I briefly worked in a public library and the problem they had was that there's no practical way to maintain sufficient numbers and depth of subject-matter exper

I briefly worked in a public library and the problem they had was that there's no practical way to maintain sufficient numbers and depth of subject-matter experts who can answer the subjective assessment such as you list. Ideally one or two shallow-SMEs in each field would be maintained over a region, but that will never happen. So you end-up with librarians making retention decisions based on what publications tell them are the "most useful" books in a topic area and trying to correlate that with check-out frequency. But those topic areas are coarse-grained like 'history of aviation' or 'web programming'. It's a crap-shoot and unfortunately a lot of good books are weeded instead of the dross. My cheap solution: non-fiction is retained until a updated edition is purchased ( or forever if that doesn't occur ), non-curriculum fiction is weeded after 50 loans regardless of replacement. Curriculum fiction such as Shakespeare probably shouldn't be weeded, just replaced.