ORCID currently supports JSON-LD via content negotiation. Meaning that if you have an ORCID iD you can pull the schema.org metadata straight from the identifier URI. Our content negotiation docs are at https://github.com/ORCID/ORCID-Source/blob/master/CONTENT_NEGOTIATION.md
We encountered some small issues implementing schema.org. Mainly that a work can have an author, but an author cannot have a work, which from an ORCID perspective is somewhat back-to-front. We did try to reach out to schema.org but didn’t get far and in the end overcome them using a JSON-LD ‘trick’ of inverse relationships. It’s not ideal, but it works.
We are considering embedding schema.org into our public landing pages. But we do not have a good idea of how this will be used, or the use cases for using it as it’s read anonymously. You can see the card on our roadmap here: https://trello.com/c/iIcOLtAE/51-hidden-trial-very-simple-embedded-schemaorg-on-public-pages-5
I’d suggest that our current record format is pretty close to best practice for representing a scholar, and we did work with linked data folks when designing it, but I’d be happy to stand corrected. Personally, I’d rather not diverge/extend the regular schema without some strong use cases to back it up.
I know Datacite support JSON-LD and we worked with them when we put our own in place. I’m sure they will have something more to say on the matter.
ORCID Product Director.