Ah of-course. I forgot the convergence requirements for pagerank. Indeed, the fact that cites are a DAG really does kind of screw with pagerank.

Note that my supposed idea was to use 'pagerank' to replace reference count as the value of a work. Not to replace peer review.