TODO

* Add documentation for new features:
  - Key generation
  - PDF file storage
  - HTML output
  - Key "private"
  - print command
  - author file

* Improve handling of duplicated keys.  
Currently, if one imports a database and a particular key is present
in both, one should not be allowed to keep both publications.

* Improve handling of allowed duplicates
Currently papers can be marked with "allow duplicates". This will skip
the duplicate test completely. Instead a list of ids to the duplicates
should be maintained and the paper should be tested against all other
papers not explicitly marked as duplicates.

* Encodings
Special characters in .pub are now assumed to be encoded in latex
style. Converting is now handled in the export script.  
- We should consider if latex is the best encoding.  
- Centralize the encoding conversion and use some standard python
  modules.