The International Gemini Project Office and the Canadian Gemini Project Office hosted a meeting in Parksville, BC on Vancouver Island in July to review the instrument program so as to determine what can be done to make instruments less costly and to ensure that they are delivered on schedule.
The first Gemini instrument is about a year late in delivery, and others are suffering varying degrees of cost and schedule overruns. Instrument programs at Keck were reviewed, and for the most part, they are experiencing similar problems. A common theme among failed instruments or those in trouble was that they were not properly managed as large projects by someone with project management experience. There was a general consensus that all too often, either project management is ignored as a frivolous expense, or "wallpaper" management is used, in which a schedule and budget are worked out at the beginning, thumbtacked to the wall, then ignored. What is needed are careful planning at the beginning of the project, diligent oversight throughout the project to determine its status, the willingness to admit that a problem exists, and the fortitude to take needed corrective action that sometimes can be painful.
Jim Oschmann (Gemini Project Manager), gave a short tutorial on the concept of designing to cost. This technique assigns to each engineer three things: technical specifications, a budget, and a schedule. To have that part of the project considered to be successful, all three goals must be met. All too often, engineers stop investigating alternatives when one or two solutions to a technical problem are found. Frequently, the first solution takes too long to implement or drives up the cost, when a simpler approach could save money and time. Rather than blindly trying to meet a specification, which is sometimes arbitrary, engineers should be willing to admit they cannot meet the technical spec with the assigned budget, but could meet 90% of the spec within the budget and schedule. Good project management includes continuing systems engineering and instrument scientist input to determine when to relax a specification to save time and money without adversely affecting the science. This concept is new to most astronomical instrumentation teams, but is increasingly important in the era of large, expensive instruments.
The consensus was that the meeting was helpful, and that similar meetings should be held in the future not only among managers and executives, but also at the technical level to help instrument teams identify good solutions to problems already solved elsewhere.
Mark Trueblood