"Server" as used in the spec vs. "Server" as used in JSON-RPC 2.0 - Request for clarification

[hesreallyhim]

Pre-submission Checklist

• I have checked that this question would not be more appropriate as an issue in a specific repository
• I have searched existing discussions and documentation for answers

Question Category

• Protocol Specification
• SDK Usage
• Server Implementation
• General Implementation
• Documentation
• Other

Your Question

One of the clearest statements declared in the MCP Specification is:

The protocol uses JSON-RPC 2.0 messages to establish communication

JSON-RPC 2.0 spec states:

The Client is defined as the origin of Request objects and the handler of Response objects.
The Server is defined as the origin of Response objects and the handler of Request objects.

Nevertheless the Architecture flow diagram clearly shows "Server Requests" and "Client Responses."

And such message types are explicitly declared in the schema. Similar issues arise with respect to Notifications, which are classified as Request objects ("A Notification is a Request object"), and therefore originate from the client. Nevertheless, MCP notifications may be sent from either party.

My question is how the reader should understand this relationship, which on its face is a plain contradiction in terms:
(a) is it a "violation"/"loose interpretation" of JSON-RPC 2.0 (perhaps just a deliberate choice);
(b) Does it fall under the following umbrella, which is suggested in the JSON-RPC 2.0 spec:

One implementation of this specification could easily fill both of those roles, even at the same time, to other different clients or the same client. This specification does not address that layer of complexity.

I find this statement a little hard to parse, but it suggests that within a given architecture, entities may exchange roles.

(c) other;

For example, in the Base Protocol > Overview, it is stated:

Requests are sent from the client to the server or vice versa

If we "unpack" the "or vice versa," this reads:

Requests are sent from the client to the server or from the server to the client.

In which case the second occurrences of the terms "server" and "client" must be interpreted from an "MCP POV", and that from the POV of JSON-RPC, roughly means something like:

Requests are sent from the client to the server or from the server (adopting the messaging role of Client) to the client (adopting the messaging role of Server).

I would appreciate some guidance in interpreting this clause. Although it may seem ridiculously pedantic or petty, it is intended to lay a foundation for my interpretation of other parts of the specification. I would also be satisfied with a response along the lines of "MCP closely adheres to JSON-RPC 2.0, with minor deviations."

Thanks!

[hesreallyhim]

ok I've researched this question, and I think I am comfortable with interpretation (b) listed above. I'll leave this open briefly and then close it.

[hesreallyhim]

The terms "server" and "client" as used in the JSON-RPC 2.0 spec do not need to map consistently to the same entities in the architecture of an application throughout the entire application lifecycle. An interesting discussion can be found here:
https://groups.google.com/g/json-rpc/c/2WyxWKSiz8o?pli=1