Reading your post it appears that a software model is required for Discovery Models to be normalized or am I misreading this? I am trying to understand if a software model is REQUIRED for the discovery model to be normalized correctly as well as receive content library EOL data.

From what I am seeing is that discovery models are being created and normalized by the content services (looking at product map and package map tables) to a publisher/product/version from the raw discovered data. I understand that Discovery Maps help with defining the conditions required for discovery models to link to software models but have also found that software models will link to discovery models even without a discovery map as long as the conditions on the software model match that of the discovery model (same publisher,product, version or edition).

Is this accurate?