Link to original WordPerfect Document here
COMMENTS OF PUBLIC CITIZEN ON NARA PROPOSAL TO REVISE GENERAL RECORDS SCHEDULE 20 59 Fed. Reg. 52313 (October 17, 1994) Contents
- I. THE SCHEDULE IMPROPERLY PROVIDES GENERAL AUTHORIZATION, TO DESTROY THE ELECTRONIC VERSION OF RECORDS IF A HARDCOPY EXISTS.
- II. THE PROPOSED SCHEDULE FAILS TO ENSURE THE PRESERVATION OF ALL INFORMATION IN THE ELECTRONIC RECORDS.
- III. THE SCHEDULE ERRONEOUSLY STATES THAT SYSTEM BACKUPS ARE NOT RECORDS.
Public Citizen, Inc., a nonprofit, public interest organization, hereby submits the following comments on the National Archives and Records Administration's ("NARA's") proposal to revise General Records Schedule 20 to give agencies blanket authority to destroy electronic mail, word processing systems, spreadsheet and other electronic records after they have been converted to hard copy. Public Citizen opposes this proposal as an unlawful effort by NARA to avoid its responsibility to appraise and, where appropriate, preserve in electronic form the electronic records that are increasingly being created by federal agencies.
Ordinarily, before a government agency may destroy its records, it must give public notice and the Archivist must appraise the records to determine whether they have sufficient administrative, legal, research, or other value to warrant their continued preservation. 44 U.S.C. Sec. 3303a. However, if records fall within one of the categories of records listed in the Archivist's "General Records Schedules," all agencies are authorized to destroy the records without such notice or appraisal. In order to approve such a General Records Schedule, the Archivist must make a determination that all the records that the schedule describes "will not, at the end of the periods specified, have sufficient administrative, legal, research or other value to warrant their further preservation by the United States Government." Id. Sec. 3303a(d).
In its October 17, 1994 Federal Register Notice, NARA has proposed to revise General Records Schedule 20 to give government-wide authorization for agencies to delete the electronic versions of correspondence, directives, studies, reports, spreadsheets, and other Federal records whenever an agency decides that the electronic version is "no longer needed," if the records have been converted to hard copy (i.e., paper or microfilm). Indeed, if the Archivist adopts the proposed revisions, agencies would not only be permitted, but required, to destroy such electronic records. See 36 C.F.R. Sec. 1228.40 (application of General Records Schedules is mandatory).
If approved, the proposed General Records Schedule would result in the destruction of the electronic version of massive volumes of Federal records. The unique features of the electronic version of these records, which make them particularly valuable for research, will be lost. Moreover, the mandate to destroy the electronic version of such records would apply to all agencies, without regard to the importance of, or public interest in, the records of an individual agency.
The proposed Schedule should be abandoned because it is unlawful, inconsistent with NARA's past statements, and contrary to sound archival practices. Many of the electronic records covered by the proposed schedule have sufficient administrative, legal, research, and historical value to warrant their preservation in electronic form, even if the agency has converted the information to hard copy. Consequently, the law prohibits NARA from giving a general, government-wide mandate to destroy such records. Rather than addressing these records in the wholesale manner proposed in this Schedule, NARA must appraise the value of the records stored on particular systems to evaluate their proper disposition in light of the importance of the information they contain and the unique value of the electronic format so that, where appropriate, the electronic records are preserved in their electronic format.
Before discussing the defects of the proposed Schedule in more detail, three general observations concerning NARA's decision to propose this revision at this time are appropriate. First, the proposed Schedule is a transparent effort to revive an unsuccessful claim that the agency first articulated as its litigation position in Armstrong v. Executive Office of the President, C.A. No. 89-142 (D.D.C.). In that action, the plaintiffs, represented by Public Citizen Litigation Group, challenged the Archivist's failure to take steps to ensure that records stored on the electronic mail systems used by the Executive Office of the President were managed and preserved in accordance with the Federal Records Act. The Archivist defended his failure to take action to preserve these records by arguing, among other things, that the General Records Schedules authorized agencies to routinely destroy electronic mail and word processing records after the information on the records was converted to hard copy.See footnote 1 The courts rejected the Archivist's arguments and held that his failure to take action to preserve the Executive Office of the President's word processing, and electronic mail records, even if they had been printed, violated the Federal Records Act. Armstrong v. Executive Office of the President, 810 F. Supp. 335, 342 (D.D.C. 1993), aff'd in relevant part, 1 F.3d 1274 (D.C. Cir. 1993). Rather than comply with the courts' decisions, the proposed Schedule suggests that NARA is seeking to circumvent the courts' mandate by returning to its discredited litigating position.
Second, NARA's proposal is particularly disturbing in light of the efforts of the Administration to promote the use of the very electronic technologies covered by the proposal. Electronic mail, word processing and other computer systems are already widely used to create federal records, and the Administration is encouraging the expanded use of such technologies through its "Reinventing Government" initiative and other programs. NARA has been an active participant in efforts to encourage the use of electronic mail and other communications technologies to improve the efficiency of government. See Report of the Electronic Mail Task Force, Governmentwide Electronic Mail for the Federal Government, April 1, 1994. Indeed, the growing use of such technologies is reflected in NARA's decision to encourage the public to send the agency comments on this proposal to revise General Records Schedule 20 by electronic mail. The expanded use of such technologies only increases the importance of proper management and appraisal of the federal records stored on these systems.
Third, this proposed Schedule is inconsistent with NARA's stated commitment to promote proper management of records created by such electronic technologies. In particular, the proposed Schedule is incompatible with NARA's proposed regulations on Electronic Mail Systems. 59 Fed. Reg. 13906-10 (March 24, 1994). In those proposed regulations, NARA acknowledged the importance of electronic mail records and the need to appraise whether the records of particular agencies were sufficiently valuable to warrant permanent preservation in electronic format. Id. at 13906. Moreover, the regulations recognized that the disposition of different record formats (paper, magnetic, microform, etc.) must be considered separately because "[i]nformation in electronic records may have greater research utility than similar information stored on another medium because it is easier to access and manipulate." Id. at 13909-10. If approved, however, the proposed General Records Schedule would severely undercut, if not wholly undermine, the proposed Electronic Mail Regulations.
Our specific comments below address (i) the proposed Schedule's provisions on destroying the electronic version of electronic records where a hard copy exists; (ii) the proposed Schedule's failure to ensure that all record data is preserved; and (iii) the proposed Schedule's statement that system backups are not records.
I. THE SCHEDULE IMPROPERLY PROVIDES GENERAL AUTHORIZATION, TO DESTROY THE ELECTRONIC VERSION OF RECORDS IF A HARDCOPY EXISTS.
Items 13, 14, and 15 of the proposed Schedule, if approved, would mandate that agencies delete the electronic version of the following types of records when "no longer needed" by the agency, if the records have been copied to paper, microform, or an electronic records system for recordkeeping purposes:
Word Processing Files
Documents such as letters, memoranda, reports, handbooks, directives, and manuals recorded on electronic media such as hard disks or floppy diskettes.
Electronic mail records.
Senders' and recipients' copies of messages, regardless of length or substance, that meet the definition of Federal records created using the electronic mail (e-mail) system and any attachments.
Spreadsheets that are recorded on electronic media such as hard disks or floppy diskettes.
Item 3.b(1) similarly provides that agencies shall delete certain database records when "no longer needed," when "hard copy records are retained to meet recordkeeping requirements."See footnote 2 Thus, each of these items provides that the electronic version of the records is to be deleted when no longer needed if the records have been copied to a "hard copy" format (paper or microfilm).
Two considerations show that this proposal is improper. First, the electronic records covered by the proposed Schedule do not have the type of common characteristics that would allow the Archivist to authorize their disposition through a general schedule. General Records Schedules are designed to cover records that have the same value, regardless of the agency that creates them. Consequently, such schedules may be used for "administrative records" but cannot be applied to record systems that contain records documenting the unique, substantive functions for which an agency is responsible. See NARA, Disposition of Federal Records, 1992 ed., Appraisal Guidelines for Permanent Records, at IV-9, 10; NARA, Disposition of Federal Records, 1981 ed., at 25-26. The proposed Schedule, however, covers all word processing, electronic mail, and spreadsheet records -- regardless of their content, or their importance to documenting particular functions of the agency that created them. Indeed, only Item 3 of the proposed revisions is restricted to administrative records.
Such records do not have "common" characteristics that would allow the Archivist to determine that all the records covered by the Schedule lack any value that would warrant their further preservation by the government. The electronic correspondence, memoranda, spreadsheet and other records of agencies with high-level policy making responsibilities are far more likely to have legal, research, and historical value, even after they are "no longer needed" by the agency, than the records of agencies without such responsibilities. Under the proposed Schedule, however, electronic records containing information on the substantive activities of agencies such as the National Security Council, the Office of Management and Budget, or the Office of the Secretary of Defense, are treated in the same way as records of the Small Business Administration and the Rural Electrification Administration. NARA's own appraisal guidelines recognize that the historical value of agency correspondence and memoranda depend on the nature and function of the agency, and that, regardless of whether the records are "in paper or electronic form," records of higher level agencies and officials may qualify as permanent records. NARA, Disposition of Federal Records, 1992 ed., Appraisal Guidelines for Permanent Records, C-3.
Second, the proposed Schedule ignores the unique characteristics of electronic records. Some of the records covered by the proposed Schedule contain information of permanent historical significance. By providing that the electronic version can be deleted if copied to paper, microform or an electronic records system, the Schedule assumes that, as long as the information is preserved in some format, NARA can leave it to the agencies to decide whether the electronic version will be preserved.
It is clear, however, that the Archivist cannot be indifferent to whether the special attributes of the electronic version, which is often the original format used by the agency, are preserved when making appraisal decisions. As NARA's own guidelines acknowledge, records in electronic form can be of particular value to researchers and historians, even if the agency concludes that the electronic version is "no longer needed" for its purposes. Electronic records can be searched in ways that hard copy records cannot. Electronic records can be used to demonstrate patterns and connections among records that it would be difficult or impossible to discover using paper or microfiche. Electronic records can also be retrieved and made accessible to multiple users in ways that hard copy records cannot.
Indeed, the special value of having information in electronic format is underscored by the fact that agencies themselves have instituted programs to convert paper records to machine readable formats so that they can take advantage of the special search, storage, and retrieval capabilities of the electronic format. See, e.g., Tab 1, The "Paperless Office": A Case Study of the State Department's Foreign Affairs Information System, 45 Amer. Archivist 142 (1982); see also Office of Management and Budget, Circular A-130, 59 Fed. Reg. 37910, Sec. 7.p (July 25, 1994) (availability of information in diverse media, including electronic formats, permits agencies and the public greater flexibility in using the information). In appraising textual records in electronic format, NARA has found that, even if the records also exist in a paper or microfilm format, there still is "compelling reason to preserve permanently" the electronic version of the records because of the special capability of the computer to sort, retrieve, search, and display the records. Tab 2, NARA Appraisal Evaluation of Foreign Affairs Information System; see also NARA, Disposition of Federal Records, 1981 ed., at 18-19; NARA, Managing Electronic Records, at 22 (1990); NARA, Disposition of Federal Records, 1992 ed., V-7, 8; General Records Schedule 20 (1982 ed.).
These considerations apply to word processing and electronic mail records, as several examples illustrate:
* From 1987 through 1992, backup tapes containing word processing and electronic mail records created by the National Security Council during the Reagan Administration in electronic format were repeatedly used by the Independent Counsel and congressional committees investigating the Iran-Contra Affair. These electronic records were also used by the Justice Department to aid in the prosecution of Manuel Noreiga, and by the National Security Council to provide information for the confirmation hearings of the Bush Administration's CIA Director, Robert Gates. To recover records from these tapes, the agencies conducted computerized searches of the data that would not have been possible if the records were not in electronic form. See Tab 3, National Security Council Memorandum describing keyword searches of tapes, Dec. 3, 1992; Tab 4, Summary of Previous Searches of the Preserved NSC Backup Tapes ("Appendix II" in Armstrong v. Executive Office of the President, No. 89-142, litigation).See footnote 3
*In order to make records from the word processing and electronic mail records of the Executive Office of the President available to the National Security Archive under the Freedom of Information Act, 5 U.S.C. Sec. 552, in Armstrong v. Executive Office of the President, C.A. No. 89-142, NARA, the Office of Administration, and the National Security Council printed and released the indices of word processing documents stored on the agencies' magnetic tapes and the logs of electronic mail correspondence for researchers to identify the records that were of greatest interest. See, e.g., Tab 5 (excerpts from Executive Office of the President word processing document inventory); Tab 6 (excerpt from logs of electronic mail). These indices are the type of "finding aids" that NARA admonishes agencies to retain because they are important to researchers trying to analyze to a large body of records. However, if the agencies had deleted the electronic version of these records after the records were converted to hard copy, as the proposed Schedule provides, it would not have been possible to generate these indices and logs.
- The indices and logs of the Executive Office of the President word processing and electronic mail records released in the Armstrong litigation contained over 40,000 entries. In order to analyze these entries and identify the records for which the public interest in prompt release appears to be the greatest, the National Security Archive converted the information back into an electronic format to create an electronic database that allows researchers to identify duplicate entries, and to arrange the data chronologically, or by sender, recipient, or topic. If the proposed Schedule were approved, however, it would impair the ability to perform such analyses by authorizing routine destruction of the electronic data from which such databases can most easily be created.
- The Canadian Archives has processed the electronic records of the Trade Negotiations Office, which consists primarily of word processing and electronic mail records stored on a micro computer network. Although many of the records had been copied to paper and preserved elsewhere, the Canadian Archives found that copying and preserving the electronic records was also warranted because the electronic version provided capabilities for full text researching and other enhanced access that is not possible with paper records. See Tab 7, Acquiring Records of TNO, Nat. Arch. of Canada Bull., Vol. 7 (1991).
- The Navy Laboratories History Program's pilot project on preserving electronic mail records compared preservation of electronic mail records in electronic format, paper, and microfiche, and concluded the electronic format had substantial advantages over the other formats because of the enhanced ability to access and store the records. See Tab 8, C. Nowicke, Managing Tomorrow's Records Today: An Experiment in Archival Preservation of Electronic Mail, 13 Midwest Archivist 67 (1988).
These considerations certainly do not mean that all word processing and electronic mail records must be permanently preserved in electronic format. They demonstrate, however, that whether the electronic format warrants preservation is an important factor in a proper appraisal of such records by NARA, and it is unlawful for NARA to abdicate this responsibility by authorizing agencies to delete the electronic version wherever they have created a copy on paper or microfiche. The Archivist cannot lawfully make a blanket determination that the electronic version of all the electronic mail, word processing, and spreadsheet records covered by the proposed Schedule do not have sufficient value to warrant their continued preservation in electronic form.
See footnote 4
II. THE PROPOSED SCHEDULE FAILS TO ENSURE THE PRESERVATION OF ALL INFORMATION IN THE ELECTRONIC RECORDS.
The proposed Schedule is also contrary to the Federal Records Act because it fails to ensure that all the information in electronic records is preserved. Electronic records often contain critical information that is not reproduced when the records are copied. In Armstrong, the court squarely rejected the Archivist's claim that neither the agencies nor the Archivist had an obligation to manage such information in accordance with the Federal Records Act. Armstrong v. Executive Office of the President, 1 F.3d 1274, 1284-87 (D.C. Cir. 1993). The proposed schedule ignores this holding in two respects.
First, under the proposed Schedule, critical information stored in electronic word processing files or spreadsheets which does not appear in the printed version of the electronic record would be lost. This defect is particularly glaring with respect to spreadsheets. The "hard copy" produced from an electronic spreadsheet typically only contains the results of the spreadsheet computations, and does not contain the information on what calculations are performed to obtain these results. See, e.g., Tab 9 (contrasting standard print-out and print-out showing cell formulas in Lotus 1-2-3 spreadsheet). Under the proposed Schedule, however, agencies would be permitted to destroy the electronic record which shows what computations are performed by the spreadsheet when it was no longer needed to produce a hard copy. In instances where the spreadsheet concerns important budgetary, funding, or other analyses, the information on the computations may warrant preservation for a longer period, or even preservation as a permanent record.
Second, although the introduction to the proposed Schedule contains an admonition that agencies must preserve information beyond the text of the electronic mail records, the instruction is inaccurate. The proposed schedule states:
E-mail records must include the identity of the sender and the recipient(s), and the date, and any other data deemed by the agency to be necessary to understand the contents or to interpret the structure of the e-mail record. Receipt data that the agency determines is needed for complete and accurate documentation must also be maintained with e-mail messages that are Federal records.
This instruction is contrary to the Federal Records Act because the Act does not limit the definition of what constitutes a record to data "deemed by the agency to be necessary to understand the contents" of the e-mail record, or to data "that the agency determines is needed for complete and accurate documentation." Agencies plainly have an obligation to preserve such information, but are not free to discard other data recorded in the record merely because they do not "deem" it necessary, or do not believe that it "is needed for complete and accurate documentation." See Armstrong v. Executive Office of the President, 1 F.3d at 1286-87. Nor can the Archivist, through a General Records Schedule, make a determination that all such information lacks sufficient value to warrant its retention, particularly since the nature of the additional information recorded in the electronic version is unspecified.
Item 8 of the current General Records Schedule 20 recognizes that a security backup file, defined as the "[e]lectronic copy consisting of data identical in format to the master copy of an electronic record or file and retained in case the master file or database is damaged or inadvertently erased," is a record, and provides that such backup file records may be deleted when they are no longer needed because (i) the master records for which this file provides a security backup have been transferred to NARA or deleted pursuant to an authorized disposition schedule; or (ii) the backup file is replaced by a subsequent security backup file.
The proposed Schedule would revise this provision so that a backup file would only qualify as a security backup file if it consists of data identical in "logical" format to the master copy. The introduction to the proposed Schedule states that this revision of Item 8 is being made because NARA is proposing to reinterpret the statutory term "record," 44 U.S.C. Sec. 3301, so that most of the backup files previously classified as "security backup file" records will no longer be considered records:
Security backup copies of files as defined in item 8 of this schedule do not equate to system backups. System backups are copies on off-line storage media of software and data stored on direct access storage devices in a computer system. They are used to recreate a system or its data in case of unintentional loss from on-line storage. System backups are merely mirror images of storage disks on which data and documents may be scattered randomly as they are on the disks themselves. Therefore, system backups are not records and are suitable for safeguarding against the loss of records only when the record copy is maintained elsewhere.
This provision is contrary to law and irrational. Backup copy files that are created in order to be able to restore lost records qualify as "records" under the statute, regardless of whether they store data in "logical format" or in the format used by the on-line system, which NARA says "may be" random. Moreover, there is no reason to distinguish between backup copies in "logical" format and those in another format, in establishing an appropriate disposition schedule. Because the purpose of such files, regardless of their format, is to permit recovery of an electronic record or file if the master file is damaged or erased, these files should be subject to a disposition schedule that ensures that they are retained by agencies to serve that purpose. The current language of General Records Schedule 23 concerning security backup files is appropriate and should not be changed. CONCLUSION
NARA should not adopt the revisions to General Records Schedule 20 proposed in its October 17, 1994 notice.
November 30, 1994
Footnote: 1 Defendants' Reply to Plaintiffs' Opposition to Defendants' Motion to Dismiss or Alternatively for Summary Judgment at 12 n.12 (March 9, 1989).
Footnote: 2In addition to the problems noted above, the language of Item 3 is, at best, imprecise on what categories of records are covered and when disposal is authorized. Rather than "clarify the coverage of this item," the proposed language is more confusing than the language currently used in item 3 of Schedule 23 and appears to expand the records authorized for destruction since it is no longer limited to databases, but covers any electronic records.
Footnote: 3 The exhibits to this document are not included here but, like other documents cited in these comments, are already available to NARA as part of the record in the Armstrong v. Executive Office of the President litigation.
Footnote: 4 We also note that, because the proposed revisions to the Schedule authorize agencies to delete records "when no longer needed," they are inconsistent with the statute's mandate that General Records Schedules authorize disposal "after the lapse of specified periods of time." 44 U.S.C. ? 3303a(d). This is a defect which is also present in many of the current Schedules. The significance of the failure to specify a time period, however, is magnified by the breadth of proposed Schedule provisions presented here.