Astronomical instruments are often fascinating cocktails combining science, engineering, economics and politics. When well designed and managed, they produce splendid science, but if handled badly, great ideas very often turn into fiascoes. This talk begins with an overview of the principles and techniques of good project design and management and continues with discussions of the history of several large instrumentation projects, both well and badly managed. These provide interesting examples of what to do (and often more importantly what not to do!) to produce state of the art, cost-effective, scientifically useful instruments on time and within budget.