When adding reference materials to the log, it would be nice if Library of Congress MARC (Machine Readable Cataloging Records) bibliographic or MODS (Metadata Object Description Schema) data could be accessed to automatically flow in a definitive descriptions of the reference item.
This would reduce redundant keying and typos. If a consistent source reference began to evolve, AI routines could be applied to predict which materials and repositories are likely to yeild specific genealogical data.
References:
http://www.loc.gov/cds/mds.html
http://lcweb2.loc.gov/cocoon/minerva/html/minerva-metadata.html