[ICA-EGAD-RiC] Technological and cultural dependencies

[Greg Bak]

On November 28 I sat down with a few other archivists in Winnipeg to hash out some reactions to RiC-CM. We were kindly joined, by Skype, by EGAD member Kat Timms. The following post offers expresses why I think that RiC-CM does not sufficiently address the interrelated issues of the hardware and software dependencies of digital records and the cultural contingency of all record and recordkeeping technologies.

I apologize for the length of this posting, and recognize that my points may not always be completely clear. I would be happy to discuss it further with EGAD or with anyone else.

Best,

Greg

Greg Bak
Assistant Professor of History (Archival Studies)
University of Manitoba




RiC-CM offers a view of information consumption in which “mediation” refers to whether or not a device is needed to render an information object (see for example  P12 Media Type).

The reality is that all information is mediated by technology and by culture: there is no such thing as culturally neutral information, or a non-mediating form of transmission. Information that does not require technology to read it (a book, for example) is not less mediated. Consider a wampum belt: as someone who is not a member of a wampum using culture I can see it but can’t read it. I require mediation to access the information represented in the wampum. Or to take another example, consider a manuscript “news” letter from the eighteenth century: I can see it and read it, but I may misunderstand its meaning if I don’t also know that this was not a private letter exchanged between two confidantes, but rather was written for the purpose of public circulation. This practice was common in eighteenth century Europe and North America, but is liable to be misunderstood today as our own epistolary culture positions manuscript letters as intimate correspondence. Or consider the standard nineteenth century practices of letterbook and docket recordkeeping, reading of which does not require the mediation of a device but which does require that the user understand the system so as to navigate the different physical locations and formats of the correspondence, and the series of indices that link them.

Perhaps this has not often been a problem in archival description in the past not because traditional archival materials do not require mediation, but because: (i) archives have tended to serve culturally homogeneous populations who could be expected to understand the implicit cultural and technological rules around archival records; and (ii) researchers who have been using materials old enough to be from culturally and technologically distinctive eras have tended to be specialists (such as historians) who understood the cultural rules at play.

Both of these factors are changing today.

In 2011 Elizabeth Yakel wrote about the “first great opening” of archives starting in the 1960s and the “second great opening” of archives starting in the 1990s. Yakel points out that archives have assiduously courted new users of archives both by acquiring records of interest to more diverse populations, and by reaching out to more diverse populations and inviting them into the archives. Archives have been working to address the growing diversity of their clientele by creating tools to introduce archival terminology and practices to new users, such as pamphlets, webpages and podcasts. Slowly but surely, these tools are expanding to include background information on obsolete forms of records and record keeping.

Secondly, whereas past bureaucratic technologies have tended to be relatively long-lived and slow to change (the transition from letterbook and docket to subject files happened over decades), we are presently in an era in which digital technologies are changing very rapidly. These changes, moreover, are not just changes in the format of information, but in the functionalities on offer in various digital applications, and the affordances of the devices used to create, manage and use the records.

Patricia Galloway’s 2011 exploration of the user culture that developed around the Kaypro II, an 1980s microcomputer, illustrates some of the constraints on word processing of this era: loading the application from floppies, for example, and being unable to run any other applications simultaneously – not even the spellcheck, which required its own set of application floppies. Content created on the wordprocessor would be stored on yet more floppies, or on some other external memory device, creating challenges for version control, information discovery and access. Someone reading records produced on a Kaypro II who is only familiar with computer usage and wordprocessing today is unlikely to be able to understand such practical constraints on using the Kaypro II, and how such constraints might affect patterns of information creation, transmission, discovery and management.

It is for this reason that the team at Emory University MARBL created their emulation of one of Salman Rushdie’s computers. By allowing researchers access not only to the content that Rushdie produced but also to the digital environment in which he wrote, MARBL signaled their belief that one cannot properly be understood without the other.

The case that I am making here is essentially the same case that has been made repeatedly in our professional literature with regard to paper-based recordkeeping systems, as in Terry Cook’s “Paper Trails” or Bill Russell’s 2013 study of recordkeeping in the office of Canada’s Superintendent General of Indian Affairs. These same points can be made with reference to many configurations of digital technologies, each with their own specific functionalities, limitations and affordances that have been in general use in the last fifty years. Unlike the transition from letterbook and docket to subject files, described so well in Cook’s “Paper Trails”, many digital technologies like the Kaypro II did not attain widespread usage, but still may have been central to the record creating and keeping cultures in which they were deployed (as described in Galloway 2011). Others, like IBM PCs (or clones) running MS-DOS (whose command-line interface mystifies many of my undergraduate students today), did achieve widespread usage, but then rapidly fell out of usage.

This brings us to the question of what are we attempting to describe, and why, through RiC-CM and other archival description standards. Are we simply providing description of the content of the records, so that people can access it, or are we using description to set the content into its appropriate context?

Consider a key issue for digital information: system dependencies.  Tracing the multiple dependencies among applications software, operating systems, firmware and hardware is a necessity for digital preservation. This point is all-too briefly acknowledged in RiC-CM in P18 Conditions for Access, which lists  “software/hardware necessary to access the record” along with other impediments to access such as intellectual property laws and security classifications. To my mind, there are multiple problems with P18, starting with the jumbling of the issue of system dependencies in with intellectual property laws and security classifications. Leaving this aside, though, this single reference to system dependencies in RiC-CM is also problematic as it suggests that system dependencies are only an issue because they may be an impediment in accessing content. There is no acknowledgement in RiC-CM that the functionalities and affordances of digital records and systems are in themselves part of the content that must be preserved and described.

As I have noted, we don’t do this with analogue recordkeeping systems either. But then, we haven’t had analogue recordkeeping systems that have changed so much and so quickly, all within the working memory of people accessing information today. MS-DOS, after all, was still in use in the mid-1990s. Constant broadband access to the Internet over wifi, another game changer, is even more recent. Social media functionalities added into various systems for information creation, annotation and access represent another major shift in the nature and content of digital communications. Each of these had a major impact on how records were created, stored, transmitted, discovered and managed.

This brings us back to the point that I started with: that all information is mediated, but that we are not always conscious of the specific ways that such mediation is woven into our culture and technologies, the fusion of culture into technology. No information communication is “unmediated”, even if no specific device is required to read the information. Our own cultural biases can make us blind to mediations that seem “natural” to us due to our own long immersion in them – remember the example of a belt of wampum (or a petroform, or a wintercount, etc.), which requires no mediating device, but which does require cultural knowledge to decode its meaning. What I have called “cultural knowledge” in the previous sentence could also be called “technical knowledge”: the line between the two is arbitrary. Now imagine that you have a client who does not understand that transatlantic mail in the eighteenth century took months, that paper was expensive or that letters might be private, familial, corporate or public. For low usage, centuries old record sets, archivists have played the role of mediating cultural expert, explaining the constraints of older systems of records creation, keeping and management. Today’s rapid transformations of technology and increasing access of records over the Internet mean that such essential contextual information must be included in our descriptive and access systems.

Such awareness of the cultural biases of all information objects and systems is not written into RiC-CM, but it is written into another key standard: OAIS.

Figure 2-2 in OAIS succinctly expresses the goal of OAIS, which is not to preserve data objects but information objects. To do this, OAIS starts building from the ground up. Foundational to all is the concept of the Designated Community, for whom the OAIS exists, whose needs are paramount in designing the OAIS. The Designated Community possesses a Knowledge Base, a set of information that is assumed to be universal within the community.  If comprehending an information object requires knowledge that is not part of this Knowledge Base, then the OAIS must deliver the necessary supplementary knowledge through the addition of Representation Information. This concept is illustrated in OAIS 2.2.1 with the example of a hardcopy book written in English. If English is not part of the Knowledge Base for the Designated Community, then the OAIS should furnish appropriate tools (such as a grammar and a dictionary) that would allow a member of the Designated Community to decode the data object and so render it into an information object. Later in the standard (4.2) a more detailed breakdown of Representation Information is presented, including its division into Structure Information and Semantic Information and the incorporation of this information into a Representation Network.

OAIS positions the challenge of preserving information as one of not simply rendering records into a form in which a human can experience them. Rather, in OAIS the goal is always that members of the Designated Community can understand them. In this, OAIS moves beyond most of the work that is done within the archival community, including the then-path breaking “performance model” of digital preservation envisioned by the National Archives of Australia in its Green Paper in 2000. Whereas NAA’s performance model adequately expressed the notion of technological mediation (as does the notion of mediation implicit in RiC-CM), it does not capture the related ideas of cultural mediation. The distinction in OAIS between the Data Object and the Information Object addresses both the challenge of rendering data objects (whether analogue or digital) and of addressing the cultural biases that are, inevitably though sometimes implicitly, part of any information object.

It can be challenging for many archives to translate the concepts and processes described in OAIS, which were developed in relation to a relatively homogeneous Designated Community such as that of space data scientists, to the heterogeneous user communities of most public archives today. Nonetheless, it may be worth examining OAIS and its concept of preserving Information Objects as a way of exploring the true complexity of preserving records and representing them as the culturally and technologically mediated information sources that they are.

References (from Google Scholar – apologies for errors)

Carroll, Laura, Erika Farr, Peter Hornsby, and Ben Ranker. "A comprehensive approach to born-digital archives." Archivaria 72 (2011): 61-92.

Cook, Terry. "Paper Trails: A Study in Northern Records and Northern Administration, 1898-1958." For Purposes of Dominion: Essays in Honour of Morris Zaslow (1989): 13-35.

Galloway, Patricia. "Personal computers, microhistory, and shared authority: Documenting the inventor-early adopter dialectic." IEEE Annals of the History of Computing 33, no. 2 (2011): 60-74.

Heslop, Helen, Simon Davis, and Andrew Wilson. An approach to the preservation of digital records. Canberra: National Archives of Australia, 2002.

Russell, Bill. "Indian Department Headquarters Records, 1844–1861: A Case Study in Recordkeeping and Archival Custody." Archivaria 75 (2013).

Yakel, Elizabeth. "Balancing archival authority with encouraging authentic voices to engage with records.” Theimer ed. A different kind of Web: New connections between archives and our users (2011): 75-101.