If we define a "failed project" as one that someone has worked on and then shelved when it didn't gain traction, then I think that perhaps a more common scenario is that of the Zombie Project; after spending so much work on something you can't bear to shut it down, so you leave it ticking over, not updating anything and not attracting any new users. Until you pronounce the project officially "dead" it's not a failure...until then it'll just keep limping along. I think recognising when a project has gone "zombie" and having the balls to kill it off is massively important, both for time, sanity and closure's sake.