Currently all subjects of the RDF returned by mementos reference the original resource rather than the memento. For example:
curl http://localhost:8080/rest/test/fcr:versions/20180101120000 will return RDF with a subject of <http://localhost:8080/rest/test> .
Assuming that this behavior conforms to the Memento Spec, consumers of Fedora events (such as the camel toolbox) may not be distinguishing new Memento resources from new non-memento resources. If this is the case, those downstream processes may break. For example, a triple store indexer will add the memento-ized triples with the original resource subject. thus corrupting the representation of the original resource in the triple store.
1. Are we implementing Memento correctly by using the original resource in the URI?
2. What do we need to do, if anything, to make sure that downstream processes can distinguish memento related events from other non-memento resources?
As a note, clients can examine the headers of a resource in order to detail whether or not it is a Memento and can process or ignore it accordingly.