List of things to do in duff
============================

Functionality
* Change to getopt_long and add long options?
  - which would require adding gnulib
* Detect duplicate file arguments?
  - is that even possible to do in a sane fashion?
* Detect duplicate directory subtrees
  - in the upcoming treeduff, not plain duff

Documentation
* Add ALGORITHM section (or similar) to manpage
  - although first we should have an algorithm worthy of the name
* Add more examples to manpage

Optimisation
* Make sampling digest use the first n bytes (nicer to cache)
* Make directory cache less stupid and slow
* Change linked lists into hash tables or trees
* Do "smart stuff" to "make it go faster"
  - people use duff on millions of files; optimise for that

