Name clash! There's already Tie::Hash::Rank (written by me) so having Tie::Hash::Ranked do something different will confuse the hell out of people. Why not just implement the extra methods in Tie::Hash::Sorted? I'm not swayed by the bloat argument in this case.
DrHyde,
Thank you for letting me know about Tie::Hash::Rank. I am seriously leaning towards a new namespace not only because of code bloat, but also because the methods I am talking about will dramatically change the way the hash behaves. If I change to a tree structure under the hood, I will need to make a new module or break backwards compatability. That is poor OO design - but as I said, it was my first real OO project.
So when and if I am ready, I will be sure to be considerate of pre-existing name spaces.