The original design of the Bibliotheca API client had some of the code in core (bibliographic stuff, which could be used by both the metadata wrangler and the circulation manager) and some in circulation (the part of the API that acts on behalf of a patron). But it turned out that the metadata wrangler can't actually do much with one library's Bibliotheca client credentials, so we took the metadata wrangler code out.

By merging the core code into the circulation code, we could reduce the size of core and put all the Bibliotheca code in one place, making it easier to understand.