parison efforts. So, one has a spectrum of people engaged, from the ones primarily engaged in the compilation to some people who like to evaluate work-the scholarly types, and the ones who are the actual researchers that actually have to evaluate who are actually using the numbers.

Mr. FELTON. What happens if the experts disagree?

Dr. HORNIG. Well, this is why one can never settle this matter until there are newer and better measurements. What the experts have to do is in effect statistically estimate the probable errors in all of their judgments and come to a best conclusion on every physical measurement at any given time. You can only have a best value as of 1966. This is the best you can do in that situation. The whole point though is to make sure that this doesn't happen just by random process and by opinions but that you have actually critically compared just these disagreements and arrived at what you consider a best 1966 value for people to use.

Mr. DADDARIO. Is that what it is all about?

Dr. HORNIG. That is really what it is all about, and it is to be sure that this process happens methodically through all of the areas in which it is needed. At the moment it is done, as I said, where one has special interests. The American Crystallographic Association for its members compiles crystallographic data. There are other compilations by various specials groups. But the problem is how do you insure that in the first place it isn't done over and over again, by the overlap, and second, to be sure that it is done in all of the areas in which it needs to be done, and in just the critical way you describe. Mr. FELTON. That is all, Mr. Chairman.

Mr. DADDARIO. Any further questions, gentlemen?

Yes, Mr. Vivian.

Mr. VIVIAN. I have a brief question.

Were there any dispute by other Government agencies as to either NBS role in this particular bill or as to the nature of the bill? I would like to know what principal questions or objections were raised by other agencies of the Government?

Dr. HORNIG. In the first place, of course, the history of this is that in the Federal Council discussions all of the research and development producing agencies felt this needed to be done. I can make available the original Federal Council policy statement. The Federal Council was unanimous in believing that the NBS ought to play the central role in setting standards and in coordinating the activities, all of the data gathering activities. Because it won't all be done in NBS. I mean, for example, the Atomic Energy Commission continues to have primary responsibility for the nuclear data.

The area in which there was discussion-I am not aware of any dissent and none occurred with respect either to the bill or to the policy statement-is the ancient question in the Government, I mean the one of the extent to which any agency should have any authority to prescribe what other agencies might do.

But that authority is not provided in the bill for precisely that reason. This is a problem we have never resolved to my knowledge in any area we have tried to coordinate.

Mr. VIVIAN. Dr. Hornig, I have one final comment. In spite of my earlier comments which might have been interpreted as being

adverse to the bill, I am not adverse to carrying on the work proposed by this bill. I think it is highly desirable because we are going to be outpaced very rapidly by other nations if we don't become more efficient. However, I would like to make sure the data produced are made available on a very open basis.

Mr. DADDARIO. Mr. Conable.

Mr. CONABLE. Dr. Hornig, has anybody made any estimate of how long it is going to take us to get to a current basis in the evaluation of data?

Dr. HORNIG. Well, I don't really know what you mean by that word "current." If one really wanted to tabulate and assemble just everything that was known, in the first place we would be talking about an order of magnitude bigger than any effort one is thinking about here. I don't believe anyone really would want to compile this massive file of all current data. So that the question really should be how long would it take us and at what expense to compile all the currently important or useful or worthwhile data, and I am afraid I can't answer the question.

Mr. CONABLE. It is a very big project, isn't it?

Dr. HORNIG. Well, what we know is that the needs are much greater than we are carrying on now. This we are sure of. One has to constantly study the question of when diminishing returns set in which is the questions we are really asking.

I am sure we can go another factor of 10 from where we are now in the efficient correlation and assembly of compiled data. But at some point the further efforts wouldn't justify the costs.

Mr. CONABLE. The reason I ask is that one of the scientists I wrote to expressed himself as simply appalled at the scope of this program. He said that he did not see how the standard reference system data could do more than concentrate on the production and yearly updating of a comprehensive annotated bibliography. Is that about all you could really hope to accomplish?

Dr. HORNIG. This sounds too black and white to me. He lives right now-I don't know who he is, but he lives right now on volumes such as the ones I have shown here and on the literature.

Now, it doesn't follow that if I want to improve this situation that I have to go to an all-out effort, that there is nothing in between, and I think my simple answer would be that I could do three times better without getting anywhere near that state. At some point, as you said, diminishing returns would set in.

Mr. CONABLE. You would envision this program as a fairly uniform yearly effort and not some sort of a massive effort at first with thereafter a simple updating?

Dr. HORNIG. Yes. I think we know right now that we need a much bigger and more comprehensive effort than we have now. So I would envision-and this was envisioned by, I might say, the Federal Council and all of the agencies concerned some years ago the condition that we would build up to a level of effort considerably greater than the one we have now.

At that point, perhaps you mean in the ordinary appropriation process, where one would consider the adequacy of the services being performed on the one hand or the possible redundancy, though I

think we are far from that, on the other hand of the services being performed and make continuing decisions as to whether continuing growth is in order or whether one levels out. I think it is a dynamic question that is going to have to be decided year by year.

Mr. DADDARIO. And as you do it you get more efficient and expend your dollars so as to increase your capability?

Dr. HORNIG. That is right.

There is both the data itself, and the thing we have to learn a great deal about, the methods of disseminating data so the right data gets to the right man.

The thing that is terribly wasteful is the thing we don't know simply how to do anything about, which is the case of the engineer who doesn't know that the number he wants is available somewhere. I have never seen any magical, easy ways of being sure that it gets to him.

Mr. CONABLE. That is all, Mr. Chairman.

Mr. DADDARIO. Mr. Brown.

Mr. BROWN. Can I just make this inquiry. Do you envision any limit, that you can verbalize, as to how far this would go in terms of either the nature of the reference data or the point at which you would not proceed further because it is becoming too complex and comprehensive or redundant? The thought that occurs to me is that all reference data is really a continuum. For example, take a simple matter of the expansion coefficient of a metal. This varies all the way from absolute zero to whatever the change of state is. Why, to define the thing precisely, for rare metals, you would have to go through this whole link. Is there some way in which you can establish commonsense limits to how far you go in this?

Dr. HORNIG. Yes, in a certain sense one has always to weigh the cost of producing and disseminating more against the value to the users, and I think this is a thing you do continuously. You are quite correct in saying one could go to absurd lengths in tabulating everything. But I don't think I would ever be willing to lay down any absolute rules because activities of the users, the engineers, the scientists, the industrialists, keep on changing, too.

So I would say that as long as there are significant needs and valuable needs not being met, it is time to expand. When you find yourself-and this perhaps relates to user charges-when you find yourself publishing and compiling things that nobody uses, then you probably have gone too far.

Mr. BROWN. In other words, it is difficult then to establish either in the law or in a statement what the limits might be. You see a gap which at present needs to be filled, and within the limits of whatever funds would be made available, you would proceed to fill that gap. Then by an empirical means determine how much further it might be useful to go or whether it has already gone too far.

Dr. HORNIG. That is essentially correct. In the discussions we had in the Federal Council, for example, the Department of Defense, NASA, all the agencies involved, were clear and unanimous about the fact that for all of the industry with which they deal, and for their own internal scientific and technical efforts, that the present magnitude of the effort doesn't fulfill the need and so I believe a substantial, perhaps several times expansion, is in order.

But I think that after the establishment of such a system, that the question you ask would be settled really in the annual appropriation process, for the future, as to whether further expansion is or is not justified.

Mr. BROWN. That is all, Mr. Chairman.

Mr. DADDARIO. Any further questions, gentlemen?

(No response.)

Mr. DADDARIO. Dr. Hornig, thank you. We always learn a little bit more when you come before this committee. This committee will adjourn until tomorrow morning at 10 o'clock in the same place. ADDITIONAL QUESTIONS SUBMITTED TO DR. DONALD F. HORNIG, DIRECTOR, OFFICE OF SCIENCE AND TECHNOLOGY, BY THE SUBCOMMITTEE ON SCIENCE, RESEARCH AND DEVELOPMENT

Question 1. In your testimony you infer that user charges would only be assessed for special services performed for particular groups. This being the case, would you discuss why the Bureau's present authority to make special service charges is inadequate (15 U.S.C. 275a) to meet the needs for standard reference data?

I

Answer 1. The inference in question #1 is correct in that I believe that user charges should be assessed for special services performed for particular groups. The difficulty comes in defining "special services" and "particular groups.” would defer to the Department of Commerce for the particulars on inadequacies of present authorities for the National Bureau of Standards. As I understand it, the particular authority it has is to make service charges for performing specific tests and measurements and special services. In the proposed standard reference data program it is constrained by other laws relating to government publications where only minimal printing costs can be recovered, and even these are not credited against the specific program but are returned to the general Treasury.

Question 2. It has been estimated that the Federal government supports in excess of three-fourths of the total research and development performed in the United States. Is it reasonable to say that at least three-fourths of the users of standard reference data will be operating either directly or indirectly under Federal grants and contracts? If not, what is your estimate, and how did you arrive at that figure?

Answer 2. I would not estimate that three-fourths of the users of standard reference data will be operating either directly or indirectly under Federal grants and contracts. Although that may be a valid ratio for scientists and engineers performing federally supported research and development, they are only one class of users of standard reference data. This standard data will be referenced and used also by several other very large classes of users-the professional practitioners, the private manufacturers of commercial goods, medicines, and products, the teachers, the under graduate students, the merchandisers, the design, manufacture and maintenance components of the whole transportation sector, and construction companies. The standard reference data system then will serve not only to further science and technology per se, but is one means for transferring the fruits of research and development into the other sectors of the nation's activities and welfare. Recent surveys by the National Science Foundation indicate that only 35% of all scientists are employed in research and development. Further, only half of the 400,000 scientists are engaged in work supported by Federal funds, regardless of field of work. It is suspected that an even smaller ratio would apply to the one million engineers, and then there are the many scientists and engineers in foreign countries who are not supported by United States federal funds, but would use United States standard reference data. A guess based on the above factors might be that perhaps one-fourth to one-half of the users of standard reference data will be operating under Federal grant or contract funds. Question 3. In your testimony you stated that "... the use of reasonable user charges is one of the best ways of determining the value of services performed . . .”

(a) Is it unreasonable to expect that the experts who advise the Bureau that standard reference data be compiled in a particular field are also competent to advise the Bureau if the compilation is of value to the scientific community?

Answer 3. It is reasonable to expect that the "expert" advice to compile data in a particular field will be based not only on the absence of such compilations, and their feasibility (state-of-the-art-wise), but also on the user needs for the data. Each of these three factors is a necessary condition and no two of them are sufficient. The same individual expert might not necessarily advise on all three factors with objectivity and competence. I believe the National Bureau of Standards can be relied on to seek the variety of expert advice that is required to reach a judicious decision based on all three factors. I understand that in each particular field an attempt will be made to estimate the number of potential users. Either the number of users or the criticality of use can be determining factors.

For the next few years, however, the situation will be one of competition for resources among many subjects fields. The technical experts advising the National Bureau of Standards would not, in my opinion, be able to advise, except in the grossest, most qualitative way, on the relative economic (as opposed to scientific) benefits to be derived from data compilations in different subject areas; until there are some standard measures of the value of information, no one is going to be able to prove specific value.

Question 4. Woud the agencies of the Federal government that make research and development grants, contracts or other arrangements with individuals or organizations be required or encouraged to secure from the recipient a commitment to turn into the system data developed with Federal funds?

Answer 4. Federal Government research and development contracts, and research grants too, provide for Federal rights to the results of the work, including rights to information. It is not, therefore, necessary that this legislation also require each contractor to "turn into the system" the data developed. The standard reference data system's principal problems are with respect to data quality, evaluation, standards and format, rather than the "ownership" of data. Question 5. What is the Administration's copyright policy regarding copyrighting by Federal agencies?

(a) What is the rationale behind the policy?
(b) Is this bill consistent with that policy?

Answer 5. The Administration's policy with respect to copyright by Federal agencies is to conform with existing law. As you are aware, the existing copyright law is under review by the Congress. I personally feel that some problems that have arisen as a result of greatly expanded scientific and technical activity and concomitant changes in information forms and activities, are not adequately covered by present laws. However, the objectives and purposes of the present copyright law remain honorable and valid.

The bill under discussion (H.R. 15638) as drafted, seeks an exception to existing general practice of copyright. It is consistent with the afore-discussed objectives and purposes, that is, in this particular case it is judged to be in the public interest to give the Secretary of Commerce certain responsibilities and prerogatives with respect to the data compilations. I am encouraged that your Committee is taking a conscientious look at the issues. Out of such an approach can come the best judgment of just what specific responsibilities and prerogatives are warranted. Appropriate safeguards (e.g., time limitations and attribution) on the necessary responsibilities and authorities may not be present in the bill as now drafted. Perhaps these authorities and safeguards cannot be fully defined and itemized in the few words of a bill, but may require the context of considerable discussion to carry the intent and meaning for future specific application and adjudication. I would regret undue deferment of needed legislation just to achieve such refinement.

(Whereupon, at 11:31 a.m., the subcommittee was adjourned to reconvene at 10 a.m., Wednesday, June 29, 1966.)