NOTE TO READER: This literature review was written by Jessica Meyerson in 2014. A revised version, with a few new citations, is currently available in the Software Preservation Network Open Science Framework project. If you would like to respond or contribute to an expanded version of this literature review, please email: contact[at]softwarepreservationnetwork[dot]org with the subject line 'Literature Review."
Introduction to Software Preservation
Experts in cultural heritage preservation have acknowledged the cultural impact and the associated preservation challenges of digital records since at least the 1980s (Peterson, 1984). Since the 1990s, the historicity of software as a particular kind of digital record has begun to reveal itself in terms of the growing volume of artifacts created using software applications (Hedstrom & Bearman, 1990; Hedstrom, 1995; Mitchell, 1996; Rothenberg, 1999; Bearman, 1999) and the development of academic disciplines dedicated to software history (Hashagen, Keil-Slawik & Norberg (Eds.), 2002; Cortada, 2002); the coupling of software to other forms of digital media (Besser, 2000; Manovich, 2013); and the critical treatment of computer code as text (Montfort, 2013). As a result, information studies scholars, cultural heritage preservationists, historians, and cultural/media studies scholars form a discourse around software preservation. In this paper, I will look closely at the demands placed on software preservation and the implications of those demands for the choice of specific preservation methods. My ultimate goal is to demonstrate that the choice of preservation methods is determined by a combination of scholarly research needs and the resource constraints of cultural heritage preservationists.
My selection of resources is hardly exhaustive, however, it is meant to be representative of the discourse around software preservation over the last three decades – a span of time in which our lives have become increasingly software-mediated. I have chosen a mixture of peer-reviewed journals, proceedings, books, reports from government-funded projects undertaken by information preservation institutions and excluded software trade publications, magazines, online forums and listservs, and other relevant, online-only resources. Aware of the broad scope of “digital preservation,” I chose to limit my digital preservation resources to those representing a major turn or disagreement in the discourse (Rothenberg, 1999; Bearman, 1999), foundational/historical examples that date the problem of “software-dependence” (Peterson, 1984) in cultural heritage preservation and several pieces from the archival literature over time to show a persistent awareness of software preservation as a key aspect of digital preservation (Bearman, 1987; Hedstrom & Bearman, 1990; Hedstrom, 1995). For the digital preservation publications, I searched within specific publications including International Journal of Digital Curation, Archivaria, and Archives and Museum Informatics. In order to focus more exclusively on software preservation or cultural studies of software, I searched within the journals mentioned above as well as IEEE Annals of the History of Computing and new media venues such as the Journal of Visual Culture. Based on my experience as a member of the software preservation community of practice, I also searched for publications by practitioner/scholars that I knew to be active in this area, including Jeff Rothenberg, Margaret Hedstrom, David Bearman and Nick Montfort.
As described above, the earlier resources (1980s – 1990s) I have selected demonstrate a growing awareness that software preservation is an integral part of the broader challenge of digital preservation. In the 2000s, software preservation begins to receive attention as a topic in its own right and there is significant effort to identify and describe its “significant properties” (Hedstrom & Lee, 2002; Matthews, 2008). The sub-discipline of “critical code studies” (Montfort, 2013:3) develops during this time within the hermeneutic tradition of cultural/media studies and broadens the scholarly user base for software archives or collections.
I have excluded works from software engineering and software development venues. Although those fields are responsible for many of the methods used in software preservation, I have chosen not to include them in my sample because of the narrower, systems-based approach to software rather than a socio-technical approach that combines an understanding of software technologies with an interest in its relationship to human experience more broadly. With that said, incorporating preservation concerns early in the software development lifecycle has become an increasingly popular preservation intervention, particularly in the United Kingdom (Barateiro et al. 2012; Matthews, 2010). Finally, at the risk of “lump[ing] all software together and treat[ing] it as a single phenomenon,” (Endres, 2002:81) I have avoided focusing exclusively on the preservation of one type of software (Ex. proprietary word processing applications v. space flight software) because the body of literature is too limited at this time to be subdivided by software type.
Software Preservation: The Breakdown
In 1990, Margaret Hedstrom reported on her attendance at a symposium on the “Preservation of Microcomputing Software,” in which conference participants’ interest in software preservation was motivated by two primary goals: 1) the ability to “retrieve and process historical machine-readable data” and 2) to preserve software as a unique source of information (Hedstrom, 1990, p.10). Major issues related to these goals included the feasibility of software collection/preservation in light of the required technological resources and “increasing legal constraints” (Hedstrom, 1990, p.10). Twenty-three years later, a similarly comprised group of scholars and practitioners came together for “Preserving.exe: Toward a National Strategy for Preserving Software” summit, resulting in a report remarkably similar in its perspectives and recommendations to those described above (National Digital Information Infrastructure and Preservation Program, 2013). As a result, the organization of this paper mirrors the long-standing, over-arching themes in the discourse: the two major problems that software preservation hopes to solve including “Software as a Utility” and “Software as Evidence and Information”; and “Preservation Methods.”
PROBLEM 1: Software as Utility
A task force report on digital archiving commissioned by the The Commission on Preservation and Access and The Research Libraries Group, reported that ‘while libraries continue to deal primarily in paper formats, archivists have seen drastic and rapid changes from paper/analog materials to a range of digital formats’ (Waters & Garrett, 1996, p.5). Evidence supporting that claim dates back at least as far as 1984, when Trudy Huskamp Peterson declared that “format makes no difference to archival principles” (Peterson, 1984, p.385). Peterson’s rhetorical goal was to remind a profession fraught with anxiety over the future influx of electronic records, that the same archival principles (provenance, original order, and respect des fonds) and activities (acquisition, appraisal, and preservation) that archivists had been trained in up to that point would guide their work into the information age. Peterson did, however, articulate several practical differences for appraisal activities in the context of digital records, including software-dependency: “A software-dependent file will print out as gibberish unless it is processed on a computer that has the right software” (Peterson, 1985, p.386). Peterson was concerned exclusively with digital records containing mostly textual information which, due to their relative simplicity, can be interpreted without the use of word processing software (Rothenberg, 1999, p.8; Besser, 2000, p.166). In contrast, increasingly complex digital records such as the “physical models” created in modern architectural practice, can only be accessed in the same software that created them (Mitchell, 1984, p.202,203), or “some closely related software” (Rothenberg, 1999, p.8). In this sense, digital documents only exist experientially through “software performances” (Manovich, 2013, p.33)
To preserve software as a means of access for other types of complex digital documents ultimately mandates treatment of this software, whose initial may have been purely utilitarian, as collection material in itself. Installation media must be migrated, environments must be emulated, and metadata has to be recorded to reduce future efforts to access the complex files that necessitate this process. As a result, these software collections become cultural records in their own right and provide an entrance point for scholars interested in making meaning from the digital/material past.
PROBLEM 2: Software as Information Itself
In 1986, the Computer History Museum (then in Boston) and the Charles Babbage Institute commissioned David Bearman to write a report that addressed the challenges of creating a national software archives. Limited research interest in software at the time led Bearman to conclude that the development of a software archives and the growth of a dedicated “user community” would be co-determined (Bearman, 1987, p.41). Additionally, Bearman describes several possible user groups, each with their own implications for archival and curatorial work: 1) scholarly research on distribution, legal protections, and the social norms of developer communities; 2) researchers interested in causal explanations of why certain applications succeeded or failed; 3) “internalist” historians of science or intellectual history; and 4) “system designers” who can not fully understand a system by merely examining its constituent parts (Bearman, 1987, p.41-46). The first two user groups, Bearman claims, can answer their research questions using materials most familiar to archival repositories such as correspondence between the company and its customers, promotional material, development documentation, material published in trade magazines, and communications between developers that collaborated on the same project (1987, p.45,46). The last two user groups, however, require access to the source code, a functional copy of the application or both (Bearman, 1987, p.45,46). Bearman’s breakdown of software collection, preservation and access was intended to inform policy and investment decisions of two major computer-culture heritage institutions, and in retrospect Bearman’s description has held up remarkably well. In the following paragraphs I will provide an updated representation of the software preservation space that builds on a de-emphasized aspect of Bearman’s documentation strategy, the retention of software as information in itself.
By the 2000s, the history of computing discourse turned towards software itself as a primary resource, or as Cortada (2002, p.77) declares in his article entitled “History of Software from the 1960s,” “Manuals are not enough.” Cortada goes on to argue that historians of software must experience software “to appreciate its birth, evolution, and function” (Cortada, 2002, p.77). The early 2000s also marked the inaugural volume of Iterations: An Interdisciplinary Journal of Software History, published by the Charles Babbage Institute as “a forum for scholarship and other discourse on the the history of software” (Charles Babbage Institute, 2002). Zabolitzky agrees that access to the software itself is essential to a history of software, but builds on Cortada’s experiential emphasis, claiming that while binaries are valuable for “accessing the sensual dimension of information processing”, the source code is the ‘only record that includes all the information concerning what was actually done’(Zabolitzky, 2002). The 2000s witnessed another hermeneutic shift towards the study of software within the field of cultural studies. In “Behind the Blip: Software as Culture,” Michael Fuller argues that a “fully fledged software criticism” places a specific program or level of a program in a broader “assemblage” of both meaning and meaning creation, necessitating access to software itself (Fuller, 2002, p.1-7). Manovich’s “software studies” focuses exclusively on the interface level of an application, placing priority on the executable form of software (Manovich, 2012). More recently, Montfort argues that writing and running software are important in any attempt to understand “code and its cultural relevance” (Montfort, 2013), implicating the source code as the object worth preserving. Although this paper emphasizes historical and cultural studies research, the value of software as information itself is not exclusive to the hermeneutic traditions. For example, researchers interested in the strategic use of information technology within business organizations benefit from the study of software source code written in-house prior to shifts towards cloud-based architecture (Barateiro, 2012, p.488).
Bearman’s prediction that the creation of software archives and the formation of dedicated user groups would coincide has proven accurate. However, the accuracy of his predictions for the artefactual/documentary needs of those dedicated user groups remains unclear. Since the early 2000s, an interest in archival documentation surrounding the development and use of software has been augmented with interest in some form of software as information in itself. The major difference in researcher attitudes or preferences towards the software artifact is represented by a desire to derive meaning from the experiential aspects of software on the one hand, and the desire to derive meaning from critical examinations of the source code on the other. The next section explores the relationship between research preferences and software preservation in practice.
Solving Problems: Preservation Methods
Denert & Lohr (2000) dismiss the notion that a museum (and presumably any other cultural heritage institution) would actually run software on a computer, that it is in fact “obvious that this approach would never work” because “the possibilities for running legacy code are very limited” (Denert & Lohr, 2002, p.247). In Elizabeth Kaplan’s (2002) response to Zabolitzky’s (2002) article on software preservation she recounts the outcomes of the First National Software Archive Conference and speculates on why there had been no new work towards the National Software Archive since the conference, namely, the lack of a “sizable” user group and the considerable resources that have to be dedicated to “conserving, organizing, researching, cataloging and presenting materials in ways that researchers can use” (Kaplan, 2002). Furthermore, Kaplan believes that efforts/expertise required both on the part of the repository and the researcher point to software-related documentation as the only “economically viable” form of software preservation (Kaplan, 2002). The value of “computer-related source materials” to the work of historians of computing dates as far back as 1980 and includes “correspondence, working papers, unpublished reports, obsolete manuals, key program listings used to debug and improve important software” and more (Sammet, 1980, p.4). In addition to preservation of textual software documentation, oral histories and other ways of capturing specialized, tacit knowledge about older generation software and its dependencies have been advocated by cultural heritage preservationists (Waters & Garrett, 1996, p.2; Besser, 2000, p.166) and scholars (Cortada, 2002, p.76) alike. Once considered a stopping point, collecting software-related documentation is now considered integral to any other software preservation method.
Evocative of Bearman’s (1987) topological description of researcher needs and the implied archival/curatorial requirements Matthews, et al describe a “performance model” that addresses two “major aspects of software preservation” – “reconstruction” and “replay” (Matthews, 2010, p.92,93). Although Matthews et al., consider software as information in itself, they feel that “adequacy” of a software performance, or the success of a particular software preservation workflow, is measured by “the processing of its input data” (Matthews, 2010, p.94), reintroducing the notion of ‘software as a utility,’ and effectively demonstrating that contemporary digital preservation methods address both problems presented in this paper as opposite sides of the same proverbial coin.
There are several well-known digital preservation methods including migration, and emulation and running legacy code on legacy hardware, and “all can in different circumstances support good preservation” (Matthews et al., 2008:, p.5). Migration has been defined as the ‘transfer of digital content from one hardware/software generation to another’ (Waters & Garrett, 1996, p.iii) and raises questions for preservationists about “whether to treat digital materials as artifacts or simply as intellectual content” (Pace, 2000, p.55). Jeff Rothenberg criticizes migration, pointing out that “computer scientists, data administrators and data processing personnel” have used migration in order to provide continued access to “documents,” “data,” and even “programs” but that the general consensus among these practitioner/users has been that is “labor-intensive, time-consuming, expensive, and error-prone” and for these reasons it does not scale (Rothenberg, 1999, p.8). Furthermore, Rothenberg argues that migration requires the preservationist to “stay exactly one [technological] cycle behind” which imposes a persistent urgency to migrate (Rothenberg, 1999, p.8). In contrast, Bearman (1999) emphasizes the advantages of migration as a metadata-heavy strategy and argues that Rothenberg overlooks or chooses not to discuss the role that metadata would have to play in any alternative preservation method (Bearman, 1999). While Bearman’s argument draws attention to the importance of metadata in the preservation process his description of migration assumes that digital artifacts/records/objects being migrated are such that their “significant properties” (Hedstrom & Lee, 2002, p.218) can be retained throughout innumerable cycles of change.
So, while migration addresses the utilitarian aspect of perpetual access to simple digital documents independent of any one generation of hardware or software, Woods & Brown argue that there are digital objects that require emulation because “migration paths are nonexistant or result in unacceptable loss” (Woods & Brown, 2008, p.86,87). Montfort describes “porting” (migration process specific to the software context) as a process whereby a programmer has to develop “an ethos of adaptation” that helps guide her in determining a “program’s essential properties” and ultimately, the relationship of the ported program to the original (Montfort, 2013:52). For the purposes of studying software as information in itself, ”porting” introduces a question about the features or “significant properties” (Hedstrom & Lee, 2002, p.218) that characterize software as a unique form of digital record.
The 2008 JISC report on “Significant Properties of Software” identifies seven categories of features that characterize software, including “Functionality,” “Software Composition,” “Provenance and Ownership,” “User Interaction,” “Software Environment,” “Software Architecture,” and “Operational Performance” which can be collapsed further into five categories of significant properties which include “Content,” “Context,” “Rendering,” “Structure,” and “Behavior” (Matthews et al, 2008, p.67,68). Manovich’s interface-centric “software studies” (Manovich, 2012) would likely consider the “Behavior” of an application to be its most significant property while perhaps not realizing that it is also the most abstract, relying heavily on the other lower-level properties and their associated feature categories. Additionally, several of the applications listed in Manovich’s categories of “cultural software” (Manovich, 2010, p.23) are both complex and proprietary, in which case migration is impossible because the source code is unavailable.
Emulation is a method in which the environment required to access older generation digital objects is re-created on a modern computer, allowing legacy software to run inside an emulation layer on current systems. Rothenberg considers emulation to be the “only reliable way to recreate a digital document’s original functionality, look, and feel” (Rothenberg, 1999, p.17), while other have argued that any preservation strategy that involves “preserving information systems in which [digital records] were generated” ignores the “functional requirements” for retaining the evidential value of digital records, and presents legal challenges associated with the collection of proprietary software that would be required in order to emulate the most popular systems (Bearman, 1999). Hedstrom & Lee consider emulation to be an appropriate method in cases where functional/behavioral/performative properties are considered most significant by the “designated user community,” but agree that it is not the “most effective or cost-effective strategy” in every case (Hedstrom & Lee, 2002, p.222). While recent developments towards cloud-based emulation known as “emulation-as-a-service” or “EaaS” (von Suchodoletz, Rechert & Valizada, 2013, p.137) may help to mitigate persistent concerns over the “significant cost, overhead, and development resources” associated with the emulation method (Carroll et al, 2011, p.78), EaaS still faces the challenge of obtaining access to “original software applications, operating systems, firmwares and drivers, including the the appropriate rights to them” (von Suchodoletz, Rechert & Valizada, 2013, p.137). Beyond its ability to preserve performative or experiential properties of digital artifacts, Carroll et al, claim that emulation has several key advantages including its ‘paratextual focus,’ its historicity as a method in engineering and computer science, and its ‘verifiability’ (Carroll et al, 2012, p.78). The first of these speaks directly to those researchers interested in studying software from the “perspectives of politics, society, and systems of thought and aesthetics” (Fuller, 2008, p.2). The second advantage addresses early concerns over emulation as a viable method in any respect, preservation or otherwise (Bearman, 1999). The third advantage introduces the issue of evaluation, or what is measured against to determine if an emulator is successful at retaining significant properties.
Although preservationists agree that attempting to keep original hardware operational will never work as a preservation in the longer term, it is considered useful in the short term for evaluating the success of emulators (Rothenberg, 1999, p.12,13; Swade, 2002, p.223,234; Matthews et al, 2010, p.95). Another method worth mentioning, but considerably less visible in the discourse, is preservation intervention at an early stage in the software development lifecycle (Matthews et al, 2010, p.101). Barateiro argues that migration and emulation are not sufficient because both methods attempt to solve software preservation “posteriori” which “results in undesirable costs” (Barateiro, 2010, p.489). In the context of in-house software development this method is feasible because preservation planning can be mandated as part of the workflow. However, in the case of software development more broadly, preservationists may struggle with ways to motivate software development companies to change their practices on behalf of cultural heritage (Peterson, 1984, p.392). Ultimately, of the preservation methods currently in use, emulation provides the most robust solution to the problems of 1) access to complex digital documents and 2) access to software as information in itself. Recent cloud-based technologies present a way to share the otherwise prohibitive costs of emulation. However, while centralization of effort promises affordability, it requires discursive agreement over the significant properties of the application that will run, not to mention access to software in binary or source code form. Evaluating the success of an emulation is an additional challenge that would require agreement among institutions and their users.
Methods and Theory in the Discourse
The texts I have chosen to represent the software preservation discourse include a sampling of scholarship from across the fields of cultural heritage preservation, history, and cultural studies. Therefore the range of methods used in this sample include: literature review/content analysis (Bearman, 1987; Hedstrom, 1995; Rothenberg, 1999; Cortada, 2002; Carroll et al, 2011) close reading (Montfort, 2008, p.198), programming and running code (Manovich, 2013; Montfort, 2013), case studies, pilot projects, and system building (Waters & Garrett, 1996; Matthews et al, 2008; von Suchodoletz, Rechert, & Valizada, 2013). Cultural heritage practitioners/preservationists are the most actively engaged in creating test suites, pilot projects and system building. They also employ literature or content analysis in order to articulate their relevance to the software preservation or digital preservation discourses (Bearman, 1987; Hedstrom, 1995; Rothenberg, 1999; Carroll et al, 2011). Theory is more heavily emphasized in texts from cultural studies which apply “actor-network theory, social semiotics, or media archeology” (Manovich, 2013, p.15) to the study of software. The literary method of “close reading” is also applied to source code (Montfort, 2008, p.198; Montfort, 2013, p.3). Considered essential to the task of “software criticism” (Fuller, 2008, p.198), scholars in the field emphasize the practice of programming and its importance in critical code scholarship (Manovich, 2013; Montfort, 2013).
Conclusions and Future Work
The small sample I have chosen for this literature review did not allow me to fully map the contours of this discourse. Although it was not discussed in this paper, there been significant schema development recently that will allow preservationists to capture the hardware and environmental dependencies required to emulate software, and questions of their effectiveness and who is involved in their development deserve closer examination. I have also chosen to not to discuss to any great extent, the preservation work done specifically on digital games, although the ‘nostalgia factor’ driving interest in game preservation is worth exploring further as means of generating interest in the preservation of more commonly used software such as word processors (Kirschenbaum, 2013, p.13).
Based on this review of the literature, software preservation can be characterized an experimental endeavor and progress has been co-determined by the formation of users interested in producing scholarship about software and its cultural relevance. One of several remaining questions is whether or not concerns over the resource investment required for preservation and the legal risks associated with collecting proprietary software will continue to be a source of uncertainty for cultural heritage institutions to the extent that it prohibits software collection. The current paradox of software preservation is that unless we collect this software, the scalable solutions for implementing otherwise resource-prohibitive preservation methods will never actually scale. One information studies scholar suggests that practitioners of software preservation should exploit their legal exemptions more heavily in order to achieve their preservation and access goals (Kraus, 2011, p.106-7). Another open question is whether it is helpful to think of the members of this discourse as a “community of practice” (Barley & Kunda, 2001, p.87), the boundaries of which may extend beyond my sample and possibly beyond the confines of the academy. For example, early work on emulation was undertaken by digital gaming hobbyists who did not have long-term preservation in mind but have been notably successful in bringing the experience of older games to life for successive generations of players (Swalwell, 2011, p.265). Is the work of defining those boundaries a worthwhile endeavor? What would the coordination of a software preservation community of practice that extended beyond academia entail? What activities or goals could a software preservation community of practice accomplish that members of a discourse could not? Furthermore, which group would be most effective at coordinating the activities of this community? There is evidence in the discourse that archivists may be effective at leveraging the skills of community members in light of their track record for stewardship of cultural heritage (Mathews et al, 2008, p.5), their professional training in appraisal and descriptive methods (Sammett, 1980, p.6), and the dedication of organizational resources that archives have commuted to the preservation enterprise (Calwell, 2011, p.264). These uncertainties present possibilities for shaping a research program that considers the relationship between the scholarly research community, the lay community of digital hobbyists and cultural heritage institutions, and the ways in which those relationships are mediated by technologies – both the technologies used for preservation and technologies being preserved.
Barateiro, J., Draws, D., Neuman, M. A., & Strodl, S. (2012). Digital Preservation Challenges on Software Life Cycle (pp. 487–490). IEEE. doi:10.1109/CSMR.2012.63
Barley, S. R., & Kunda, G. (2001). Bringing work back in. Organization science, 12(1), 76–95.
Bearman, D. (1987a). Collecting Software: A New Challenge for Archives & Museums. Archives and Museum Informatics. (Archival Informatics Technical Report [August 1985].1987;1, no.2)
Bearman, D. (1999b). Reality and chimeras in the preservation of electronic records. D-Lib Magazine, 5(4) Retrieved from http://www.dlib.org/dlib/april99/bearman/04bearman.html
Besser, H. (2000) Digital Longevity. In S. Maxine (Ed.) Handbook for Digital Projects: A Management Tool for Preservation and Access (pp.155-166) Andover, MA: Northeast Document Conservation Center
Carroll, L. et al. (2011) A Comprehensive Approach to Born-Digital Archives. Archivaria. 72, 61-92.
Charles Babbage Institute (2002) Iterations: About Retrieved from http://www.cbi.umn.edu/iterations/
Cortada, J. W. (2002). Researching the History of Software from the 1960s. Annals of the History of Computing, IEEE, 24(1), 72–79.
Denert, E. & Lohr, K. (2002) Towards a Software Museum. In Hashagen, U., Keil-Slawik, R., Norberg, A.L., (Eds.) History of computing: software issues: International Conference on the History of Computing, ICHC 2000, April 5-7, 2000, Heinz Nixdorf MuseumsForum, Paderborn, Germany. Berlin ; New York: Springer (pp.245-259)
Endres, A.. (2002) Towards a Software Museum. In Hashagen, U., Keil-Slawik, R., Norberg, A.L., (Eds.) History of computing: software issues: International Conference on the History of Computing, ICHC 2000, April 5-7, 2000, Heinz Nixdorf MuseumsForum, Paderborn, Germany. Berlin ; New York: Springer (pp.77-81)
Fuller, M. (2002a). Behind the blip: Software as culture. Nettime mailing list, 7. Retrieved from http://noemalab.eu/wp-content/uploads/2011/09/fuller_sw_as_culture.pdf
Fuller, M. (2008b). Software studies a lexicon. Cambridge, Mass.: MIT Press. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&scope=site&db=nlebk&db=nlabk&AN=228271
Hedstrom, M., & Bearman, D. (1990). Conferences. Archives and Museum Informatics, 4(1), 10–15. doi:10.1007/BF02770069
Hedstrom, M. (1995). Electronic Archives: Integrity and Access in the Network Environment. The American Archivist, 58(3), 312–324. doi:10.2307/40293923
Hedstrom, M., & Lee, C. (2002). Significant properties of digital objects: definitions, applications, implications. Presented at the DLM Forum. Retrieved from http://www.ils.unc.edu/callee/sigprops_dlm2002.pdf
Kirschenbaum, M. (2013) The Executable Past: The Case for a National Software Registry. National Digital Information Infrastructure and Preservation Program, Preserving.exe: Toward a National Strategy for Software Preservation (pp.12-22) Retrieved from http://www.digitalpreservation.gov/multimedia/documents/PreservingEXE_report_final101813.pdf
Kraus, K. (2011). “A Counter-Friction to the Machine”: What Game Scholars, Librarians, and Archivists Can Learn from Machinima Makers about User Activism. Journal of Visual Culture, 10(1), 100–112. doi:10.1177/1470412910391576
Manovich, Lev (2013). Software Takes Command. Retrieved from http://www.eblib.com
Matthews, B., McIlwrath, B., Giaretta, D., & Conway, E. (2008). The significant properties of software: A study. JISC report, March. Retrieved from http://www.jisc.ac.uk/media/documents/programmes/preservation/spsoftware_report_redacted.pdf
Matthews, B., Shaon, A., Bicarregui, J., & Jones, C. (2010). A framework for software preservation. International Journal of Digital Curation, 5(1), 91–105.
Mitchell, W. J. (1996). Architectural archives in the digital era. American Archivist, 59(2), 200–204.
Montfort, N. (2013). 10 PRINT CHR$(205.5+RND(1));:GOTO 10. Cambridge, Mass.: MIT Press. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&scope=site&db=nlebk&db=nlabk&AN=513654
National Digital Information Infrastructure and Preservation Program, Preserving.exe: Toward a National Strategy for Software Preservation (2013) Retrieved from http://www.digitalpreservation.gov/multimedia/documents/PreservingEXE_report_final101813.pdf
Peterson, T. H. (1984). Archival Principles and Records of the New Technology. The American Archivist, 47(4), 383–393. doi:10.2307/40292703
Rothenberg, J., Commission on Preservation and Access, Council on Library and Information Resources, & Digital Libraries. (1999). Avoiding technological quicksand: finding a viable technical foundation for digital preservation : a report to the Council on Library and Information Resources. Washington, DC: Council on Library and Information Resources.
Sammet, J.E. (1980) Preserving Computer-Related Source Material. Annals of the History of Computing, 2(1)
Swade, D. (2002) Towards a Software Museum. In Hashagen, U., Keil-Slawik, R., Norberg, A.L., (Eds.) History of computing: software issues: International Conference on the History of Computing, ICHC 2000, April 5-7, 2000, Heinz Nixdorf MuseumsForum, Paderborn, Germany. Berlin ; New York: Springer (pp.227-235)
Von Suchodoletz, D., Rechert, K., & Valizada, I. (2013). Towards Emulation-as-a-Service: Cloud Services for Versatile Digital Object Access. International Journal of Digital Curation, 8(1). doi:10.2218/ijdc.v8i1.250
Waters, D. & Garrett, J. (1996) Preserving Digital Information, Report of the Task Force on Archiving of Digital Information. Council on Library and Information Resources (CLIR) Report (Publication 63)
Woods, K., & Brown, G. (2008). Migration performance for legacy data access. International Journal of Digital Curation, 3(2), 74–88.
Zabolitzky, J. G. (2002). Preserving software: Why and how. Iterations: An Interdisciplinary Journal of Software History, 1(13), 1–8.