In the 1980s, Don Swanson of the University of Chicago (https://en.wikipedia.org/wiki/Don_R._Swanson) was a pioneer of advancing science through literature review -- going back over the published literature to see what had been missed.  Thought of him when I saw this:  

Enter text and data mining and artificial intelligence.  I particularly like the part about looking at the literature pre-2009 to see if the method would have predicted discoveries post-2009 and it seems to have done so.  This is relevant for the need to ensure reasonable TDM terms in the licenses libraries sign with publishers.  We need their content for more than traditional current *or* historic awareness.

Jim O'Donnell

ASU