Multi-User-Domain / mud-jena

0 stars 0 forks source link

POST Scene ... Action options #33

Closed calummackervoy closed 3 years ago

calummackervoy commented 3 years ago

We have an existing endpoint for POSTing a Scene, a graph of objects, and receiving back a graph of content/perspective for that scene

The proposal is to do something like this for an Action Server so that we can deduce from a scene the available actions, their endpoints and expected parameters. This could be quite generic, e.g. I might post it a dialogue graph and expect to receive options for interactions

I think that if the user POSTs an empty graph, or if they make a GET request then we should return all endpoints available on the server, as a means of GETing the configuration (a server configuration then only needs to explain where to find this endpoint, and can remain static - e.g. if I define my servers' actions dynamically or in a database rather than in code)

calummackervoy commented 3 years ago

I think that for now we can just return where the action is, and rely on the client implicitly knowing the parameters.. "ah I know how to request this one"

It's a step forward. For communicating the parameters, follow #44