Results from a survey concerning 400 IT projects:

What were the surprises? > > > * Most projects that had no schedule were successful > > * Requirements are needed for project success, but not necessarily early in the project > > * Projects often continue successfully for some time with unclear requirements > > * The choice of requirements methodology does not matter; UML was "no help" >

Success/Failure Criteria: Some Surprises by Robert L. Glass - developer.*, Developer Dot Star

I find the results interesting. I’m not as surprised by the preceding points as the author (Robert Glass, whose excellent book Software Conflict 2.0 I’m currently reading). I won’t start a rant about UML. I think its best use is for sketching some ideas.

Software engineers should find the first point, projects with no schedule were successful, to be of particular interest. How many times have you heard the results from the Standish Group that says 75% of projects were unsuccessful, meaning that they were late and over budget? Is that a real failure of the project, or do we just suck at estimating?