On Friday 13 March 2015 05:47:38 Markus Jung wrote: > Further, should the metadata encoded if a resource is retrieved? If yes, a > simple key/value encoding of attributes is not possible anymore and the > payload might become quite big. Another way would be to use the resource > type definitions based on JSON schema as proposed in the draft and to > provide a separate interface to query the metadata.
Hello Markus The spec does not discuss how the metadata is transmitted. It only describes the schema for what the metadata can be. It's up to the implementation to transmit that payload in the form most convenient to it. Since we'll have a structured key-value payload encoding anyway, transmitting the metadata should be as easy as adding an extra attribute to our payload which contains itself a dictionary of key-value pairs. If we're transmitting JSON, then the dictionary is a JSON object; if we're transmitting CBOR, the dictionary is a CBOR object. -- Thiago Macieira - thiago.macieira (AT) intel.com Software Architect - Intel Open Source Technology Center
