J.L. Eveleens and C. Verhoef, IEEE Software, 27 (1), Jan-Feb 2010
This is an article that analyses the frequently cited Standish reports that state how many SE projects have failed and generally present software engineering to be in a bad state. The paper shows that Standish’s method is completely unreliable and the Standish figures simply have no credence at all. I’ve always maintained we’re not that bad at software engineering.