duced some 30 computer science and technology special publications reporting on our investigations into such topics as computer software; computer network interconnection; computer security and risk management; memory and storage technology; and auditing of computer systems.

Mr. Chairman, I would like to concentrate now on the biggest part of our mission, which is the development of Federal information and processing standards.

We're the only agency charged with the responsibility for developing Government-wide ADP standards. Since the passage of the Brooks Act, we have developed and promulgated 31 such standards. We have issued standards in the areas of high-level programing, languages, software, data encryption, data codes for information interchange, computer-magnetic storage media, and computer hardware.

The process of developing standards is a long one. I have described it briefly in my written statement and will not repeat it here.

In spite of our conscientious efforts, there have been criticisms of the Federal ADP standards program. And these criticisms have charged that we haven't produced enough standards; we haven't produced the right standards; and we have relied too much on voluntary commercial standardization activities.

We've taken these criticisms seriously. And we have taken stringent steps to overcome them. We've been strongly supported by the Department of Commerce and by the Office of Management and Budget. We initiated a major effort in 1976 to strengthen the standards development program and achieve a more meaningful implementation of the Brooks Act. This effort produced a 5-year program plan centered on the standards development function.

This plan, covering fiscal years 1979 through 1983, was the basis for the President's significantly expanded budget request for the ADP standards program for fiscal year 1979.

I have, in the written statement, included a characterization of major features of the 5-year plan, and will pass it now.

The 5-year program plan places heavy emphasis on the standards in the computer software area. This reflects the overriding importance of software problems, and the growing ratio of software to hardware costs in Federal ADP expenditures. In fact, 75 percent of our expansion requests for fiscal year 1979 for developing standards is targetted on software-related problems. Our total expansion request for the Brooks Act program in 1979 is $13.4 million.

The planned development of high-level programing language standards is aimed at reducing the problems of converting software from one set of equipment to another; reducing the costs associated with program conversion; and eliminating software conversion as an obstacle to competitive procurement of competitive computer systems. In September, last year, the GAO reported to Congress, and estimated that software conversions cost the Government about $4.5 million each year.

These high-level programing language standards on which we are focussing, and their potential impact on software conversion costs, take on even more importance when viewed in the context of one estimate that over 80 percent, or about 8,500, of the general management computers in the Federal inventory will be replaced by 1985.

The 5-year plan that I have referenced and earlier described contains several provisions for strengthening the management of the overall ADP standards program. The particulars of this strengthening are described in my written statement.

The 5-year plan itself is focused on 10 priority families of standards to be developed. And again, I have those listed in my written statement, and will, in the interests of time, pass them here.

In summary, Mr. Chairman, the NBS 5-year program plan is a significant and necessary step toward a more meaningful and effective implementation of the Department of Commerce's Brooks Act responsibilities. In addition to constituting an action plan, the 5-year plan places a price tag on overcoming identified deficiencies in the Federal ADP standards program, and meeting the demand plans for a more meaningful implementation of the act.

We will update and extend this plan annually. We intend in the upcoming plannig cycle to broaden the plan; and to strengthen our research and advisory service program.

And finally, Mr. Chairman, I believe that it is entirely fitting that the NBS, which helped launch the era of the computer, should have responsibility for this program, which has such great potential to influence computer technology for public benefit. I believe that the Institute of Computer Sciences and Technology and its Brooks Act mission are well placed at the Bureau, and that in the coming years can make contributions to the Nation that are in keeping with the Bureau's tradition of service to the country.

Thank you, Mr. Chairman.

[The statement follows:]

STATEMENT OF ZANE THORNTON, ACTING DIRECTOR, INSTITUTE FOR COMPUTER SCIENCES AND TECHNOLOGY, NATIONAL BUREAU OF STANDARDS

Mr. Chairman and Members of the Subcommittee: I am pleased to have this opportunity to describe to you the activities of the National Bureau of Standards' Institute for Computer Sciences and Technology. The Institute will continue to exist as an entity in NBS under the reorganization. The Institute is responsible for providing the Federal computer standards, guidelines and scientific and technical advisory services aimed at helping achieve economy and effectiveness in the Government's acquisition, application and use of computer technology. Our job is a big one because we are attempting to gain leverage on the world's largest user of automatic data processing resources. These resources cost the Government over $10 billion annually and comprise more than 10,600 computers staffed with more than 150,000 technical personnel. The problem of bigness is compounded by the complexity and rapid change of computer technology.

The Institute's mission is encompassed by Public Law 89-306 of October 1965. This law, more familiarly known as the Brooks Act, has the purpose to achieve economic and efficient purchase, lease, maintenance, operation, and utilization of automatic data processing equipment by Federal departments and agencies. It authorizes the Secretary of Commerce to recommend the establishment of uniform automatic data processing standards, provide computer scientific and technological advisory services to Federal departments and agencies, and to undertake necessary research to support both the standards-making and advisory services roles.

I would like to turn now to some of the specifics of what we do.

For nearly three decades, the Bureau has been a source of advice and technical assistance to Federal departments and agencies in their efforts to select, acquire and apply computer technology. The Bureau achieved preeminence in the computer field by designing and building SEAC, one of the world's first internally programmed electronic computers. This work has been done mainly by the Institute for Computer Sciences and Technology and its predecessor organizations. Almost all of this advisory and technical assistance work has been funded by

reimbursements from the Federal agencies requesting help; these reimbursements have averaged $1.8 million annually over the last five years. We have tried to manage this work so that, in addition to benefitting the sponsoring agency, the end product will help other agencies and contribute to the Institute's standards-making and research programs.

Our technical assistance to other agencies covers a broad spectrum of ADP activities, including assessing the technological feasibility of automating functions, designing ADP systems, evaluating vendor proposals for computer systems to be acquired, providing technical monitoring of systems development, and advising on computer security and risk assessment.

I would like to turn now to our research activities. These activities are oriented along two lines, namely, that required as a foundation for the development of standards and that to support our advisory services role.

I can best describe to support standards-making with a few examples. When we were getting ready to develop the Data Encryption Standard to protect sensitive information in computer systems and networks, we had to conduct research into cryptography, the criteria for selecting suitable encryption algorithms, and how to use such algorithms in a computer system.

One of our most important research efforts is aimed at improving the quality of computer software-the most costly and complex aspect of computer usage. Some have estimated that software-related activities account for as much as 80 percent of total ADP costs. We are working to advance knowledge and understanding of the deficiencies and problems in the tools for developing software. Concurrently, we are working on theories and methods for proving the correctness of software, and assuring that computer systems perform their intended functions accurately.

As I mentioned earlier, some of our research supports our advisory services responsibilities to keep Federal agencies abreast of developments in computer technology. During 1977, we produced some 30 computer science and technology special publications reporting our investigations into such topics as computer software, computer network interconnection, computer security risk management, memory and storage technology, and auditing of computer systems.

Mr. Chairman, I would like to concentrate now on the biggest part of our mission which is the development of Federal Information Processing Standards. We are the only agency charged with responsibility for developing Governmentwide automatic data processing standards. Since passage of the Brooks Act, we have developed and promulgated 31 Federal Information Processing Standards. We have issued standards in the areas of higher level programming languages and software, data encryption, data elements and codes for information interchange, computer magnetic storage media, and computer hardware. The development of a standard is a protracted process because

1. Obtaining a consensus position on a standard is time consuming.

2. Standards development task forces are frequently staffed by agency volunteers whose time must be divided between standards and their regularly assigned duties.

3. We have depended frequently upon voluntary commercial standards development efforts that are often prolonged by divergent industry views and interests. The recent General Accounting Office audit of the Federal Information Processing Standards Program cites an average of six years to develop a standard.

In spite of our conscientious efforts, there have been criticisms of the Federal ADP standards program. These criticisms have charged that:

We haven't produced enough standards.

We haven't produced the right standards.

We have relied too much on voluntary commercial standardization activities. We have taken these criticisms seriously and have taken strenuous steps to overcome them. We have been strongly supported by the Department of Commerce and the Office of Management and Budget.

We initiated a major effort in 1976 to strengthen the standards development program and achieve a more meaningful implementation of the Brooks Act. This effort produced a deailed five-year program plan centered on the standards development function. This plan, covering the Fiscal Years 1979-1983, was the basis for the President's significantly expanded budget request for the ADP standards program for Fiscal Year 1979.

The major features of the five-year program plan include:

1. Expedited development of ten families of standards with expected high payoff in cost savings, cost avoidance, and improved productivity in the Government's acquisition and use of computers.

2. Annual updating of Federal ADP standards priorities through working arrangements between NBS, the Office of Management and Budget, General Services Administration, and Federal agencies.

3. Implementation of effective technical and administrative means for measuring and reporting on compliance with ADP standards by Federal agencies.

4. Production of technology forecasts to guide Federal agencies and particularly the central management agencies (NBS, OMB and GSA) in their Brooks Act responsibilities.

5. Development of technical impact assessments and cost benefit analyses for each standard.

The five-year program plan places heavy emphasis on standards in the computer software area. This reflects the overriding importance of software problems and the growing ratio of sofetware to hardware costs in Federal ADP expenditures. In fact, 75 percent of the expansion request for Fiscal Year 1979 for developing standards is targetted on software-related problems. Our total expausion request for the FY 1979 Brooks Act Program totals $13.4 million.

The planned development of high level programming language standards is aimed at reducing the problems of converting software from one set of equipment to another, reducing the costs associated with program conversion, and eliminating software conversion as an obstacle to the competitive procurement of computer systems. In a Sepember 1977 report to the Congress, the General Accounting Office estimated that software conversions cost the Government about $450 million each year. These high level programming language standards and their potential impact on software conversion costs take on even more importance when viewed in the context of one estimate that over 80 percent or about 8500 general management computers in the Federal inventory will be replaced by 1985.

If funded, the five-year plan should lead to the more timely development of standards by strengthening management control of the development process and reducing reliance on purely volunteer, part time labor to staff standards development task forces. Moreover, the plan and adequate funding will give the Government the latitude to reduce dependence on voluntary commercial standardization activities whenever it becomes clear that those activities are not moving at a pace or in a direction that will satisfy essenial Federal goals and requirements. We plan to continue, however, an effective participation in voluntary commercial standardization activities wherever such participation will contribute toward achieving Federal ADP goals in a cost-effective and timely manner.

The five year plan, contains a management mechanism for an annual updating of standards priorities. Equally important, the plan contains requirements for the development of technical impact assessments and cost-benefit analyses to accompany each standard. These assessments and analyses will strengthen the process of setting priorities and quantifying the benefits of Federal ADP standards.

Another key management feature of the five-year plan is the provision for an annual Federal-wide ADP standards assesment and program report. This report will serve as a principle vehicle for assessing the impact each year of the computer standards in place. The report will asses compliance by Federal agencies with each standard, thus filling a serious void in the present program.

I mentioned earlier that the five-year plan is focused on ten priority families of standards to be developed. I would like now to describe briefly some of those ten families:

1. We will develop software quality control standards against which to measure software quality at time of procurement and to improve quality in software development, and high level programming language standards to reduce the cost of software conversion.

2. We plan to develop a family of performance assurance standards to assure that computer systems perform their functions correctly and do not perform any unintended functions. These standards are aimed at eliminating the computer errors that waste Government resources or adversely affect the public's privacy, health, or safety.

3. We will establish computer system functional specification standards to permit agencies to procure computer systems and services on a fully competitive basis with regard to end use.

4. The family of computer system and network interface standards will permit the interconnection of competitively procured components in computer systems and networks. These interface standards will enable the Government to select among alternative suppliers and take advantage of competitive pricing.

5. We will develop data base management standards to reduce data conversion problems and permit the efficient use of files.

In summary, the NBS five-year program plan is a significant and necessary step toward a more meaningful and effective implementation of the Department of Commerce's Brooks Act responsibilities. In addition to constituting an action plan, the five-year plan places a price tag on eliminating identified deficiencies in the Federal ADP standards program and meeting the demands for more meaningful implementation of the Brooks Act. We will update and extend this plan annually. In addition, we intend in the upcoming planning cycle to broaden the plan to strengthen our research and advisory programs.

Finally, Mr. Chairman, I believe that it is entirely fitting that the National Bureau of Standards, which helped launch the era of the computer, should have responsibility for this program which has such great potential to influence computer technology for public benefit. I believe that the Institute for Computer Sciences and Technology and its Brooks Act mission are well placed at the Bureau and that in the coming years can make contributions to the nation that are in keeping with the Bureau's tradition of service to the nation.

Senator STEVENSON. Thank you, gentlemen.

Dr. Ambler, could you explain more about what the rationale is for this reorganization? It sounds, at least superficially, like changes in the names of institutes, job titles, creation of the new Center for Cooperative Technology, but otherwise cosmetic.

What's the effect of this reorganization really going to be? What's its rationale?

Dr. AMBLER. I think there are three substantial benefits that NBS will derive from this reorganization, Mr. Chairman.

The first is that we shall be organized along major functional lines. I have described the dual functions of the Bureau as the Central Reference Laboratory and as the Laboratory that applies its expertise to the solution of particular national problems. Reflecting those roles in the National Measurement Laboratory and the National Engineering Laboratory, respectively, and in clearly identifying the organizations responsible for the Organic Act responsibilities will enable our management to clearly focus on these basic functions of the Bureau of Standards.

Second, this reorganization gives us the opportunity to consolidate competences. We have been operating under the existing structure for many years now, and there is a need to consolidate our existing competences in appropriate parts of the Bureau of Standards. Such consolidation will also give us an excellent basis for building for the future and for adding competences.

The third effect that will be achieved is that we shall achieve a greater flexibility than we have in the present organization. The number of formal organizational units will be substantially reduced, from 249 to 106. The intent is to provide flexibility for changing programs, for changing assignments, without the need for formal transfers and becoming involved in a great deal of paperwork.

Thus, I see these three benefits as being much more than cosmetic, Mr. Chairman, in allowing us to manage the Bureau in a much more efficient wav.

Senator STEVENSON. In your submission to OMB, it was stated that "sections will be eliminated as an organizational level within the reorganized Bureau." I also understand that internal Bureau documents state that sections are going to be retained.

Can you explain that apparent discrepancy?