It appears that Martin Fowler agrees with a long-held belief of mine that, despite industry analysts reports, software projects are almost never truly late or over budget: if the team is working effectively, then projects take about as long as they are supposed to, and cost about what they should. It's unrealistic expections that are to blame. As Fowler says, "Rather than saying that a project is failed because it is late, or has cost overruns - I would argue that it's the estimate that failed."

How many projects have you been on, where the development team comes up with an initial estimate, and the business stakeholders come back with "no, try again", until the estimate fits with the marketing (or other artificial) milestone?