Developing large software systems can be done with a variety of approaches. Some of these emphasize fully analyzing all of the requirements up front; others emphasize building a little, testing a little, learning what's needed and how to build and maintain it as you go along. That's just one of many sets of variables a software project manager must make choices about. How would you, as a software project leader, want to do this? What other factors would you consider in choosing your approach? How would the nature of the desired system you're trying to build influence your choices