Why all of this matters with respect to #JsonLD:
JSON-LD is a hugely, massively complex beast of a spec that is incredibly feature rich (c.f., https://hachyderm.io/@hrefna/110894609163744434 ).
That complexity, however, comes at a cost and, unlike say JSON schema or some other options in this space, that cost is borne by the _implementer_. Especially problematic since a lot was punted _to_ JSON-LD: #ActivityPub didn't want to define a metamodel and wanted to lean on #JsonLD for a lot.
So what do we do about it?
1/
Option 1: We can define an #ActivityPub2 specification (or equivalent) that moves wholesale away from #JsonLD and adopts something that pushes the complexity to a different level. Preferably something with a more defined metamodel.
Bluntly: I do not think this is feasible. Not from a technical standpoint, but rather from a _cultural_ one. Too many people too invested in JSON-LD, and too much already built on top of #ActivityPub to do this.
2/
Ok, I am not a technical expert but surely what we have works so why break a model that works and does the job. More complexity = more work for people too.
