In the digitally transformed college or university, IT work will thrive for technology teams that develop and leverage the powers of persuasion and leadership.
Alfréd Rényi (not Paul Erdős) said, "A mathematician is a device for turning coffee into theorems." Coffee also fuels conversations about digital transformation.
My colleague Charley Kneifel and I were describing the layout and rationale of Duke's research computing services to overseas visitors, and the following question arose: "How do you get researchers happily using centralized resources?"
"Well, it involves large quantities of coffee," Charley said as glanced over to me. Charley was not being flippant, or at least entirely so. He was right. In my role as the director of research computing, I've spent a fair amount of time in Duke's various coffee establishments and have come to realize that my earliest uttered and most repeated research computing question is, "When can we meet for coffee?" The coffee cup is the tool I grab as I do my work in digital transformation. It is good to remind ourselves that IT work—especially in the digitally transformed college or university—will thrive for technology teams that can harness the powers of persuasion and leadership. All-too-human features also matter when it comes to using bits and bytes for science and scholarship. It is easy for people involved in information technology to overlook them.
Launches and Landings
The pairing of infrastructure and what might be called "fitment" that is shown in Charley's role (he leads technical infrastructure development and maintenance) and my role strengthen research at Duke. These two IT roles do, in fact, blend, though each has a distinct emphasis: one creates and tends to research services and infrastructure, whether on premises or in the cloud; the other faces researchers and their distinctive projects and seeks to respond with services that fit their aims. One role is concerned with "launches"; the other is concerned with "landings." Both roles reflect a challenge of digital transformation as IT roles shift, blend, and expand. In effect, information technology is assuming a broader role in academe—one that is deeply technology-focused but is also profoundly directed toward the activities of individuals who make up academic life: students, faculty, researchers, alumni, staff, and communities. While technology professionals have become quite good at technology "launches," our technology focus will increasingly become more centered on "landings"—the actual use of services—and on people's experience with them. Professional, scholarly, and educational benefits will become more pressing and emerge as the whole point of our efforts in information technology—a shift in emphasis from the hard work of keeping the infrastructure running and maintaining the engines of administration and education that characterize many of yesterday's and today's IT operations. (It could be argued that the benefits of academic information technology have always included the educational and professional, but the habits of many IT professionals might suggest otherwise.) IT work, in effect, spills from the data center into labs, classrooms, dorms, and offices, where people labor, learn, and exchange ideas over cups of coffee.
The kind of landing that successfully and productively applies digital tools to research aims is, of course, complex, and even emotionally fraught since multiple hands control the descent and the touchdown of the research craft, so to speak. To guide researchers through the process, IT professionals will have to know the research terrain as well as the technological capacities that have long been at the center of IT organizations. We will have to understand, sometimes with depth and sophistication, the research processes and analytics that naturally vary in method and in scope from field to field.
Scalability and Nonscalability
That variation of research across fields is indeed a big part of the challenge of discovering pathways toward digital transformation. Variation complicates scaling, and much of the expansion of centrally managed research computing has been driven by the desire to scale up and make infrastructure more expansive and efficient. Indeed, technologies are often evaluated by their scalability, and a primary activity in devising a scalable product is, of course, removing (or at least reducing) variation. Herein lurks a conflict. In her essay "On Nonscalability: The Living World Is Not Amenable to Precision-Nested Scales," Anna Tsing rather categorically points out, "Scalable projects," broadly speaking, "are those that can expand without changing. . . . Scalable projects banish meaningful diversity, which is to say, diversity that might change things."1 IT scalability's complaint, naturally, is that academic fields and their research projects are diverse and have differing IT requirements, and researchers work from widely varied perspectives and with different levels of computational skill. Researchers' diverse projects threaten the digital uniformity needed to scale an infrastructure.
Over a cup of coffee, I often plumb the resonance and the dissonance of researchers' desires and needs against the backdrop of "what might easily be done in the infrastructure"—the product of Duke's rather remarkable and thoughtfully scaled infrastructure. Sometimes, as in the case of researchers using genomics data, the match is made as easily as falling off a log. Other times, not so much. One physicist pondered a wiki page full of installation incantations that his international research community prescribed for server setup. The list made his machine unique, at least in configuration, but a centralized virtual machine can still do the trick. The reason for the successes (and partial successes) of a centralized virtual machine has been Duke's extensively virtualized research computing infrastructure and the reduction to CPU-core, RAM, and storage as basic, modular "raw materials."
Modularized building from such raw materials allows the idiosyncratic grammar of a research workflow and the diverse peculiarities of research habits to be reflected in IT systems. For the most part.
Differences between the needs for scalability and the facts of nonscalability cause the friction found in centralized (i.e., scaled) research computing as it is applied to research questions, and the friction is likely a reason why some researchers view centralized research computing resources skeptically. To surmount that skepticism and ease the friction requires thoughtful negotiations between the technical virtue of scalability and the diverse virtues of novel and innovative research. Charley's comment about the abundance of coffee points to a method of initiating transformative relationships, which are "the medium for the emergence of diversity."2 Negotiation and exchange accommodate a need for IT economies of scale and help to match research tasks with tailored computing resources. Coffee helps (dark Brazilian, if you please).
Know Your Field (and Everyone Else's)
Digital transformation changes research behavior and the behaviors of IT professionals. "Researchers will need to learn new research techniques, adopt a research approach that includes students . . . and collaborate more frequently and deeply with other faculty and with other professionals," write EDUCAUSE's Susan Grajek and Betsy Reinitz. "There will be a blurring and re-forming of academic disciplines."3 Technology is a powerful agent of this change, but as most people know, that "blurring and reforming" may not be the most welcome development in many fields. "Information technology" skills that will come to bear during digital transformation will take on more of a diplomatic than a technological hue.
The reason? In academic fields, transformation arises from practitioners, not from the IT guy. Art history, for example, might be digitally transformed by its professors, many of whom have already embraced digital visualization tools in their teaching, but the adoption of that technology might be only modestly transformative since its use might be simply to replace old instructional tools like projectors and slide stacks with tools like PowerPoint or Photoshop. More deeply transformative changes, such as those that employ machine learning or "big data," certainly would affect art historians' research practices, making typical and traditional practices more efficient or opening doors that previously were closed or even invisible to scholars. Still, such changes can and do threaten existing practices and structures within academic fields, as well as the people who use them, because new practices can (and probably will) interrupt, re-scope, and perhaps even redefine the processes of scholarship and the competencies required in a field of study. These are not trivial matters.
Uncovering opportunities for transformation requires an understanding of a field's language and habits. It means delivering a message that not only is understandable but also is consonant with the aims of a field and could be embraced by its current practitioners. Knowing a field's practices, rehearsing them if possible, and experiencing them from the inside out are essential elements of digital transformation—especially in some fields with firmly rooted habits and traditions. In effect, to be a better IT professional, you must acquire an identity as a research colleague, someone who is knowledgeable about a field and its aims and who understands its history and values.
Questions raised over a cup of coffee, of course, almost always have to do with outfitting research information technology so that the technology serves the ends of research. The answers that arise from conversations reveal different paths toward digital transformation. You have to drink a lot of coffee, have a lot of exchanges, and have deep research involvement.
And so, the coffee shop will continue to benefit from my patronage until the application of computing technologies becomes a normal practice of conducting research. Such practices do not emerge from information technology alone; they happen as a result of how scholars and scientists choose their studies, what they count as their data, and how they move toward their conclusions and results. Paradoxically, a cultural transformation like that will indeed "scale up" the digital habits of and scholarly conversations within transformed academic fields.
Only then, I might have to find other things to talk about over a cup of coffee.
- Anna Lowenhaupt Tsing, "On Nonscalability: The Living World Is Not Amenable to Precision-Nested Scales," Common Knowledge 25, no. 1–3 (April 2019): 143–162. ↩
- Tsing, "On Nonscalability," 143–162. ↩
- Susan Grajek and Betsy Reinitz, "Getting Ready for Digital Transformation: Change Your Culture, Workforce, and Technology," EDUCAUSE Review, July 8, 2019. ↩
Mark DeLong is the Director of Duke Research Computing at Duke University.
© 2020 Mark DeLong. The text of this work is licensed under a Creative Commons BY-SA 4.0 International License.