To implement the new OPDS selection spec (https://github.com/io7m/opds-acquisition-spec), we need a single central location that can answer various questions such as "Should I display this OPDS feed entry?" and "Which acquisition should I pick?". Currently, we answer these questions in the `BookFormats` class and the `BookAcquisitionSelection` class. These classes both have deficiencies; `BookFormats` doesn't know about optional services such as Adobe DRM, and `BookAcquisitionSelection` implements an inferior algorithm that doesn't give ideal results and also doesn't calculate acquisition paths.