If you were to choose the scientific method for diagnosing network problems, what trade-offs (i.e. compromises) would you consider? (HINT: Think of evaluating theoretical models, simulations and experimentation in terms of cost, complexity and accuracy).