It covers a different use case. JSON, MessagePack and CBOR fall into the "schameless" format which mandate a common but useful enough data model (CBOR is novel in that this data model can be somehow extensible too). Cap'n Proto and Protobuf fall into the "schametic" or "schemaful" format where you always need the correct schema to encode and decode the thing, but is possibly more efficient in encoded size and general performance.
Thanks for clarifying! I thought that Cap'n Proto allows for evolving schemas, but I guess it's true that if each of your messages is completely different, it's not going to benefit you as much perhaps.
It covers a different use case. JSON, MessagePack and CBOR fall into the "schameless" format which mandate a common but useful enough data model (CBOR is novel in that this data model can be somehow extensible too). Cap'n Proto and Protobuf fall into the "schametic" or "schemaful" format where you always need the correct schema to encode and decode the thing, but is possibly more efficient in encoded size and general performance.
Thanks for clarifying! I thought that Cap'n Proto allows for evolving schemas, but I guess it's true that if each of your messages is completely different, it's not going to benefit you as much perhaps.