V. VI. VII. e) None of the interim measures face the fact that Computers become obsolete in ten years. Stretching beyond ten years is possible but increasingly difficult and costly. Adequacy of Acquisition Plans Plans are well thought out. Good advisers were b) Seem to be conservative and careful which is proper. Accuracy of cost estimates a) Cost estimates for hardware are most certainly too b) Cost estimates for software are too low (a very easy A. Lessons to be Learned from other Large Government Computer a) Make sure requirements are well understood and b) Proceed in on orderly fashion c) Structure the software. step-by-step a) Avoid too much avant-garde technology. e) After you do all this, you will still have software errors. f) Buy more hardware if you can simplify software by so g) Software errors are insidious difficult to find h) Fault Tolerant computers are on the horizon perhaps two years away. i) Voting Mechanisms are beginning to be used and seem B. Conclusions drawn from other large systems a) Software costs 70 % of a system acquisition (not large scale production). c) Scftware is almost always the item taking longest time. d) Go slowly at first - carefully pick architecture design it thoroughly, check it by an independent source. Examples of Large Scale Government Acquisitions with which I am familiar (4) World Wide Military Command and Control Systems. d) Integrated Logistic Systems. e) Automated Test Equipment Systems for Aircraft Maintenance. Mr. GLICKMAN. The next witness is Mr. Donald Scantlebury, Director of Accounting and Financial Management Division of the GAO. Mr. Scantlebury, it's a pleasure to have you here. Your complete statement will appear on the record or summarize it as you deem appropriate, and we will appreciate your identifying the people with you at the table today. STATEMENT OF DONALD L. SCANTLEBURY, DIRECTOR, AC- I would like to read my testimony, if that is all right with you. The people accompanying me today are Dr. Carl Palmer on my right, who is Group Director of our Division, the Accounting and Financial Management Division, and Mr. Tony Csicseri, on my left, who is a supervisory evaluator of the staff. Chairman and Members of the Subcommittee: We are pleased to be here today at your hearings on the ral Aviation Administration's Air Traffic Control Computer rnization Program. In response to your request, we have ed to testify in three areas: (1) our analysis of the FedAviation Administration's reply to the Senate report; (2) possibility of direct replacement of outmoded computers; (3) the lessons learned about planning, acquisition, and estimating for large computer systems. At the outset of my testimony, I would like to stress that n the General Accounting Office have made no independent assessment of FAA's Modernization Program or of the en route air traffic control computer system. Our comments are based on the extensive review conducted by the Senate investigations staff, the GAO analysis of FAA's reply to the Senate report, the GAO report on direct replacement of economically obsolescent and outmoded computers in the Federal Government, and lessons learned from our reviews of other large Government computer system acquisitions. In response to reported concerns over the safety and reliability of the equipment being used by FAA for en route air traffic control, the Chairman of the Senate Appropriations Committee authorized and directed the Senate investigations staff to determine if a problem existed and, if so, what corrective action could be taken to improve the situation. Federal Aviation Administration officials had proposed acquiring a new $2.8 billion en route computer system sometime in the late 1980s. The replacement system would be designed to perform in the same manner the functions of the existing computers, be based on the advanced technology at the time, and be 10 times larger so as to later incorporate desired futuristic air traffic control functions. In its report, the Senate investigations staff found serious management and planning deficiencies in the en route air traffic control computer system. Because of weaknesses in reporting equipment outages, lack of planning, and a poorly defined approach to managing system operations and software changes, the report concluded that FAA could not be certain that the |