7 Steps to Failure with Your Next Software Project
For a little while now I've been trying to decide how Rational (now part of IBM) could be so successful at pushing their monolithic, process-laden software development methodology, especially in the face of increasing evidence that such methodologies aren't particularly useful. The answer to this came from Jeff Sutherland of SCRUM fame. He recently posted a link to a PDF titled How to Fail with the Rational Unified Process: Seven Steps to Pain and Suffering. Basically, my understanding of the RUP was incorrect and that most implementations of RUP are faulty.
At only 14 pages, it's a relatively short article, but I'm going to summarize it here anyway. If your software development methodology is modeled after the now famous sequential, "waterfall" lifecycle, then you're likely to fail. Failure in this case means missing deadlines or exceeding budgets. The article goes on to explain that the most compelling reason for software project failure is the assumption that most requirements can be defined in the first project phase.
This article is definitely worth reading. What's even more amusing is that you can probably replace the RUP part of the title with your company's name and the article will still ring true. So now I want to know is where are all the companies that use agile processes? Are they so successful that they don't want their competitors to know?