The Message Key Format Is Not Formally Defined

[ahassany]

The draft states the Message Key is composed of hostname, datastore name, and absolute-schema-nodeid plus optional list key, but never gives a formal syntax for how these components are concatenated. There is no definition of:

• The delimiter between components (the examples imply / but this is never stated).
• Whether the YANG module prefix is included or excluded from the absolute-schema-nodeid portion of the key.
• Whether the datastore name is the NMDA identity string (e.g., urn:ietf:params:netconf:datastore:operational) or a short form.

Without a normative format definition, two independent implementors will produce incompatible Message Keys for the same data, which completely defeats the purpose of deterministic hash partitioning — consumers computing expected partitions would get different results from producers.