[sigsergv]

Semantic web is not that hard to understand. But facebook/google/etc will NEVER adopt it. They are not interested in opening (meta)data in highly precise and machine readable form to 3rd parties. So the only adopters are geeks/scientists.

[abusque]

I can see where you're coming from, and your point of view is certainly not baseless, but I just thought I'd point out that at least Google has pushed for the development and adoption of linked data formats like JSON-LD [0] and standardized vocabularies like schema.org [1]. They make use of it for "knowledge graph" [2] features, as well as in Gmail for what they call "actions and highlights" [3] (things like displaying flight reservation details, for instance).

[0] https://json-ld.org/

[1] https://schema.org/

[2] https://developers.google.com/knowledge-graph/

[3] https://developers.google.com/gmail/markup/getting-started

[OkGoDoIt]

Yes, Google would love for you to mark up your data so that they can better consume it. But good luck trying to get Google to make any of their data more interoperable. Google Plus, YouTube, Google Photos... they do have somewhat limited APIs, but they are not federated and standardized. Semantic web in, limited proprietary access out. Walled gardens are a business tactic, no semantic web technology can change that.

[smadge]

I think you are correct. Semantic web is a decent technology. It has some rough edges, but it has solved the technical aspect of the data interoperability problem. The only barrier is a social/political/business one: privatizing and monetizing user data is the business model of most of Silicon Valley. I always say that a federated protocol like email would never be adopted today, the business incentives just do not exist.

[pmlnr]

Don't trust google on metadata recommendations, they change it way too often: https://aaronparecki.com/2016/12/17/8/owning-my-reviews

[meh2frdf]

The W3C docs are very developer unfriendly additionaly there is a significant issue with OWA vs the default mindset of CWA. Also OWL and RFFS not being the type of schema a dev is used to and until very recently no option to actually verify and control the shape of the data (there is SHACL now). Also no sensible examples of using the stack, using FOAF to publish open data isn’t the job for most devs! Finally the rubbish tooling, clunky DBs and only viable option is Java makes the whole thing a quagmire... however it has tons of value if you can get past all that friction, Enterprise Linked Data is an order of magnitude better than alternatives.