Documentation for extraction from the many sources is weak, and embedded in comments and code in each extraction. These comments are often incomplete and all of them lack discussion of Daniel's rationale for each element's/attribute's treatment.

This is a long term issue, but we can shoot to have it resolved by December 2015