The 2014 annual meeting of the American Council of Learned Societies (ACLS) took place on 8–10 May at the Sheraton Society Hill Hotel, Philadelphia, PA. Although the continuing existence and entire activity of the ACLS supports all of us in our lives as scholars and teachers, I have noticed that every year's meeting brings some particular issue to the forefront of my attention and I think reporting on that in some detail, briefly at the business lunch and at greater length in my report for Speculum, is the most valuable service I can give the Academy as your delegate. Each year a few scholars report on their ACLS-supported research. The work being done by Stephen Berry, Professor of History at the University of Georgia, supported by an ACLS Digital Innovation Fellowship, is exceptional for its intrinsic fascination and originality. Together with Professor Claudio Saunt, he founded the University of Georgia's Center for Virtual History whose aim is to produce historical research projects in entirely digital formats on interactive websites that involve “‘citizen historians’ in the amassing and analyzing of historical data,” breaking down the expert/amateur hierarchy and opening the construction of history to a broader public, “and the resulting ‘citizen history,’ we [they] believe, better reflects the way knowledge is created and consumed in our increasingly digital world” (quoted from the University of Georgia's eHistory website).
Stephen Berry's project is called CSI Dixie: a project based on coroners’ inquests from South Carolina between 1840 and 1880, proceedings that used the testimony of women and slaves to investigate deaths connected to spousal and child abuse, master-slave murder, and violence among slaves, offering an extraordinary entry into life and death in the Old South. CSI:Dixie and other ongoing eHistory projects at the University of Georgia will be “published” only as very complex websites with advanced interactive features, with the ambition to be kept open and ongoing as contributors other than the original author/researcher offer further discoveries from inquest records. All the eHistory researchers are committed to “the belief that new technologies make possible a new kind of research in the humanities, one in which students, scholars, and a broader public are full partners and collaborators.” This “citizen history” is a notably idealistic application of digital technology driven by democratic collaborative values in the service of deepening and enlivening the creation of history.
During the question period, someone from the audience raised the issue of the durability and stability of website “publication,” but as usual time was short and the tough questions have to be left for later consideration. This is one tough question I have been aware of and troubled by for some time, but thought is nearly impossible on an issue that seems to concern next to no one and is virtually never brought forward in academic or public forums, or in terms we can understand. The issue of the endurance of digital information concerns us all—medievalists have been frequent winners of the ACLS digital innovations awards. Editing the Sage Handbook of Historical Theory, with Sarah Foot, brought me a chapter on digital information by Valerie Johnson, Head of Research, and David Thomas, Director of Technology at the National Archives of the United Kingdom. Their discussion of threats to digital sustainability should be read by everyone for whom scholarship and ephemera are not compatible ideas.
Johnson and Thomas point to crucial but seldom adequately discussed issues that face every digital publication, whether it is a relatively simple article or book uploaded to a website or an elaborate digitally created project with multiply layered information and interactive features: how long will it really last? I do not notice that the permanence of digital documents is ever placed in the foreground of considerations when scholars are invited or encouraged to produce work for digital-only publication. We are so accustomed to the astonishing longevity of the published book or journal—they are not always easy to obtain, but once published and shelved in a library they are there in near permanence. Anticipating that digital letters on the page will metaphorically disappear or turn into undecipherable gibberish, and fairly soon, is not a thought that occurs naturally to people who know that a parchment codex from the eighth century is good for another millennium. Yet it turns out that digital materials can have a startlingly brief shelf life.
Johnson and Thomas point out two significant threats to digital materials: financial and technological:
Large funding bodies insist that grantees have to guarantee that their material will be available for, typically, seven years. Guarantees of the long-term availability of materials given with great enthusiasm and genuine commitment when the grant application is being completed can ring a bit hollow seven years down the line when the money has long since run out. The National Archives of the UK has recently been approached by one large grant-funded project for which the funding has now been used up and the team is dispersing. The organizers would now like The National Archives to take responsibility for the resources they have created. In fact, the funding model for academic digitisation is contributing to the problem: it is relatively easy to obtain funding for a new, exciting project which pushes the boundaries of scholarship and creates significant digital resources. There is simply no source of funding to maintain those resources once they have been created.
The digital longevity track record is dismal. Quoting again from Johnson and Thomas:
Analysis of some statistics reveals the scale of the problem. Of the 155 projects funded by the New Opportunities Fund in the UK between 1998 and 2003 at a cost of £55 million, 25 can no longer be found, while there have been no changes or enhancements to a further 83. Of the 155, there are only 30 which have been enhanced or added to since the launch. So in less than 10 years, 16 per cent of resources have been lost and 53 per cent have, at best, stagnated.1
Commercial organizations that fund digital projects may go out of business, and organizations change their policies about ongoing projects. Johnson and Thomas cite Google's major plan, begun in 2008, to digitize newspapers, which was stopped in 2011 when Google announced that the site would no longer accept any more microfilm or digital files.
The sustainability problem is not a theoretical computing issue—it is a live problem now. A 2004 University of Illinois study examined website citations in three top online journals and found that about half of the URLs cited in their articles no longer pointed to the authors’ source material. Web addresses have become so unreliable that the Modern Language Association recently stopped requiring authors to cite URLs when referencing web-based resources.2
But the real existential threat to digital information is technological. In instances where a digital project has to be transferred to an institution different from the one where it was created, technology problems arise. “This produces particular difficulties when images are stored and delivered using complex proprietary software. Even if projects are created with open-source software, the survival of the system itself is subject to instability and disappearance of support.”
A BBC online article (13 February 2015) reported on a speech made at the annual meeting of the American Association for the Advancement of Science by Vint Cerf, an internationally celebrated internet pioneer, now in his seventies, regarded as one of the two “fathers of the internet” for the design of its basic communication protocols, who says bluntly that “all the images and documents we have been saving on computers will eventually be lost” as hardware and software become obsolete. He anticipates entering a “digital Dark Age” (so sadly ironic, thinking of those indestructible parchments) eradicating documents stored on hard drives or in the so-called Cloud, as accelerating change in basic formats makes information unreadable and “backwards compatibility” is uncertain. Mr. Cerf anticipates vast digital archives of unreadable content.
All of us know this perfectly well because we are among the statistically minuscule cohort of computer users who create content (from our files of reading notes and images to articles and large projects) that we want and need to keep currently legible permanently. The vast majority of computer users only need to access content created by other people and having to “upgrade,” that is, change, their machinery and software to do that is no more than a mild nuisance. Scholars have known for a long time about the ominous threat of digital self-erasure that Mr. Cerf is trying to warn others about. Every operating system “upgrade” I am forced to install presents the threat of losing my document files to backwards incompatibility; even little applications I like whose makers are not rewriting them for the next-generation operating system show me in small what may well happen on a massive scale.
I had finished the first draft of this report when a sad example came my way. I was browsing one of our oldest go-to resources for undergraduate teaching, the Internet Medieval Sourcebook, housed at the Fordham University Center for Medieval Studies, looking for a few primary sources to download or link to my course website, and found that document links I’m sure I had used before without any problem now led to blank screens. The home page explains this, noting that the project, begun in 1996, is old by internet terms and that “at the time it was instigated (1996), it was not clear that web sites [and the documents made available there] would often turn out to be transient. As a result there is a process called ‘link rot’—which means that a ‘broken link’ is a result of someone having taken down a web page.” Or websites have been altered without forwarding links being created. “Since 2000, very few links to external sites have been made. An effort is under way to remove bad links.” But I have to note that the date on the home page for the last modifications is Nov. 4, 2011. “Transient” is not a term we are accustomed to using for published medieval sources, and how many of us even know the term, “link rot,” self-explanatory as it is?
I’m glad that someone of Vint Cerf's stature is paying attention. He is promoting an ambitious and idealistic plan to preserve every piece of software and hardware in digital perpetuity, content and application and operating system and everything together made permanently and universally reconstructible by means of some not yet invented standardized descriptions. He calls this concept “digital vellum”—the irony and pathos of which term are fully graspable only by the people in this room. There is a project called Olive Archive at Carnegie Mellon working on it. I was heartened to find the report of Mr. Cerf's speech, to see that someone high in the ranks of masters of our computerized universe recognizes the threat of cultural oblivion contained in volatile nonstandardized information technologies. Maybe computer people will pay attention to someone of Mr. Cerf's importance, but before entire generations of scholarship are willingly and naively entrusted to digital-only format, we should be paying attention now.
Report on the ACLS Fellowship Programs
In the last competition, medievalists won the following awards:
- ACLS Fellows—2.
- Collaborative Research Fellows—2.
- New Faculty Fellows—1.
- Charles A. Ryskamp Research Fellows—1.
- ACLS/Mellon Dissertation Completion Fellows—5.
The Charles Homer Haskins Prize Lecture
The 2014 Charles Homer Haskins Prize Lecture was delivered by Bruno Nettl, Professor Emeritus of Music and Anthropology at the University of Illinois, Urbana-Champaign.