I'm with you about holding management accountable and avoiding crap management, but I think it is orthogonal to sprints. Sprints might be a solution for some teams, but I have seen them fail far more than succeed. Actually ... I've never seen them succeed.
I've seen them succeed. I think the differentiating factors were 1. everyone worked on lots of different bits of the code; there weren't bits that were "owned" by someone, and 2. the end goal was pretty obvious.
