code4lib thread on ranking
Colleen Whitney posted to the code4lib mailing list yesterday: Question re: Ranking and FRBR. She asks, “In a retrieval environment that presents ranked results (ranked by record content, optionally boosted by circulation and/or holdings), how could/should FRBR-like record groupings be factored into ranking?” The whole thread is in the code4lib archives. Thom Hickey of OCLC says they add up all the holdings of all the manifestations of the given work. OCLC’s enormous WorldCat catalogue lets them to do things regular libraries can’t, and there’s more discussion about relevance rankings and, unavoidably, Google.