> If there was a planning process that was consistently reliable for new, complex software I think people would use it. And I think the adoption of agile method

> If there was a planning process that was consistently reliable for new, complex software I think people would use it. And I think the adoption of agile methodologies in software development is an example of people trying to implement an improved planning process. There are such processes. But they require understanding of the problem, which takes time. I am not sure "agile" is really attempt at an improved planning process, it's at best not doing planning and just marching onto the problem, seeing what happens at regular intervals in the hope of better understanding it, and at worst, things like "planning poker" where you put a bunch of people into a room and let everybody make a (possibly educated but usually not very much) guess from the top of their heads.