Together, these are interesting as they are responsive to the particular characteristics of the materials in Scholar and are a departure from the general Google approach. It raises a couple of immediate questions: (1) what use Google actually makes of the structured data that is increasingly available to it, and (2) what approaches to ranking 'hidden web' materials it comes up with. And a more general longer term issue: will we see a growing differentiation depending on the materials being handled, as the approach developed for an interconnected world of web pages applies less to other types of material? Whatever happens, we can be sure that we will see innovation.
OCLC is now providing volume of holdings data to the search engines as part of its Open WorldCat data exposure (although it may not yet be used). It is doing this to provide more 'intelligence' as they make ranking decisions. This is a simple count, the number of libraries holding an item: we are not releasing any data about who actually holds an item.
There is a definite family resemblance between the page rank approach (aggregating linking decisions), citation (aggregating citing decisions), and holdings (aggregating collection development decisions).