We have been trying to deal with the use case where the user or the questionnaire modifies either the model name or the simulation name in a CIM document. This has actually happened a lot. Since the feed works by creating new versions that fix the offending strings, not removing older versions, our system the way it currently does harvesting on total triple store restarts will grab the old as well as the new instances.

We need to change this behavior so that it tracks just the lastest version of a document and grabs only that one rather than harvesting them all.