First usable python version:
      o Port what's there to python
      o Automate generation; delay implementing search?
      X I need a CGI for submitting changes?  Probably not.
      o Make an HTML file with many anchors to hold the BiBTeX stuff.
      X Cached parsed BibTeX, if we have any reason to do so.
      o Integrate two files: search for duplicates, return everything else.
          o Matching
              o Very dumb match: is everything the same? (For all fields)
              o Dumb match: Can we find something that has the same title
                and the same year and the same authors?  If so, and some
                other fields differ, dump out both version in third file.
              o Less dumb match: As above, but with some fuzziness in
                authors and titles.
            o CLI and real code
          o Cleaning
      o CLI
      o "Important" note.
      o Remarks

Later:
        - Know about @book
	- Maybe uncrossref in tmp.bib
	- Maybe pull important papers to the start of their sections?
	- Clean \{}~ when going from note to url; add \{}~ when making
	  note from url.
	- Look for urls in wherepublished.
	- Forgive newlines in wherepublished, note.
	- "Systems, Zero Knowledge"?