specifically how the client needs to deal with converting returned id values into full urls.

integrated servers

When integrated services a returned id value (e.g. would be a relative id on that server); and that relative id would be understood by the other integrated services.

independent servers

However when services are on independent servers (e.g. Patient management on one server, Directory on another, DocumentRef on another, and QEDm results on another); then a returned id; would need to be converted into a full url when hitting another server.

consistency

Note that a client should always work to convert id values into full urls.

Note that a server should always work if it only returns full urls.

Not sure which is better, or if there is a way to constrain this.

questions

This is my theory.

Iryna Roy asked on linked in