I have implemented many APIs in this style, both read and write. Depending on the design of the storage layer, the server may or may not have to parse ids out of URLs. The client never has to; for the client, the URL is the only Id.