I’m wondering if Hydra could be used to describe an API that a solid client could use to discover what ai and nlp capabilities a solid server may have. I’m not that familiar with Hydra so I need to find out more about it. If anybody has info or ideas on the feasibility of that, they would be much appreciated.
I am participating in the Hydra Community Group and I think Hydra could be a good fit for a solid server. It could e.g. be used to discover where to find or store specific kinds of data, in a more flexible way than type index. But it is an area that needs to be explored further
Personal opinion: in-band hypermedia is great for doing this. We could use it on all graph-based content types, or with profile-based content negotiation.
However, Tim is in favor of leaving documents as-is (which we could still support with profile-based conneg).