[sumnole]

Pedantics aside, not much reasoning against quality. Perhaps I've lucked out, but I've worked in many sectors and do not at all agree with sentiment here about DOD software quality. There is significant formal investment/research in DOD to improve operations, including taking the best of practices in commercial. In my experience, the worst of software is written by teams with little experience improvising under Agile and taking on tech debt with no time/resources to get things done the right way.

[Jtsummers]

Can you point to a successful Waterfall project? A multi-million or billion dollar, 3+ year software development effort where a team figured out all the requirements correctly before writing a single line of code. Where they wrote every line of code correctly before testing it. And where testing was so spectacularly successful that they didn't have to go back and renegotiate the project requirements or dates to get extensions or reduced project scope.

If you can do this, then I might believe you about Waterfall being the best approach out there.

Right now your counter example is "teams with little experience" which is not much of an argument. Teams with little experience fail all the time, because they are inexperienced. Give them a $100 million Waterfall project to plan and execute over 3+ years and their failure would be even more spectacular.

[ghc]

I just attended a DoD "Scrum of Scrums" meeting.

> In my experience, the worst of software is written by teams with little experience improvising under Agile and taking on tech debt with no time/resources to get things done the right way.

Sounds like every DoD software project I've worked on for the past 5 years.