Materiality and Immateriality
Digital resources have another interesting characteristic from an historical perspective. These resources provide users with an experience analogous to, but not precisely emulating, physical, material resources. They therefore give the perception of “immateriality”. Everything in the digital realm seems to reside in “cyberspace”. Users are accessing “data”, which are immaterial, in contrast to “physical”, material objects and artefacts. The perception of immateriality results from the transparent access that users have to the material: users generally do not think about, nor do they need to be concerned with, the physical storage media on which the data are stored, the microprocessors on which the algorithms execute, or the physical memory into which the “immaterial” data are loaded. The digital nature of the archive therefore abstracts away materiality from the user’s viewpoint. However, maintaining the perception of immateriality and transparency is possible only because of many unseen background processes. For instance, programs are modularized, or written in discrete components. Error-correcting algorithms are employed to reduce the “noise” inherent in physical media. And degradation of physical computational resources, which result specifically as a consequence of their materiality, are hidden from the user through system migration, physical resource redundancy, “garbage collection” to recover unused but still occupied memory locations, and additional error-correcting and noise removal measures. These processes to ensure a transparent virtual experience has a negative effect on performance, although this effect is negligible to most users. The trade-offs of “overcoming materiality” are balanced by performance penalties to ensure the perception of immateriality (Ries & Palkó, 2019).
Consequently, although materiality can be hidden from users, immateriality is only a perception provided by multiple, complex background tasks. Digital resources and environments, appearing as immaterial, are built atop a physical, material substrate, and are therefore subject to degradation. Hardware and software inevitably age and become obsolete. Bitrot, the gradual loss of data on physical media, and diminution of the performance of these media, is caused by physical factors, including the flipping of polarity in bits stored on media due to background electromagnetic radiation (recall that all digital data are represented in the binary code of 0s and 1s). Network environments that operated in the past no longer function. Although not directly related to material considerations, many online services that were part of the original environment of born digital material are now offline, and no longer in service. All these factors raise the question of what and how much needs to be preserved and maintained. Another central question is whether the human-computer interaction experienced by the original creators and users of the digital born resource must be precisely recreated for users engaging with it now. Clearly, text and metadata are integral and necessary components of born digital material. However, the operating system, structure of the file folders/directories, and other aspects of the original digital environment may also be just as essential. The question is then raised to whether it is necessary to preserve the original hardware environment, including the network context, or whether a precise system image that can be emulated at runtime is sufficient (Ries & Palkó, 2019).
One implication is that to preserve and maintain the born digital historical record as completely as possible, all the material facets that were utilized in creating, using, and modifying the digital resource should also preserved and maintained to the greatest extent possible. This implication raises many questions about stability, authenticity, and the degree to which the digital materials are dependent on the computing context in which they resided (Ries & Palkó, 2019). The “materiality” of digital resources and the born digital historical record itself achieve a new importance for preservation and research, archival curation, and digital forensics, described below. The question arises as to whether this context-dependency necessitates the comprehensive preservation of operating systems associated with the digital resource. The straightforward approach to such preservation is to maintain the original physical computational hardware and environment, and the operating systems, programs, and data that run on it and work in it. Preservation can also be effected through emulation – modeling older hardware and/or software on current systems, and virtualization, in which multiple complete computing environments are encapsulated and executed transparently on a single physical computing system. Emulation is important in protecting digital resources, including legacy operating systems and software, from obsolescence. It allows running an operating system or program or reading data in a specific format in a different computational environment from the one in which that operating system, program or data format was originally designed. The result is that older systems can still work properly, and older data can still be read, in current environments. From the user’s viewpoint, the older operating system or program is running in its original environment. In virtualization, a single physical environment is partitioned into “virtual environments”, which gives the appearance of each virtual machine running its own dedicated hardware and software. Conversely, in storage virtualization, multiple physical storage devices are pooled and treated as a single, logical unit by a server system (Emulation or Virtualization). Given these factors, curation again emerges as a vital consideration. Memory institutions must employ standards and criteria for deciding what material requires inclusion to constitute authentic preservation. They must decide which aspects of these digital resources are relevant to future investigations, and what can be left out as an acceptable loss (Ries & Palkó, 2019).