So, I think this is an _excellent_ idea for a tool for Debian.
Some other ideas for data sources that might be worth thinking about:
* BTS. How many bugs, and maybe ratio of bugs to installations,
etc.
* Dependencies. How many other packages depend on this one? Gives
a rough estimate of popularity among developers as well as
users.
~ESP
--
Evan Prodromou
evan@debian.org