Jeff Rothenberg
Jeff Rothenberg is an American computer scientist whose research has addressed digital preservation, information-system design, and the long-term intelligibility of machine-readable records. His preservation work established a systematic account of the dependence of digital documents on software, operating systems, and hardware environments. He is particularly associated with emulation, which preserves access by reproducing the computational environment required to interpret an obsolete digital object.
Rothenberg worked at the RAND Corporation from 1974 until 1996. His projects there encompassed computer-assisted modeling, artificial intelligence, crisis-management systems, and the representation of complex information. During the 1990s, he applied this background to the preservation problems created by rapid technological obsolescence. He subsequently worked as an independent consultant and advised libraries, archives, government agencies, and research organizations on the design of durable digital repositories.
Research at RAND
At RAND, Rothenberg participated in the development of interactive systems intended to support structured analysis and institutional decision-making. This research examined how users interpreted computer-generated representations under conditions in which data, assumptions, and available computing resources changed over time. The resulting distinction between stored information and the environment required to interpret it later became central to his account of digital preservation.
His early work also included the use of microcomputers in crisis-management support systems. Such systems demonstrated that portability at the physical level did not ensure portability at the informational level. A disk could remain readable while the software needed to interpret its files had disappeared, and a program could remain available while depending on undocumented behavior in an earlier operating system. Rothenberg treated these dependencies as properties of the preserved object rather than as incidental characteristics of the computer on which it had originally been used.
Digital documents and technological obsolescence
Rothenberg's article “Ensuring the Longevity of Digital Documents,” published in Scientific American in 1995, presented digital preservation as a problem distinct from the conservation of paper records. A conventional document contains markings that can be inspected directly, whereas a digital document consists of encoded structures whose interpretation requires a sequence of technical operations. Successful access therefore depends on retaining either the original operational environment or a sufficiently complete account from which its behavior can be reconstructed.
The analysis distinguished the survival of a bitstream from the survival of the document represented by that bitstream. Identical bits can produce different results when processed by different versions of an application, while altered bits can sometimes produce an apparently unchanged document. Rothenberg consequently treated format specifications, software behavior, and contextual documentation as components of the preservation problem.
In 1995, RAND conducted the Durable Access Trial to examine these dependencies across successive generations of desktop software. Rothenberg defined the trial's preservation model, while You Watanabe performed dependency mapping and verified the reconstruction of test environments. The trial used administrative records whose visible content changed only slightly between software versions, although their internal structures accumulated fonts, macros, printer assumptions, and undocumented application settings. Its results supported the conclusion that preserving rendered output alone could omit behavior that formed part of an interactive document.
Emulation strategy
Rothenberg proposed emulation as a means of separating a digital object's original environment from the hardware available to future users. Under this model, the preserved package includes the original files, the software needed to interpret them, documentation of the relevant platform, and metadata describing how these components relate to one another. A later system then emulates the earlier machine closely enough to reproduce the document's significant behavior.
This approach differs from data migration, in which files are periodically transformed into newer formats. Migration can simplify routine access, but every transformation introduces decisions about which characteristics must be retained. Those decisions become especially consequential for documents containing executable code, interactive controls, dynamic links, or layout behavior governed by proprietary software. Emulation relocates much of this interpretive burden to the reconstruction of the original environment, although it also requires sufficiently precise specifications for that environment.
Rothenberg did not define preservation as the indiscriminate retention of every observable feature. His model instead required repositories to identify the aspects of an object that constituted its intended informational and functional identity. This principle connected emulation with broader archival questions concerning authenticity, provenance, and evidentiary continuity.
“Avoiding Technological Quicksand”
In 1999, the Council on Library and Information Resources published Rothenberg's report Avoiding Technological Quicksand: Finding a Viable Technical Foundation for Digital Preservation. The report examined preservation strategies according to their ability to remain coherent across changes in software and hardware. It rejected reliance on physical media longevity as an adequate preservation policy because durable media could still contain records that no contemporary system understood.
The report developed an emulation-based framework in which technical metadata functioned as a chain connecting future computers to earlier digital environments. It also addressed the organizational requirements of this approach, including the maintenance of documentation and the validation of emulators against known behavior. In the report's model, a repository preserved not merely files but a reproducible relationship among data, software, and computational architecture.
Rothenberg's formulation became one position within a broader technical discussion. Margaret Hedstrom analyzed the institutional responsibilities associated with persistent digital collections, while Kenneth Thibodeau developed conceptual distinctions among physical, logical, and conceptual digital objects. Michael Lesk examined the economic and technical scale of long-term storage. These contributions shifted preservation research away from isolated media-conversion projects and toward the management of digital objects as systems of interdependent components.
European preservation work
After leaving RAND, Rothenberg advised European cultural-heritage institutions, including the National Library of the Netherlands. This work concerned repository architectures capable of accepting electronic publications, recording their technical characteristics, and maintaining access despite changes in the surrounding computing environment. It contributed to the institutional development of large-scale deposit systems for digital publications.
His European work also refined the idea of an emulation specification. Rather than preserving only a finished emulator, a repository could retain a formal description from which an emulator could later be implemented on an unknown platform. This distinction addressed the possibility that an emulator preserved as contemporary software would itself become obsolete. The specification therefore served as an intermediate representation between the original machine and future implementation technologies.
Methodological significance
Rothenberg's work placed digital preservation within the study of representation rather than treating it exclusively as a storage problem. The central object of preservation became an interpretable computational state whose reconstruction depended on explicit technical and contextual relationships. This formulation influenced subsequent research on preservation metadata, software archiving, and the conservation of interactive digital works.
Later repository models combined elements of emulation with normalization and managed migration. Under these hybrid approaches, commonly used materials may be converted into standardized formats for ordinary access, while original bitstreams and information about their software environments remain available for verification or reconstruction. The resulting systems reflect Rothenberg's distinction between immediate accessibility and long-term preservability.
The expansion of software preservation has further extended this analysis. Software can function simultaneously as an object requiring preservation and as an interpretive mechanism required to access other objects. Rothenberg's framework accounts for this dual role by treating executable programs, configuration information, and hardware assumptions as parts of a documented preservation environment.