We need to have some way of generating endpoint documentation automatically if we want this to be useful. HAL and JSON-LD are two ways of presenting data that are quite lightweight. They don't seem to be able, intrinsically, to signal some fields as required or not. Hydra seems much more complicated, but it seems to do what I want it to. Meh.
Maybe writing an HAL generator based on tables + capabilities would make sense in terms of making a program aware of the endpoint? Maybe we could also slap a pretty interface on top of the HAL representation? Maybe get Colin implicated?
------Copied from #1------
For automated data consumption, there needs to be an as-automated-as-possible json-ld mapping based on table, field names.
for reference: http://www.markus-lanthaler.com/hydra/
Then again, there's really nothing at all preventing me from implementing some, none, or all of those.
As an exercise in futility, I don't see why I shouldn't.
I would probably need to have some filename-based sleight of hand if I wanted to have both json-ld AND hydra active, or you know, I could allow a user to pick and choose at runtime.
Dammit now I have to reopen #1.