2008 EMMETT LEAHY AWARD WINNER, KENNETH THIBODEAU

As Director of the Electronic Records Archives Program of the National Archives and Records Administration, Dr. Thibodeau defined an extraordinary vision of the management and preservation of electronic records, overseeing the development of an unprecedented system that enabled the transfer, ingest, establishment of item-level control, and faceted searching of more than 200,000,000 electronic records from the George W. Bush White House. He has had major leadership responsibilities within the archival community, both nationally and internationally. His written publications and presentations on the preservation of electronic records have had a profound effect on the records and information profession and have raised this awareness in other disciplines, including library and information science, museum studies, computer science, electronic engineering, statistics, natural science, and the humanities. He has played a leading role in the development of a variety of standards, including the Department of Defense Standard for Records Management Applications, the ISO Open Archival Information System Reference Model, and the Object Management Group’s Records Management Services Specification. 

Balancing Archival Practice and Information Technology.  

Keynote Address. International Council on Archives.  Section on University and Research Institution Archives.  Annual Conference.  2013.

Wrestling With Shape-Shifters: Perspectives on Digital Preservation.  

Opening Keynote Address. UNESCO's Memory of the World Program.  The Memory of the World in the Digital Age: Digitization and Preservation.  2012. LINK

The  Perfect  Archival  Storm: The Transfer of  Electronic  Records from the  G.W.  Bush  White  House to the National  Archives of the United States.   

UNESCO's Memory of the World Program.  The Memory of the World in the Digital Age: Digitization and Preservation.  2012.  LINK

At the Institute of Certified Records Managers Annual Business Meeting, Las Vegas, Nevada October 19, 2008

Good evening. I am Charles Dollar, Chair of the 2008 Emmett Leahy Award Committee. 

For more than two decades, the Institute of Certified Records Managers has graciously included the presentation of the Emmett Leahy Award as part of its Annual Business Meeting. Through these years, the Committee has welcomed the opportunity to annually recognize a distinguished leader of our profession before this gathering of records management expertise and experience that the ICRM embodies. The Emmett Leahy Award Committee looks forward to continuing its association with the Institute of Certified Records Managers in 2009 and beyond. 

The Emmett Leahy Award Committee is made up of the previous ten recipients of the Emmett Leahy Award. Tonight we have five members of the 2008 Emmett Leahy Award Committee present: Christine Ardern, Eugenia Brumm, James Coulson, Charles Dollar, and John Philips. Mary Robeke, Luciana Duranti, John McDonald, Bruce Miller, and Anne Thurston could not be with us.      

As some of you already may know, in 2008 the Huron Consulting Group began underwriting the expenses associated with the Emmett Leahy Award. The 2008 Emmett Leahy Award Committee believes this support establishes a solid foundation for the Committee to continue recognizing individuals whose impact on the information and records management profession at the global level perpetuates the information and records management legacy of Emmett Leahy. 

In addition to selection of the recipient of the 2008 Emmett Leahy Award, the Committee undertook an initiative to promote a greater awareness of the Emmett Leahy Award. One feature of this initiative is the design of an Emmett Leahy Award Logo that is being incorporated into the Emmett Leahy Award web site. The logo, which is displayed on the brochure in your seat, captures the global dimensions of the Emmett Leahy Award. Another feature of this initiative is the placement of advertising in the Information Management Journal and distribution of the Emmett Leahy Award brochure, which I just mentioned. Please note that the brochure displays information about the 2009 Emmett Leahy Award, including February 1, 2009 as the deadline for submission of nominations. More information about the nomination process, award selection criteria, and other background information is available at leahyward.com. 

The first Emmett Leahy Award was made in 1967 and except for the years 1981-1984 it has been presented annually since then. We have in attendance tonight several previous recipients of the Emmett Leahy Award: Bill Benedon, Fred Diers, Mark Langemo, Don Skuptsky, and Bob Williams. Incidentally, Bill Benedon is the 1968 recipient of the Emmitt Leahy Award. 

As in previous years, the Emmett Leahy Award Committee continues its existence as an independent entity and the selection of the Emmett Leahy Award recipient is the exclusive responsibility of the Committee. 

It is now my pleasure to turn to the important business at hand, the presentation of the 2008 Leahy Award.

Named after Emmet J. Leahy, the Emmett Leahy Award is presented annually to recognize an individual whose contributions and outstanding accomplishments have had a significant impact on the records and information management professions. It is differentiated from other awards in the following ways:

•  Membership in a professional organization such as ARMA, the SAA, AIIM, or the ICA, among others, is meritorious but it is not a threshold requirement.
•  Demonstration of service to a profession also is meritorious but it is not a differentiating requirement.
•  Authorship of professional and presentations is commendable but it is not a threshold requirement.
• Original contributions that fundamentally moved the information and records management discipline in a direction that it might not otherwise have gone is a threshold requirement.

The Emmett Leahy Award Committee selected Ken Thibodeau, Director of the Electronic Records Archives Program at the National Archives and Records Administration, as the 38th recipient of the Emmett Leahy Award because of his contributions to the preservation of electronic records. A member and Fellow of the Society of American Archivists. Thibodeau has had major leadership responsibilities within the archival community, both nationally and internationally, in addressing the management and preservation of electronic records. Over the last decade he has written more than ten seminal articles and essays and made scores of presentations on the management and preservation of electronic records in professional conferences in a variety of other disciplines, including library and information science, museum studies, computer science, electronic engineering, statistics, natural science, and the humanities. 

Ken Thibodeau had a major role in the development of DoD 5015.2 and conceptualization of ISO 14721, the Open Archives Information System. In conjunction with computer scientists and engineers at the San Diego Supercomputer Center he established a conceptual architecture for the persistent preservation of electronic records. This architecture included an archival system that is independent of its information technology infrastructure so that hardware or software components can be replaced with minimal impact on the system as a whole and negligible impact on the records preserved in it. All of this work culminated in the design, development, and successful roll out of the Electronic Records Archives (ERA) Program of the National Archives and Records Administration. The success of the initial deployment of ERA has enabled NARA to take on the challenge of ensuring the preservation of more than 100 Terabytes of electronic presidential records of George W. Bush. The Emmett Leahy Award Committee believes that Thibodeau's work has established a practical and technical basis for electronic records preservation that will persist for into the future.

The first thing that might come to mind, on hearing the topic of the survival of records in the twenty-first century, is the problem of digital preservation. Obviously, the preservation of records in digital form is crucial, but while preservation is necessary, it is not sufficient. Given the open-ended proliferation of new forms for creating, capturing, and combining information in digital systems, of ever wider, more diverse, and more highly specified ways of applying digital technology in the conduct of business, of the expanding capacity to re-use and even re-purpose digital data to satisfy a variety of both planned and spontaneous needs, it is also necessary that records continue to be recognized not only as a distinct class of information assets, but also as one which merits special attention. It is not sufficient for records keeping to be seen as a necessary part of doing business. Necessity of this sort is all too often the daughter of laws, regulations, and other external requirements. Records that are kept as the result of such external forces are easily relegated to the sidelines and perceived as having marginal value in the accomplishment of practical objectives, in strong contrast to the immediate value of real time online transactional processing, multidimensional analytic processing, geo-referencing, experiential computing, and other powerful tools made available by the application of computer science and engineering. If records slide to the margins of the conduct of business, records management will diminish with them. To survive and prosper in the twenty-first century, records and methods for managing them need to have a vibrant, organic relationship to the conduct of business. To be a vital contributor to corporate, institutional, individual and societal success in the twenty-first century, records management must deal cogently and comprehensively with the increasing permeation of digital information in the conduct of business.
    

Cyberspace is so different from what has gone before that it has been characterized as the fifth dimension, one which enables us to create and do things which are not possible in the space-time continuum. The creation, capture, and communication of information in cyberspace is drastically different than in four dimensions, because the digital dimension breaks the bounds of space and time that constrain the information technology of hard copy documents. To face the challenges and take advantage of the possibilities offered by digital technology, we need a richer and deeper understanding of the nature of digitally encoded information and how such information can be and can be managed as records. We need to be better able to apply the knowledge we have of records and records management in the digital realm; translating it in terms that make it effectively operative in cyberspace; adapting it where necessary; and also abandoning those concepts and techniques that are not viable in cyberspace. If we fail to do so, we run the risk of seeing records management become an increasingly esoteric exercise, divorced and isolated from the mainstream of affairs.
    

The first thing records managers need to do is to acknowledge that established knowledge and methods have limited applicability in the digital dimension. The second thing - logically but not necessarily chronologically - is to recognize the opportunities that digital technology creates for managing records in ways that might exceed by far anything possible in four dimensions. Consider just two basic topics: What are electronic records, and how should we organize them? For both topics, let us consider how well or how far established knowledge applies in the digital realm; some obvious limits to its applicability, and alternatives that the digital dimension opens up.
    

Conceptually, electronic records are not radically different than traditional records. Traditional definitions and concepts identify fundamental characteristics that are independent of how records are constructed, encoded or stored. A record

•     is a unit of information;
•     is made or acquired in the course of activity;
•     contains evidence or information about that activity;
•     is kept for use in subsequent activity or for reference, and
•     is related organically to other records of the same records creator and activity.

But problems arise when we try to deal with electronic records empirically. From an empirical perspective, we are not yet able to characterize them in a way that would enable us to articulate surefire methods and processes to maximize their value in the conduct of business. Records as we know them are artifacts of a particular genre of information technology, the technology of hard copy, where information is affixed to a physical medium in a hard and fast way. Digital encoding fundamentally alters the relationship between medium and message. In the digital realm for all practical purposes the relationship between medium and message is immaterial. Digital data move frequently and repeatedly from one physical medium to another: from silicon chips to magnetic drives, solid state memory, optical discs, copper cables, optical fibers, and electromagnetic waves, but the information objects the data constitute persist unchanged across such physical transformations.
    

The message here is not the medium, and it is certainly not about the durability of the medium. In the realm of hard copy, people could, did, and do use both permanent and short-lived media: from the clay tablets of the Babylonians to the wax tablets of the Romans, from cheap newsprint to archival quality paper. The same is true in the digital realm. There is no physical or chemical barrier to permanent digital media. You could write digital data on clay tablets. It’s just that the business case for doing so is lousy.
    

The basic difference in the relationship of electronic records to their physical carriers is but one of many ways that electronic records substantially differ from traditional records. Additional differences will emerge as digital technology evolves. Menne-Haritz suggests that the shift from 'written' to electronic communications is as epochal as the shift from oral to written communications in past millennia. Articulating a variation on the theme that form follows function, Menne-Haritz points out that changing forms enables and impels changes in functions and that, conversely imposing old forms may constrain our ability to function effectively. New biological species emerge through the gradual accumulation of mutations operating on a very small scale, but widespread speciation, as well as its evil twin, extinction, is often driven by large scale disruptions, such as global changes in environmental conditions. We should expect that in the twenty-first century some older forms of records and ways of managing them will not survive. They will either become extinct or retreat into ecological niches in the information landscape. We should also recognize that both marginally different and radically new types of records have already emerged in cyberspace, including new genres that have no parallel in the world of paper and other hard-copy records. With types of electronic records which appear to be counterparts of familiar hard-copy documents, we must be able to recognize and respond to even minor mutations that either imperil or promise to improve the value of records. New species of records will continue to emerge apace with the evolution of digital technology.

A basic difficulty we encounter in trying to apply knowledge derived from experience with traditional records in cyberspace is how to identify an individual record, an item of information that cannot be further decomposed without loss of 'recordness.' In physical space, unit records are often congruent with the physical media on which they are inscribed: a piece of paper, or several pieces stapled or clipped together, or a roll of motion picture film. In cyberspace, what appears as a single document may actually consist of data that are stored in numerous separate objects, each with a different structure and semantics; assembled by means of an intermediate object, such as a view on a database; organized according to the specifications of a form; and presented according to the dictates of one or more style sheets. There is nothing – no single object - stored in the computer system that corresponds to the document presented to a human in such a case. Such a document - which I call a "pseudo-document" or "pseudoc" for short - fails to satisfy one of the defining characteristics of a record: it is not kept.
    

How then do we deal with the case where a pseudoc is the exact counterpart of a traditional record and serves the purposes the traditional record would have served? The first InterPARES project studied numerous cases of this sort and concluded that it is literally not possible to preserve such records in electronic form. It is only possible to preserve the ability to produce copies of the records. This led to the articulation of sets of benchmark and baseline standards for maintaining the ability to produce authentic copies of such records. But, if we are not preserving the records, but only the ability to reproduce them, what is the stuff we do keep in digital form? Analyzing cases which encompassed both digital replicas of hard copy records and electronic records which have no traditional counterparts, Duranti and I arrived at the necessity of distinguishing two different classes of records in such cases: the information kept in digital form, which we designated as a 'stored record' and the rendering of this information in a form suitable for human use, characterized as a 'manifested record.'
    

Keeping a manifested record in the manifested form would be redundant given the stored record. Thus, the term, 'manifested record,' is shorthand for "a copy of the record we would have kept had we decided to keep it in human readable, rather than digital, form." The manifested record may exactly reproduce the content and form that a human author saw when the record was created, but it may also be a document that a computer application produces de novo from data extracted or derived from the contents of one or more other documents created either by human authors, by processing of externally originated data, or by system to system interactions. The output from one or more stored records is not necessarily a record. It may be a temporary, evanescent display which is not saved. Nevertheless, given proper procedures and controls, a manifested record can be an authentic copy of a record.
    

How do we identify and delineate individual digitally stored records? In simple cases there can be a one-to-one correspondence between a digital file and a manifested record. But in many cases, the relationship between stored and manifested records can be many-to-one, one-to-many, and even many-to-many. Some pieces of digital data can be used in many records of many different actions. For example, data identifying a customer will appear in every interaction with that customer. They will appear even in multiple records related to a single transaction, such as the order, shipping manifest, and invoice for a sale of merchandise. While the data identifying the customer are necessary parts of the contents of such records, the chunk itself carries no clue of the actions or records in which it participates. The InterPARES project distinguished pieces of data that occur in or contribute to the reproduction of manifested records, calling them "digital components" of electronic records. A digital component is a bit string that is necessary to produce a manifested record. We need to manage at the level of digital components, as well as at the level of records, in order to ensure that we can reproduce manifested records from their digital components.
    

Digital components should not be limited to portions of content. A digital component might not consist of content, but define what content should be included in a manifested record, such as in a database view. It could also specify the semantics, syntax, or presentation of the record; for example, a statement in an XML Schema, an Xpath query, or a Cascading Style Sheet. Basic objects, such as the dynamic load library for a typeface or a color space should be treated as digital components if they are necessary for the output of an authentic copy of a record. Overall, there are four categories of digital components of records: (1) composition data, which tell the system what form and content data belong to a document, (2) the content data, (3) the form data, which determine how the content is arranged and presented, and (4) rules. Several different types of rules can shape the production of a manifested record. For example, rules may define the conditions or circumstances in which the record can be reproduced, or they may exclude or include certain elements of content, depending on a user's privileges, or they may define links or hyperlinks between documents or parts of a document. Again, they should be retained as long as any manifested record which they control is needed and they should be managed to ensure that authentic copies can be produced.
    

But we are still left with the question of how to distinguish individual stored electronic records in cases where there is not a one-to-one correspondence between a stored and a manifested record. A record is something that is kept. Therefore, a stored electronic record must be a persistent object that is maintained in a computer system. A record provides information or evidence about an activity or a state of affairs at a particular time; therefore, the persistent object must contain fixed, invariant data. But it is not necessarily the case that the stored record itself provides complete information about one or more actions or a particular state of affairs. This is not peculiar to the digital realm. At least since World War II, it has been extremely common that the complete 'record' of a single action is contained in many different documents. Furthermore, a stored electronic record does not necessarily provide such information or evidence directly in the form in which it is stored. There is no a priori reason why an actor could not create records that require some combination with other records, or some processing in order to deliver meaningful information about past actions or situations. Again, this is not peculiar to cyberspace. Many governments, for example, do not keep birth certificates as distinct documents. Rather they capture the necessary data about each birth in a registry that contains the same data elements about all births within the jurisdiction. An individual's birth record can be produced on demand as a separate document by copying the relevant data from the registry onto a blank birth certificate. This is essentially analogous to the production of a manifested record from a stored digital record, with notable differences that computers provide much more flexibility in how the data about individual activities or situations are recorded and – because they not only capture, organize, and store data but also participate in the execution of business processes – can more reliably enforce data quality by embedding business rules in the execution of processes. Thus, a stored electronic record can satisfy the requirement that records provide information or evidence if it can be used to produce one or more manifested records that directly communicate such information or evidence.
    

We can, then, formulate a simple set of three criteria for identifying a stored electronic record: it must be (1) a persistent digital object that (2) contains fixed information about an activity or the state of affairs at the time when the action was done, and that (3) can be used to produce one or more manifested records. The stored record may be a single information chunk or digital component, but it might just as well contain many thousands of such elements. There is no a priori limit to the structure or content of a stored electronic record and it may contain data in any one or more of the four types of digital components; that is, content, composition, form, or rule.
    

The properties of records inevitably impact methods for managing them. Methods that we have come to think of as fundamental to managing records can be seen on reflection also to be artifacts of hard copy technology. Traditional filing systems are based on the physicality of hard copy records, on location or more specifically on collocation in file folders, filing cabinets and file stations. It is certainly possible to import such approaches into cyberspace, at least as icons, as we see in Windows file management and email management products. This approach is embodied in records management applications that implement the Department of Defense standard, DoD 5015.2-STD. These applications effectively create virtual filing cabinets in cyberspace and allow us to manage electronic records as if they were in metal cabinets. But possible does not equate to optimal. What is important in the digital realm is neither the media on which information is recorded, nor the physical place where it is stored, but the possibilities that digital technology creates both for the forms of information that can be created and the ways they can be organized and used. The diversity of ways in which data can be organized; for example, in relational and object-oriented databases, data warehouses, and geographic information systems, are obvious advantages of digital technology, as are the possibilities for multiple simultaneous arrangements and for virtually instantaneous recombinations.
    

Do records exist in such applications? Can they exist? We know that it is certainly possible to produce manifested records from such applications. Selected content of databases is commonly output in the equivalents of traditional forms and reports. But the output of a system is not the same as what it contains. If so, how can we manage the records? One possibility might be to take them out of such applications and put them into virtual filing systems. But to do would diminish their usefulness in the conduct of business for the basic reason that they would exist apart from the systems used to execute business processes and most likely in formats that would not be useful in these processing. Breaking the organic links between records and activities would also diminish their value as records. If you want to know what an organization did and you had two places to find out: one the system the organization used to conduct its business and the other a repository where it put special forms of information that satisfy some abstract criteria for evidence, which would be the better source? Other things being equal, the system that contained the information used in the conduct of business, in the forms in which it was used, would be the better source. But isn't the fundamental purpose of creating and keeping records to provide a privileged source of information about prior activity? Does this represent a dilemma? How can we manage records if we don't put them in filing systems? There are other ways, made possible by digital technology. One method that is emerging is called Records Management Services. It was initiated by the National Archives and Records Administration, articulated in collaboration with 19 other federal agencies and is now being developed, with much broader participation, as a voluntary standard under the aegis of the Object Management Group. In brief, Records Management Services provide methods for identifying records in practically any type of computer application and form managing them within their native applications.
    

Can we discern a path for records management that will enable it to survive and prosper in cyberspace? Certainly, the basic role of helping an organization to determine what records it needs to create and keep remains essential in cyberspace, with some adaptation, such as determining what records the creator needs to be able to manifest, rather than to keep. While records managers cannot personally have sufficient IT expertise even to implement decisions about what records to keep, they must be able to work with and guide a range of IT specialists in developing and maintaining systems that meet requirements for records creation and retention. But activities like this essentially translate traditional activities into the digital realm. That is not likely to be sufficient for records management to truly prosper.
    

Records management can contribute to realizing the potential value of digital technology in a way that best satisfies the needs of organizations and individuals in the conduct of their business. To do so it must be able to identify and show how records exist in business systems and it must offer methods for managing electronic records that convincingly add value in the conduct of business. For this, the discipline must move beyond established knowledge and methods. Not only the specific knowledge, but even the kinds of knowledge that have stood us in good stead in the hard copy world cannot migrate into cyberspace. We should not seek to develop the digital equivalents to the kinds of expertise we had, for example, in the relative merits and drawbacks of end digit filing, or the differences between diazo and silver halide films. For one thing, the digital realm is too big and complex for us to develop sufficient knowledge about the different ways information can be encoded, organized and used. For another, by the time anyone could develop such mastery of any particular digital data type, the technology would have changed.
    

What, then, is the potential for records management to add substantial and readily recognized value in this context? Just as the basic definition of ‘record’ does not need to change in cyberspace, the basic knowledge that records managers have of the most important information assets an organization has, its records, is a solid foundation for helping the organization transition to and operate in cyberspace. After all, records managers are not alone as neophytes in the digital dimension. Indeed, records managers are advantageously positioned to help organizations navigate in cyberspace because of the simultaneously broad and deep insight they have into what types of information are used in what parts of an organization and for what purposes. They can use this advantage to promote one of the basic objectives of records management: to ensure that all of the right information and only the right information is delivered to the right person at the right time to meet organizational needs.
    

In order to leverage the advantage that knowledge of the organization, content and uses of records provides, records managers need to develop their skills both in applying technical judgment and in relying on technical expertise when appropriate. Technical judgment is the ability to evaluate technological solutions both when they are proposed and after they have been implemented. This evaluation is not of information technology itself, but of the benefits and deficits of particular technologies in relation to given business needs. Thus, the evaluation has an external orientation: focusing not on what technology is or does, but how it is, or could be, applied to meet business requirements; what is entailed in implementing a solution; and what difficulties, risks or shortcomings a technology poses. Technical judgment requires broad familiarity with three areas of technology: (1) technologies used to create, keep, and organize records, (2) technologies used to perform records management work, and (3) technologies used to discover and deliver records.
    

Technical judgment is not equivalent to technical expertise. Indeed, a hallmark of sound technical judgment is the recognition of the limits of one’s own technical expertise, combined with astuteness in identifying experts in different areas of technology, and aptitude in working with them. Records managers may acquire considerable technical knowledge in exercising their profession, but they should avoid assuming that their knowledge is sufficient for two reasons: (1) technical knowledge rapidly becomes outdated and (2) technical specialists know more. In contrast to technical judgment, technical expertise is the ability to adapt, deploy and apply technological solutions. Its focus is on information technology itself, with emphasis on in-depth knowledge of specific technologies. But no amount of technical knowledge is sufficient to determine how well any solution meets business needs. That determination requires in-depth knowledge of the business and its needs for acquiring, producing, recalling and applying information, knowledge that records managers can provide.
    

Thus, a recipe for records management prosperity in cyberspace is for the discipline to expand into the fifth dimension, keeping and applying established knowledge or techniques in cyberspace when the they are valid independently of the context in which applied; discarding any knowledge or technique that is not applicable beyond the context of hard-copy records; adapting concepts or techniques that are fundamentally sound but have not been articulated appropriately for cyberspace; and developing new concepts and techniques that respond to what is new and different in cyberspace.