This is just a dumping ground for ideas and vague plans.

- Some trackers (bytemonsoon.com is notable. (warning: mature content)), don't
  support scrape.  Can we get some sort of useful information from them using
  the announce interface?

- POD documentation

	- pod2usage?

- If we end up with lots of marginally compatible announce/scrape servers, we
  may need to guess at multiple scrape URLs and try them all.  Rewrite to
  support?  Wait until it comes up?

- Clean up code

	- Ensure errors are passed around, not just dumped on STDERR.

- BitTorrent::BDecode 

	- Massive cleanup?  It's scary code.

	- Currently spews errors to STDERR.  Send somewhere useful.

	- Refactor into a generic BEncode library supporting both encoding and
	  decoding
	  	
		- Submit to CPAN?

- Can other parts of code be refactored into BitTorrent::* tools?

	- Provide tools necessary to write a BitTorrent client in Perl?

- Reimplement the lot into Python? (I love Perl and all, but given that most
  users will already be using the Python based BitTorrent client...)

