the Nation's foremost scientists and engineers, (b) results of a survey of the needs for numerical data among 100,000 scientists and engineers, a survey conducted by the American Chemical Society, (c) numerous discussions with individual consultants and leaders of existing data compilation projects, (d) consultation with research and development program leaders in mission-oriented agencies, and (e) awareness of foreign and international data compilation activities and plans.

The proposed Standard Reference Data Act would enable us to apply user charges to the costs of the program. Therefore, the appropriations required would be reduced by the amounts realizable from such charges. Although these amounts cannot be estimated in advance with precision, our present estimates indicate that a significant, but perhaps not large, fraction of the total cost of operation would be recoverable.

In the development of the program of the Standard Reference Data System several alternative approaches have been considered. These alternatives may be summarized in the following way: (1) Continue the present pattern of uncoordinated response to especially urgent needs; (2) set narrower goals for NSRDS; (3) set broader goals for NSRDS; and (4) rely on the other approaches to the scientific information problem such as is envisioned in programs of Chemical Abstracts, the American Institute of Physics, the American Society for Metals, and other nongovernmental professional societies. Each of these possible alternatives will now be discussed.

(a) Continue present system: It was the deficiencies of the existing situation that led to the establishment of the Standard Reference Data System. Nevertheless, the technical effort of the United States was making progress in the absence of the SRDS, and it can continue to do so. A properly operated Standard Reference Data System may increase the overall efficiency of the Nation's technical activities by perhaps 1 percent or perhaps as much as 5 percent.

An inefficiency of this magnitude is tolerable in the total effort; indeed, it is hardly detectable. It cannot be claimed therefore that an essential goal of the United States will not be achieved if the SRDS is not fully implemented as rapidly as possible. It is claimed, however, that hundreds of millions of dollars will be unnecessarily spent in the achievement of those goals unless the SRDS is effectively imple

mented.

(b) Set narrower goals: The goals of the Standard Reference Data System have been circumscribed by the definition of the technical scope of the program. That is, the program is to be concerned with data on the intrinsic properties of well-defined substances and their interactions.

The goal of the 5-year program is to achieve 90 to 95 percent coverage of all the properties which are determined to be of high priority as determined by surveys and consultations with specialists. This goal has been set because it is achievable with the expenditure of modest funds, it is probably within the capability of available manpower in the United States, and the scope is harmonious with the traditional areas of expertise within the National Bureau of Standards.

Clearly, the goal could have been set at a more modest level. The practical realities of funding within the National Bureau of Stand

ards force us to the conclusion that it is unlikely that we shall attain the resources to carry out the program at the level stated. One of the very attractive features of this program is the fact that the benefits to be obtained at all levels of expenditure up to four to five times the present level are proportional to the amount expended. The region of diminishing returns is not reached until activities begin on those properties which are identified as of less than "high priority."

(c) Set broader goals for SRDS: The benefits to be obtained from systematic data compilation activities are not exhausted when the scope defined for SRDS have been covered. Many properties of imprecisely defined materials-concrete, plastics, steels, et cetera-are of enormous importance for technological purposes. Many types of data gathered in the earth sciences are now outside the scope of SRDS, but make vital contributions to understanding of the environment.

A program to include these other types of data would double or triple the resources in money and manpower required. Some of the areas involved are outside the traditional range of activity of NBS. The availability of technical manpower to accomplish such broader objectives is also questionable.

For these reasons, no plans have been made within NBS for enlarging the scope of our planned program. As additional experience is gained through operation of the program under its present restrictions, it will be necessary to review the scope regularly in order to determine whether broader goals are appropriately within the mission of NSRDS.

(d) Rely on other approaches to the scientific information problem: The problems of making scientific and technical information of all kinds more readily available to the technical user who needs it are receiving great attention from various professional societies in the United States and throughout the world. In particular, the American Chemical Society, the American Institute of Physics, the American Society for Metals, and the Engineers Joint Council are beginning to plan extensive programs.

These programs are all concerned with the entire substance of the disciplines within their purview-concepts, experimental techniques, interpretations, data, and practical applications. Ultimately, the programs now being formulated by these societies will, if fully implemented, have a profound effect on the operations of the numerical data centers that comprise the Standard Reference Data System. To implement the concepts being developed, however, requires the solution of intellectual problems of great complexity, as well as the solution of management and funding problems of equal difficulty and complexity.

It is unlikely that the work of these organizations can make any significant contribution to the numerical data problem for a period of 5 to 10 years. To delay the opportunity of achieving clearcut, immediate gains through the well defined programs of SRDS while awaiting as-yet-to-be-conceived solutions to larger problems is not in the interest of scientific and technological efficiency or progress. Thank you, Mr. Chairman.

Mr. DADDARIO. Dr. Astin, if we could touch for a moment on the section that you just completed. You refer to the work proposed by

the American Chemical Society and these others indicate that they can do the job, but that it will take time. I wonder if you could give us some indication as to how they feel about this bill, and whether or not any thought has been given to supporting their efforts so that they could achieve their goals in a lesser time?

Dr. ASTIN. The primary support of these programs, according to my understanding, already is provided by the National Science Foundation. But everybody, including these groups involved, appreciates the great importance and complexity of the total scientific information problem and are trying to find ways of accelerating meaningful solutions to the total problem. It is recognized by the people involved in those programs however, that the solution of the numerical data information problem is a much more straightforward and simpler one and one with immediate benefits. So, according to our understanding, all of the people concerned with the broader information problem are anxious to have us go ahead with the solution of the numerical data information problem, because nothing we will do on that will be of a form that cannot ultimately be integrated into the broader information activities.

Mr. DADDARIO. Then your feeling is that they believe this effort will be complementary to theirs and not in conflict?

Dr. ASTIN. That is correct.

Dr. HOLLOMON. Mr. Chairman, what this group is trying to do, which we refer to here, is to develop an effective system for the information storage and retrieval scientific information. This is an extraordinarily complex task. The business of critically evaluating the specific numerical values and publishing it is much more direct and can be accomplished immediately.

To answer your second question, we have worked out arrangements with these societies-the American Chemical Society to be more explicit-and where they are competent to provide the critical evaluations, we will cooperate with them. That is intended to be a part of the program.

Mr. DADDARIO. What is the nature of these arrangements?

Dr. ASTIN. Well, our people in the Office of Standard Reference Data have close contacts with the information programs of these other agencies. There are many common problems.

For example, the first phase of the Standard Reference Data problem is the search for literature. It is also a problem that is common with the broad scientific information problem. So, in improving techniques of searching for literature and in making indices of what is in the literature, we have a common interest. In this we are working together closely.

Mr. DADDARIO. Then we should consider your explanation here as not being an alternative through which this can be accomplished but rather an supplementary kind of activity?

Dr. ASTIN. That is correct.

Mr. DADDARIO. Is it in that nature?

Dr. HOLLOMON. I quite agree.

Mr. DADDARIO. The matter of sharing costs, which you both touched on, seems to be of some concern to the committee. Mr. Vivian and others went into that yesterday. I am sure there will be further

questioning about this. I would like to ask how we reconcile the present authority, that you have, as shown in the United States Code, with a charge for services.

Dr. HOLLOMON. The question is what is meant by services. Publications are not considered to be services. We can charge currently under our present authority for such things as calibrations, standard reference materials, but we do not charge for initial publications under that authority.

Mr. DADDARIO. Would the recoverage from service charge be something less than the cost of a full publication if you were asked to provide specific information of a specific nature?

Dr. HOLLOMON. We would wish to recover the significant part of the cost of the materials, the printing, and the reproduction, and some part of the editorial and compilation costs.

Mr. DADDARIO. What is your view of that Dr. Hollomon and Dr. Astin? What do you consider it to be appropriate? Do you feel it to be helpful rather than harmful?

Dr. HOLLOMON. We have under the present authority the ability to recover substantial direct costs and some of the associated costs. If we couldn't recover that cost, that is if it didn't go back into a trust fund, the more service we rendered, the less appropriation would remain available to develop new methods and means for providing essential calibration services. Under such conditions, the better we did our job, the less money we would have to do it with. That is the real intent and thrust of the trust fund activities for calibration and standard reference materials. It allows us to recover for the services supplied to these people who need it, a fraction of the cost.

The second problem is how do you determine really what the user need is. How do you determine that we are doing something that is really useful to the scientific and technical community. One of the ways is to have them pay for part of it. Take the case of a compilation of data, which would be very expensive intellectually-by this I mean having scientists go in and evaluate it and so on-and for which there are only a relatively few possible customers. I think in this instance you could argue that the Nation will be better served to be sure that data is available. But if the compilation is made available at no cost, or at a relative insignificant fraction of cost, you would have no easy mechanism to determine whether you were doing something that was of service.

In the selling of reference materials and calibration services, and one of the real measures to find out whether this service is useful, is whether the user is willing to pay a share of the cost. So, my own feeling is that some fraction of the costs of these reference data services should be borne by the specific user who can be identified. In large measure, these are going to be institutions and libraries for these kinds of data.

Another aspect of the problem has to do

Mr. WAGGONNER. Would you stop there just a moment? Where do nonprofit institutions and libraries fit into this picture of paying user

costs?

Dr. HOLLOMON. Nonprofit institutions and libraries would under the proposed system pay a part of the charge necessary.

Mr. WAGGONNER. Less than the private institutions?

Dr. HOLLOMON. No difference. There would be no difference. The charges would be the same to all comers. We would not try to discriminate between one person and another.

Mr. DADDARIO. Mr. Waggonner has a good point. Are you preventing people who need the information from getting it?

Dr. HOLLOMON. I don't believe that is the case. In no case do we intend to recover full costs. In no case can we recover full costs, nor do we propose to.

Mr. DADDARIO. I am persuaded somewhat by the argument which was raised yesterday by Dr. Hornig, but I am not convinced that it is so. I wonder will people believe that the information is good because they pay for it, or would they ask for it anyway even if they didn't have to pay for it?

Dr. HOLLOMON. There are two points to the question which I will repeat. First, it is my view that some fraction of the cost should be recovered in order to have a measure of whether or not this service is useful. I think we should have some mechanism of making that determination. It has been successful in two other services that the National Bureau of Standards renders. One is calibration services, as I say, and one is standard reference materials.

You calibrate an instrument to make sure it makes the same measurements throughout the country. This is not unlike the business of having a number which you are sure is reasonably accurate. We do the same thing with standard reference materials.

I would not believe that we should recover all costs, this is because in some instances the cost of the evaluation and compilation may be very high, particularly in a highly complex technical subject. But on the other hand, we do believe that something more than just recovering paper and printing costs, would be appropriate.

Mr. MOSHER. Do I understand correctly that the user charges will be made on some data but not on others?

Dr. HOLLOMON. No. We would intend to have user charges on all data, but the degree to which you would apply the original intellectual cost may be different depending on how much that relative investment is against the total amount that you would expect to sell.

Mr. MOSHER. The discretionary authority would lodge in the Secretary?

Dr. HOLLOMON. Yes.

Mr. MOSHER. Would the charge be whatever the market will bear? Dr. HOLLOMON. No; I do not think that is the case. I think a valid judgment would have to be made-Dr. Astin may wish to speak to itas to the degree to which this data were critically needed and, to some degree, how many possible customers there are. I don't think it is what the market will bear in this sense. Let me turn it around in another way.

Mr. MOSHER. In other words, the price you charge will be based on the demand?

Dr. HOLLOMON. I would put it a little bit differently than that. Let me give you a specific example. Here is a book of mathematical tables. It cost, as I remember, approximately, $500,000, to produce. This includes the compilation of the tables, the editorial work, the