What not to solve?

Every now and then I keep mentioning the people who I really enjoyed working with. One name which I kept omitting was my past manager of three years - Trey Marshall. This MBA from Carnegie Millan - a critique of overkilled solutions is one who leaves his impression. The two things which I learned working with him are : what not to solve? and what a running system means? Two not so simple questions as they seem to be. Sometimes as developers, as analysts and as business owners we don't understand the problem statement clearly and usually get carried away on what the system should do? And knowing what not to solve is as difficult of a question as of knowing what should be solved. The second is the value of a running system. A running and a scalable system is better than a utopic one on papers and discussions. The system which runs is the one which sustains. Being practical is one more reason of my admirations for him.