by the National Archives in 1964 concluded that the disposition of machine-readable records had to be brought under control and recommended that a government-wide committee be established to consider the proposal of the Social Science Research Council for a federal data center. 13 The ensuing debate over a federal data bank prevented significant action until 1967 when the archivist of the United States established the Committee on Disposition of Machine-Readable Records, which was to undertake a detailed study of machine-records in the possession of federal agencies. 14 In January of 1968 the committee presented its findings, including a recommendation that a special unit to deal with machine-readable records be created.15 The following year the archivist of the United States established a data archives staff to preserve and to make available the computer records of the federal government. The new organization began to survey the government's magnetic tape libraries, preparing inventories and identifying for preservation files of permanent value. Because of the magnitude of this task, the staff expanded first to branch and later to division status. 16

The task facing the Machine-Readable Archives Division is awesome. The federal government now has the equivalent of eleven million reels of magnetic tape in use or in storage. Computer-based information in the federal government is increasing at the rate of at least 6 percent each year. The fundamental problem confronting the division is not so much the eleven million reels of magnetic tape and the annual increase as it is the fact that magnetic tape as a storage medium has different properties than paper.

Generally, paper is used only once because the cost of restoring it to its original condition is prohibitive. Magnetic tape, however, can be used repeatedly without its condition being seriously affected. The ease with which magnetic tapes can be erased and reused has given rise to situations in some federal agencies whereby inactive magnetic tapes containing

13 "Report of the Committee on the Preservation and Use of Economic Data to the Social Science Research Council" (Social Science Research Council, 1965), p. 1.

14 'Data Archives Staff Report," p. 2.

15 Ibid.

16 Gerald J. Rosenkrantz was the first chief of the branch and served until August 26, 1974, when the writer became division director.

valuable information have been erased to supply tapes for other projects. A second difference between paper and magnetic tape is that the latter is very fragile. Paper records exposed to a wide range of temperatures and humidity, rough handling, and being left untended for years generally will not become unreadable. The reverse is true of magnetic tapes. Fluctuations in temperature and humidity can seriously affect the readibility of magnetic tapes. Fingerprints on magnetic tapes contain moisture and body oils, which, along with dust, may cause permanent damage. Even when magnetic tapes are stored under controlled temperature and humidity they still must be cleaned and rewound periodically. Finally, magnetic tapes in long-term storage must be read by computer periodically to ensure that no erosions have occurred.

Obsolescence is another problem of tape that differentiates it from paper. Information recorded on paper hundreds of years ago is still readable, or usually can be made so, because the translating medium-the human eye-remains the same. However, rapid advances in computer technology have made obsolete some magnetic tapes on which the 1960 census returns were recorded. There are some sixty-five hundred of these tapes now being preserved by the Bureau of the Census that can be read only by using a type of tape drive that is no longer being manufactured. It is reasonable to anticipate that other magnetic tape files will become obsolete as industry develops better tapes and tape drives.

The last significant difference between magnetic tape and paper is that the information recorded on tape is not visible to the unaided eye. Paper records can be scanned to determine their general content and even if information about the filing system used in creating and maintaining them is not available, a trained researcher can still browse through them and sort out the important records from the unimportant ones. Though the information on tapes cannot be seen, it is there and can be read only by a computer. However, for a computer to read the tape it is necessary to know the technical aspects of how the records were stored on the tape. Furthermore, additional information is necessary before the data on the tapes can be deciphered. The fact that a computer can read a tape and then print out the contents is no guarantee that the printout will be intelligible.

It becomes intelligible through the use of a codebook that specifies what values the printed characters represent. This is especially true where numeric codes are used to represent items such as race, sex, marital status, and the like.

The characteristics of magnetic tape pose a number of problems for the division, the first of which is simply that of identifying tapes that contain information of long-term value. The experience of the National Archives has indicated that only a small portion of records have enduring value. Therefore, a set of guidelines has been developed to assist in identifying tapes that should be considered for permanent retention. 17

Because magnetic tape requires a controlled storage environment, federal agencies are encouraged to offer to the National Archives tapes that may have long-term value as soon as they are no longer in active use. Most federal agencies are cooperative once they learn of the value of their tape files. Sometimes the peculiar circumstances of an agency do not permit such cooperation. If administrative and legal requirements prevent transfer of tapes to the National Archives, they remain in the custody of the agency. Few agencies have the resources to preserve inactive tape files so as to ensure that when they are eventually offered to the National Archives they will be readable. A case in point concerns the Internal Revenue Service. In 1970 the archivist of the United States notified the Internal Revenue Service that the statistics of income tape files could not be erased. Since then some of the files covering the late 1960s and early 1970s have been offered to the National Archives. It is estimated that 40 percent of the remaining files are either unreadable or now lack adequate documentation. Had these files been donated six or seven years ago it is likely that valuable economic research data could have been preserved.

Before a magnetic tape file is accepted as part of the National Archives permanent holdings, a thorough appraisal of its value is undertaken. This appraisal in part is similar to that performed on paper records offered to the National Archives. In addition, certain technical infor

17 This includes general records schedule 20, Data Automation Program Records, for machine-readable records, and "Guidelines for Identifying Machine-Readable Records of Potential Archival Value." The second one is much more specific.

mation about the condition of the tape file must be obtained. This requires examining the documentation to determine if it is detailed enough to permit its use in the future by persons unacquainted with the file. It is equally important to establish whether the computer can read the information on the tape. If the documentation will not permit a computer to read the tape or if the tape is in such deteriorated condition that it cannot be read, then it is valueless. Assuming that the physical condition of the tape is good, it is important that its readability be validated against the values specified in the codebook. The validation is made by comparing a partial printout of a tape with the codebook to identify inconsistencies or

errors.

An appraisal such as this requires an archivist who is expert in the subject matter, statistical analysis, and computer documentation. This expertise must be complemented by the ability to sense a variety of possible uses of the same data by researchers asking different questions. For example, the appraiser making this kind of judgment might conclude that an economist, a historian, and a sociologist could use the same data to answer different questions.

Preservation of magnetic tapes that contain information of archival value remains a problem. Presently, the division has a fireproof vault at the Washington records center in which the temperature and humidity are carefully controlled. As noted earlier, long-term preservation requires that tapes be rewound and cleaned annually. Since the division's holdings now number less than fifteen hundred reels, that is easily accomplished. It will become more difficult as the number of reels approaches 2 percent of the tapes used and stored by federal agencies. 18 The number would then be about 220,000 and would double every fifteen years. The resources required to preserve permanently holdings of this magnitude will be enormous. Another equally problematic aspect is that major changes in computer technology are occurring about every ten to fifteen years. One way to avoid obsolescence would be to convert the existing tape files so they could be used with the newer computers. This, too, would soon become unmanageable. Hence, realistic

18 Based upon the experience of the National Archives, between 1 and 2 percent of paper records created each year are eventually selected for permanent retention.

long-term preservation of machine-readable records demands a storage medium other than magnetic tape. Several systems that use laser beam technology to record and compress data on a photo-sensitive medium are now on the market and a review of their strengths and weaknesses is under way.

As do other divisions of the National Archives, the Machine-Readable Archives Division provides reference service for its machinereadable files. Though the division is relatively young, it has identified a potential clientele of over six hundred users. Recently more than two thousand persons were sent a questionnaire concerning their interest in obtaining a catalog that describes in non-technical terms the magnetic tape files available from the National Archives. About 30 percent requested that the catalog be sent to them. The catalog, which is the first official effort of the division to publicize its holdings, is available free of charge from the National Archives. 1

19

The level of reference service the division offers consists of providing copies of tapes to as many users as possible in a format that conforms to the computer hardware requirements where the data will be processed. At present the policy of the General Services Administration and the resources of the division do not permit special purpose programming or creating subsets of data files if either requires a unique computer program. When a computer program developed for a file in the division's holdings is available, it is included at no further cost to the user when he orders the tape file. The fee now charged for copying tapes is $60 per reel regardless of characters per inch, data compression, or number of tracks.20 A printout of all or part of a reel of tape may be

19 A request for a copy of the catalog should be sent to the Machine-Readable Archives Division (NNR), National Archives and Records Service, Washington, DC 20408.

20 Under authority of the National Archives Trust Fund the cost of copying documents plus 10 percent may be charged

obtained at a machine-time cost of $150 per hour.

The limitations on the division's capacity to provide subsets of data files as well as the relatively high fees raise questions about the quality of reference service the division can offer compared with other divisions. For example, unlike other divisions the Machine-Readable Archives Division must charge a fee for search time on a computer. Furthermore, if a user wants only part of a tape copied, he must pay for the entire reel. Some people believe that the fees now charged for copying tape files or printing out the contents of a tape file may result in only the well-funded researcher having access to the information in the division's holdings. A parallel concern is that some researchers do not have access to a computer that can process the data. It would be ironic if in spite of the Freedom of Information Act, the expansion of computer-generated information in federal agencies resulted in fewer researchers having access to information. Certainly the questions of what level of reference service the division will provide and how much to charge for this service merit serious attention.

The premise underlying this article is that the National Archives stands between the proliferating computer-generated information within federal agencies and the increasing number of researchers whose skills and interests necessitate use of this information. Consequently, the National Archives has entered fully into the change in information processing. Systematic efforts now are being made to ensure that through proper identification, appraisal, and preservation machine-readable records of archival value being created today will be available for use by researchers of

tomorrow.

to users. For machine-readable records this includes computer time, magnetic tape, mailing, and staff time. Since the National Archives does not have its own computer, it has little control over rates for computer services.

KENNETH MACDONALD JONES

On August 1, 1945, the Baltimore Sun com

mented on the "recent rash of bills" in Congress for federal support of science and asked, "Why this sudden outburst of interest in science?" The answer, according to the Sun, lay in a "plain-spoken" report entitled Science, the Endless Frontier which had been released on July 18 by Dr. Vannevar Bush, director of the Office of Scientific Research and Development. Bush had prepared it with the assistance of four committees of prominent men in response to a request the previous November from President Roosevelt for recommendations on the following points:

First: What can be done, consistent with military security, and with the prior approval of the military authorities, to make known to the world as soon as possible the contributions which have been made during our war effort to scientific knowledge? Second: With particular reference to the war of science against disease, what can be done now to organize a

1976 by Kenneth MacDonald Jones

This article was prepared largely on the basis of research in a file of newspaper clippings in the President's Personal File of the Harry S. Truman Library. It is doubtful that so comprehensive a collection of newspaper sources could be found in any but the largest research libraries. Its appearance in Prologue is due, therefore, not only to its intrinsic merit but also because it helps to demonstrate the extensive possibilities for research in the presidential libraries.

program for continuing in the future the work which has been done in medicine and related sciences? Third: What can the Government do now and in the future to aid research activities by public and private organizations? Fourth: Can an effective program be proposed for discovering and developing scientific talent in American youth so that the continuing future of scientific research in this country may be assured on a level comparable to what has been done during the war? 1

In his report Bush offered two immediate and two long-term recommendations. He wanted the armed services to discharge as soon as militarily possible those men with proven scientific ability so that they could resume their training. Bush also advocated the early release of scientific information developed during the war that would be of use to industry. Of more lasting significance, however, were his recommendations for a postwar national science policy. He believed that government scholarships in science and an independent agency to provide "a focal point within Government for a

'The impetus behind Roosevelt's letter to Bush was complex. For a discussion, see Milton Lomask to Science, Oct. 12, 1973, p. 116; reply by Daniel Kevles, ibid., Mar. 1, 1974, p. 798. An article by Kevles will soon be published in Isis. A more contemporary account is Lyman Chalkley, "Prologue to the U. S. National Science Foundation," series 8, box 1, folder 1, Kilgore Papers, University of West Virginia Library. Vannevar Bush, Science, the Endless Frontier (Washington, 1945), p. vii.

The Nazi secret weapon: the wingless V-2 rocket, against which no defense other than capture of its launch sites was developed.

Flugbahn einer Fernrakete mit Tragflügeln

concerted program of assisting scientific research conducted outside of Government" were required. In addition to furnishing funds for basic research, this agency, the National Science Foundation, should "formulate a national policy for the Government toward science." 2 Although others had made similar suggestions previously, Bush's report was a more comprehensive study and its publication generated a great amount of public discussion.

This discussion revealed the image of science and scientists Americans held before the atomic bomb was used. Four major questions should be raised. What image of science emerged from the World War II experience? What did those who accepted Bush's contention that scientific research should be supported expect science to provide in the postwar world? Given the wartime contributions and the potential of science, what was its proper relationship to government? Finally, what was a scientist, and what was his relationship to the society he would serve if Bush's recommendations were accepted? The intellectual framework established by the answers forms a base against which the impact of the atomic age and the cold war on the policy of the United States government toward science may be measured.

2 Bush, Science, p. 25. Bush's original title was "National Research Foundation."

Reichweite: 750 km