimeters ?t*2.54

Since RDF isn't fit for the purpose it was designed for, people get really confused about it.

The strange thing is that semweb work has been massively overfunded in zones that face severe multilingual problems (Europe), but it seems to be almost banned in academia in the United States. (except for a few enclaves I can enumerate on the fingers of a few hands.) I know quite a few Africans who are semweb believers because they are looking at markets fragmented by languages, but they don't have the overfunding that Europeans have. (Funny enough I am working on a standards doc for ISO 20022 which is badly wanted by Chinese authorities who are interested in semweb tech because they want to feel included in financial messaging despite language barriers.)