Comments by the Society of American Archivists on NARA's "Proposal for a Redesign of Federal Records Management"

Prepared for SAA by its Electronic Records Section, and endorsed by its Government Records Section and SAA Council

11 September 2002

Executive Summary

SAA would like to commend NARA for the considerable research and analysis represented in this document. Realistic recordkeeping strategies require an understanding of the current practices and constraints faced by government agencies. NARA openly recognizes that many aspects of the current recordkeeping regime are not working as intended. In order to improve this situation, NARA must present procedures, services, tools and facilities to agencies that are perceived as adding value, rather than simply imposing additional administrative burdens.

Given NARA's limited resources and the vast number and heterogeneity of recordkeeping systems in the federal government, we agree that it is essential to "empower agencies to determine for themselves how best to manage their records," but we would also like to emphasize the importance of the qualifier that NARA includes with this statement: "within general parameters provided by NARA." While NARA cannot take part directly in the design, procurement and implementation of every new electronic information system in the federal government, it can develop, disseminate, maintain, and actively promote detailed guidance on how best to build recordkeeping requirements into new systems. Consistent with NARA's proposal to move away from a narrow focus on the distinction between records and nonrecords, such guidance could adopt language that is more inclusive of the considerations involved in developing and maintaining trustworthy information systems. In addition to some degree of direct involvement by NARA staff, this may be an area in which NARA could apply its "training, certification, and monitoring program for contractor staff who offer technical assistance."

We would warn NARA against the concept of "pre-accessioning" electronic records in the federal records center (FRC), and instead treat the FRC service more like full-fledged accessioning. Apart from the important stipulation of leaving the reference load and Freedom of Information Act (FOIA) fulfillment responsibilities with the creating agencies, it is not clear what it would mean for NARA to offer an electronic records FRC as a means to "take on the preservation responsibilities for the agencies prior to legal accessioning." Rather than seeing ingest into the Electronic Records Archive (ERA) as a separate and distinct process, NARA should address the ERA's requirements with any transfer of electronic records to NARA, even if those records are not considered permanent.

We would recommend that NARA further clarify several issues. First, it is not clear what sort of retention period might qualify records as candidates for transfer to the FRC. We specifically recommend that NARA consider and explicitly decide whether or not the FRC will provide preservation services for records with temporary but not permanent value. Second, we raise some questions related to the details of FRC transfer procedures. Third, we would find it helpful for NARA to further elaborate on how agency and NARA responsibilities for FRC services would be identified and assigned. Fourth, and perhaps most importantly, we raise some cautionary notes about the large financial and technical implications of offering a free FRC for electronic records. Finally, we suggest a reconsideration or refinement of the term "continuing value (CV) records."

Most of all, we applaud NARA for looking with fresh eyes at the issues associated with modern records management. We hope that this redesign of records management can serve as guidance not only for NARA and federal agencies but also to many others struggling with an increasingly complex records universe with very limited resources. Many of the ideas expressed in the paper need to be tested to see if they can actually improve the records management process, and we hope that NARA will continue to actively test and refine its thinking on these important issues.

1. Introduction

In response to the request for comments on the National Archives and Records Administration's "Proposal for A Redesign of Federal Records Management" available at <http://www.archives.gov/records_management/initiatives/rm_redesign_project.html>, the Society of American Archivists (SAA) submits the following document [1]. SAA is the oldest and largest association of professional archivists in North America. Representing more than 3,000 individuals and 400 institutions, the SAA is the authoritative voice in the United States on issues that affect the identification, preservation, and use of historical records.

The SAA commends NARA on its recent and continuing efforts to face the daunting challenge of managing and preserving the records of the federal government. NARA's call for comments on the "Proposal for A Redesign of Federal Records Management" represents not only a bold move in taking leadership on these issues but also a willingness to engage with the larger archival community in their discussions. We appreciate the opportunity to comment on such a groundbreaking effort.

2. Comments on the Records Management Background and Challenge

SAA would like to commend NARA for the considerable research and analysis represented in this document. Realistic recordkeeping strategies require an understanding of the current practices and constraints faced by government agencies. NARA openly recognizes that many aspects of the current recordkeeping regime are not working as intended. According to the "Proposal for a Redesign," "a vast majority of retention schedules are out of date, inaccurate, or nonexistent. A large majority of electronic records series of continuing value are not coming into archival custody." In order to improve this situation, NARA must present procedures, services, tools and facilities to agencies that are perceived as adding value, rather than simply imposing additional administrative burdens.

3. Comments on Breaking Down the Barriers

Given NARA's limited resources and the vast number and heterogeneity of recordkeeping systems in the federal government, we agree that it is essential to "empower agencies to determine for themselves how best to manage their records," but we would also like to emphasize the importance of the qualifier that NARA includes with this statement: "within general parameters provided by NARA."

3.1 Addressing Recordkeeping Systems, Not Just Records

One proposed change that the SAA can heartily endorse is the decision to move away from distinctions between "record" and "nonrecord" in records management systems. In archival theory, the definition of what is a record is of great importance, as Mark Greene has recently argued in his article "The Power of Meaning: The Archival Mission in the Postmodern Age."[2] From a practical records management perspective, however, the distinction between records and nonrecords is often more confusing than it is enlightening. NARA proposes quite properly to change its focus to record keeping systems, rather than individual records.

Consistent with NARA's proposal to move away from a narrow focus on the distinction between records and nonrecords, such guidance could adopt language that is more inclusive of the considerations involved in developing and maintaining trustworthy information systems. A long-standing maxim of electronic records work is that archivists and records managers must get involved as early in the record life cycle (or continuum) as possible. Ideally, this should involve active guidance on the design of systems before any records have even been created. While NARA cannot take part directly in the design, procurement and implementation of every new electronic information system in the federal government, it can develop, disseminate, maintain, and actively promote detailed guidance on how best to build recordkeeping requirements into new systems [3].

We believe that it is important to distinguish between two areas of consideration:

1. NARA's priorities for focusing its own resources.
2. The content of services, assistance, policies, guidance and tools that NARA offers to agencies.

For the first, we agree with NARA's suggestion in the document that it should focus more precisely on areas of high accountability, rights, risk, and preservation value. This should ideally be based on what NARA knows or can discover about the operations, business and functions of the agencies and the federal government more generally. How such an approach would work in practice is one of the great unanswered questions, and we therefore support NARA's efforts to investigate further, ideally through pilot projects, how exactly a more high-level functional approach to records management would work.

As regards the services, assistance, policies, and guidance that NARA can offer to agencies, we would contend that much of the focus should be on systems. Agencies tend to plan, design, manage and migrate data at the system level. As NARA points out this document, agencies are often either unable or unwilling to prioritize their own efforts in terms of a strict definition of the word "records." From the perspective of agencies, the implementation of recordkeeping requirements makes the most sense if it can be framed in terms of designing, developing, managing and using the agencies systems. While the ultimate goals of NARA are to ensure the capture, management, preservation and access to the essential evidence that the systems contain, these larger goals can best be achieved through language and methods that address the specific system characteristics that agencies confront on a daily basis.

In order for such guidance to serve its intended purpose, it should be advocated and supported by individuals with sufficient expertise in both records management and the design and implementation of information technology. Since such human resources are limited, involvement in system design and management could conform to the same prioritization and graduated approach that NARA is proposing more broadly in this document. The most attention could be paid to government functions and systems that involve the most rights and accountability, continuing value and risk. It is important, however, not to frame these priorities primarily in terms of the records that meet these criteria. In order for NARA's strategy to be scalable, it should attempt to address the most important recordkeeping systems before the records themselves have been created. Contributions could include, among others:

• model language for requests for proposals (RFPs),
• suggestions for the purchasing and bidding process,
• tools for introducing recordkeeping into an agency's system development life cycle,
• guidance and training on addressing recordkeeping in standard project management methodologies.

In addition to some degree of direct involvement by NARA staff, this may be an area in which NARA could apply its "training, certification, and monitoring program for contractor staff who offer technical assistance."

These concerns are particularly important if NARA is to take the more active role in the custody of electronic records that it proposes. According to the "Reference Model for an Open Archival Information System (OAIS)," [4] the "ingest" of digital objects into an archival repository requires that the digital objects follow certain submission information package (SIP) conventions. An OAIS ingest scenario is much more feasible if NARA has provided agencies with specific guidance on the forms and formats of digital objects.

3.2 Rethinking the Concept of a Federal Records Center for Electronic Records

NARA should establish clear conventions for the submission of digital objects, not only to the Archives for long-term preservation, but also to its proposed electronic federal records centers (FRC). In an environment of paper records, the FRC plays a dual role: (1) providing agencies with intellectual and physical control of their semi-active records, and (2) serving as a point of transition from agency custody to NARA custody of records. For electronic records, both of these roles require that records conform to certain format, encoding and metadata requirements before they are transferred to the FRC.

It is for this reason that we would warn NARA against the concept of "pre-accessioning" electronic records in the FRC, and instead treat the FRC service more like full-fledged accessioning. Apart from the important stipulation of leaving the reference load and Freedom of Information Act (FOIA) fulfillment responsibilities with the creating agencies, it is not clear what it would mean for NARA to offer an electronic records FRC as a means to "take on the preservation responsibilities for the agencies prior to legal accessioning." When NARA accepts paper records into the FRC, it takes certain measures and issues certain guarantees about the physical protection that those records will receive. When accepting electronic records into the FRC, these measures and guarantees become much more involved and resource-intensive (persistent naming, media refreshment, periodic data transformations, etc.), in order to ensure the accessibility and understandability of those records. Once such measures and guarantees are made, it would seem that the records have then been ingested by NARA into its Open Archival Information System (OAIS) [5]. Thus, rather than seeing ingest into the Electronic Records Archive (ERA) as a separate and distinct process, NARA should address the ERA's requirements with any transfer of electronic records to NARA, even if those records are not considered permanent.

With paper records, transfer from agency to FRC to Archives involves the movement of one physical object. In a digital environment, one can store any number of copies of the same object in multiple locations at very little cost. Disposition of records can, therefore, be treated quite differently. It is not necessary to wait until the end of a retention period for a disposition act of "transfer to Archives" to be triggered. If the long-term preservation value of given records is already known, copies of those records can be transferred to the Archives any time after they have been created. During their period of active use within the agency, it is quite likely that the agency would continue to rely on its own internal copies of records that reside in its active systems, rather than relying on the copies [6] that are being maintained within the FRC. The difference between the copies in the agency and the copies in the FRC would be that the former are being managed in a way that optimizes their short-term use, while the latter are being managed in a way that ensures their ongoing reliability, authenticity, integrity and usability to future users both within and outside of the creating agency.

While we agree with NARA that the descriptive metadata created for long-term records should be more detailed than that for short-term records, there are many recordkeeping, administrative and preservation metadata elements that should be created or captured for all records within the FRC. Earlier in the document, NARA states that "many electronic records are being kept as backup tapes in tape libraries until systems are replaced or shut down. At that point, if not sooner, records of continuing value are likely to be lost." It is important that NARA's FRC for electronic records not simply replicate such libraries of backup tapes. In order for the FRC to serve what NARA states as its goal of "preservation work," it must make a significant commitment of resources to the proper ingest, management and dissemination of digital objects. Otherwise, the FRC will simply serve as a "bit repository" with no guarantee of appropriate management, preservation or access. It is possible that NARA could decide to offer only bit-level preservation services itself, while partnering with third parties to provide the other components of the repository. Such an arrangement, however, would take a great deal of explicit planning and agreement. It would also be important for NARA to identify and manage the risks associated with such a dependence on third parties.

3.3 Further Elaboration of FRC Functions, Responsibilities and Costs

We would recommend that NARA further clarify several issues related to the electronic records FRC. First, it is not clear what sort of retention period might qualify records as candidates for transfer to the FRC. Most of the electronic records that are at risk in government are those with a moderate-term retention (more than a year but less than permanent). If it were to live up to the standard definition of records center, then those moderate-term records should be transferred to the FRC in order to save them from loss within the agency. If agency systems are updated every 3-5 years, the records within them can be at risk. Apart from obsolescence, deletion and migration issues, mismanagement within the agency is also likely to become an issue once records have existed for some time within the agency. Permanent records are obviously very important, but they also represent only a tiny portion of the records universe. In order to add value to agencies in the sort of "virtuous circles" that the document describes, we recommend that NARA consider and explicitly decide whether or not the FRC will provide preservation services for records with temporary but not permanent value.

Another issue related to the FRC is that of transfer procedures. How soon after creation should records be transferred to the records center? 2 seconds? 6 months? In batches?

Just-in-time? Since there can be any number of copies of the same records, transfer could be implemented in many different ways. The transfer process should also include provisions for avoiding widespread duplication. If NARA were to ingest copies of records from active agency systems, future ingests from those same systems could result in a great deal of redundancy. This would seem to lend additional support to the argument that NARA should, to the extent possible, assist agencies in ensuring that their internal recordkeeping systems are appropriately designed, implemented, managed and maintained. In order to address some of these issues, perhaps NARA could identify several different transfer scenarios, depending on the needs of the agency and the recordkeeping context.

We would also find it helpful for NARA to further elaborate on how agency and NARA responsibilities for FRC services would be identified and assigned. Provisions should include not only technical submission agreements [7] but also legal issues such as who would be responsible for satisfying legal discovery motions and, as mentioned previously, how responsibility for reference and Freedom of Information Act (FOIA) duties would be fulfill and possibly transferred over time. This may be an area in which the traditional conception of "official records" needs further refinement. Since exact copies of electronic records are both feasible and likely, it could be case that the agency copy will be considered the official records for purposes of FOIA but that the NARA copy is considered the archival copy, i.e. the master from which preservation efforts and long-term access will derive.

The final area, and perhaps the most important, is cost. Throughout this document, NARA is careful to emphasize that its resources are quite limited. We would caution NARA about the large financial and technical implications of offering a free FRC for electronic records. If a free records center for paper records is no longer a sustainable proposition for NARA in the current political context, then it would seem that the same would be true for electronic records. Perhaps NARA should consider making either some or all of its electronic FRC services fee-based or tied to some sort of in-kind contribution from agencies. Ultimately, we feel the federal government should be providing NARA with considerably more funding to fulfill these essential government services, and we see future advocacy of this position as vital to NARA's mission. Even if more funding is allocated to NARA in the future, however, there will always be limits to its resources, and we recommend careful consideration of the sustainability of any services that might be offered in this area.

3.4 Clarification of Continuing Value

If NARA is going to offer FRC services for electronic records, we would also suggest a reconsideration or refinement of the term "continuing value (CV) records." Strictly speaking, any records that are still being purposely (as opposed to accidentally) retained—either for a limited period as per a retention schedule or for an indefinite time in the Archives—have been deemed to have continuing value. As NARA is using the term in its discussion of resource allocation strategies and agency profiles (e.g. Figure 6), it would seem that it is roughly synonymous with "records with long-term retention value." This seems to contradict NARA's claim elsewhere that "the term 'records of continuing value' is used interchangeably with 'records documenting the national experience' and 'archival records.'" In other parts of the document, the "national experience" is listed as a consideration separate from business needs, rights and accountability. Does this mean that CV is assessed on the basis of considerations apart from business needs, rights and accountability?

Should one assume that all the records that NARA designates to be CV are to be retained indefinitely, or are there CV records that have limited retention periods? Is there a general range of retention periods that would qualify as CV, e.g. more than 25 years? If identifying records as CV is about more than simply their period of retention, what other factors would be considered?

4. Comments on New Approaches to Scheduling and Appraising Records

As stated above, we laud NARA for its efforts in this area and its informed recognition that "the current approach is not working." We agree that a higher-level and more functional approach to the appraisal and scheduling of federal records may make more effective use of NARA's limited resources, and is certainly worth exploring further. We would suggest that this Appendix might be enhanced by providing some examples, particularly those that illustrate how the proposed retention standards would differ from existing general retention schedule (GRS) or agency retention schedule entries. It is also not clear what meaning and rationale are behind the following statement: "Keep GRS 20 items 13 and 14 but continue to add these items in at a high level to agency specific schedules." Perhaps NARA could elaborate further on this point.

5. Conclusion

SAA lauds NARA's continuing efforts in redefining its approach to federal recordkeeping. Historically, many other archival institutions have looked to NARA as the model for how to approach records management and preservation. "Proposal for a Redesign of Federal Records Management" represents a bold new step in this long tradition. In addition to NARA's vital work in contributing to the development of standards, policy, and legislation to bolster the greater archival mission, we hope that this redesign of records management can serve as guidance not only for NARA and federal agencies but also to many others struggling with an increasingly complex records universe with very limited resources. While we encourage and hope for further clarification on several issues raised in NARA's document, we see it as a promising step in the right direction, and eagerly await the results of NARA's future investigations on the topic.

End Notes:

[1] The Electronic Records Section, a constituent unit of the SAA, is responsible for most of the comments included here. The following individuals contributed texts and comments during the process of drafting this document: Charles Arp, Ohio Historical Society; Rosemary Pleva Flynn, University of North Dakota; Robert Horton, Minnesota Historical Society; Geof Huth, New York State Archives; Cal Lee, University of Michigan. The Electronic Records Section includes members who are employees of the National Archives and Records Administration. Those members have abstained from the formal process of endorsing this document.

[2] Mark A. Greene, "The Power of Meaning: The Archival Mission in the Postmodern Age," American Archivist 65:1 (Spring/Summer 2002)

[3] For one excellent example of such guidance, see Shawn P. Rounds and Mary P. Klauda, "Trustworthy Information Systems Handbook," Saint Paul, Minnesota: State Archives Department, Minnesota Historical Society, Version 4, July 2002, http://www.mnhs.org/preserve/records/tis/tis.html.

[4] "Reference Model for an Open Archival Information System (OAIS)," Washington, DC: Consultative Committee for Space Data Systems, 2002, http://www.ccsds.org/documents/pdf/CCSDS-650.0-B-1.pdf.

[5] According to the "Reference Model for an Open Archival Information System," an OAIS is "An archive, consisting of an organization of people and systems, that has accepted the responsibility to preserve information and make it available for a Designated Community" (p. 1-11).

[6] It is our assumption that the FRC would rely on redundant storage.

[7] According to the "Reference Model for an Open Archival Information System," a submission agreement is the "agreement reached between an OAIS and the Producer that specifies a data model for the Data Submission Session. This data model identifies format/contents and the logical constructs used by the Producer and how they are represented on each media delivery or in a telecommunication session" (p.22).