I don't think it's depressing. I can use emacs (others would give different examples) for free and I can use it in my daily work quite efficiently. So given that reality, I have a hard time paying for an editor, even if it is a key tool of my trade.
What I was trying to say is that every tool should be evaluated using 'productivity gain vs cost' framework. It may well be that for some people productivity gains of TextMate vs emacs/vim/other editor are small and therefore paying any price for TextMate is unreasonable, since the alternatives are free.
But in a case where TextMate does yield to substantial productivity gain, "$60 is too much" is a wrong mental model.
For java development? I don't have a recommendation really. I've tried Intelli-j and I think it is def. better but at least for my workflow at the paying job, eclipse's notion of project works out much better.
Few people are saying that $60 is too much.
Many people are saying, I paid $60 w/ the promise of an upgrade to X. I expect that commitment to be honored.