I think the reason these kinds of things aren't done very well for a long time is because of existing implementations. Would-be creators do some research and decide the market is taken, not realizing how much room there is for a great implementation.

Geni, Gmail, Flickr, Facebook, and the new Evite competitors come to mind as examples of people ignoring that affect.

Do you mean the bit about a user providing others' email addresses when putting them into a tree? Only, family tree making sites, including ones that spot possible "same people" between trees and tell both parties, have been around for years.

definitely - ideas are cheap and plentiful, no matter how cool they are.

Just remember that implementation/execution (no matter how 'simple' an idea is) is a bitch

... and initial ideas aren't often that special - they often change and evolve.

I'd get more excited about finding the people who u can trust, work well with, and who also have the craziness to either jump off the corporate/college safety net or to work their nights and weekends towards completing something worthwhile