#SI5: The State of Digital Humanities Software Development Roundtable
The State of Digital Humanities Software Development
1Carnegie Mellon Unviersity; 2University of Virginia; 3Princeton University; 4Performant Software
In 2013, the University of Virginia Library Scholars’ Lab convened an NEH-funded symposium exploring the often-tacit intellectual labor in the practice of digital humanities software development. Called “Speaking in Code”, this event drew attention to issues faced by coding practitioners in the field. But where do we stand now, five years on? This roundtable will bring together DH software developers from a variety of different roles and professional stages, including in-house developers associated with institutional DH centers, independent consultants, and faculty who produce and publish their own code.
This range of perspectives is crucial for understanding how institutional contexts affect all aspects of DH development, from tool building, to research projects, to teaching programming. Some universities have adopted the “Virginia Model” pioneered by UVA and George Mason, establishing a centralized library DH shop. How does this structure compare to practices around individual faculty/grad-students-as-coders? Or to the the lab-based PI organization with staff research programmers prominent in the sciences? Or programs that contract with independent consultants?
This roundtable will also grapple with the relationship between industry-based “best practices” and DH development. Taking advantage of tools and techniques such as unit testing, code coverage, and pair programming ought to result in better software products with fewer bugs that are easier to maintain. But how well does that correlate with better scholarship? While software developers often strive to develop minimal components that can be used in a variety of applications, many DH projects pride themselves on context-specificity. In which projects do these priorities complement each other, and in which do they conflict? Larger software engineering teams benefit from the ability to do things like code sprints, code review, etc. Can one-person or smaller DH developer shops do the same?
Different organizational contexts also complicate the rhythms of DH development. How do we determine when something is complete, in absence of a publication deadline that requires us to stop refining something? Grant-driven development can be quite different from developing for industry; planning cycles are much longer, and projects can be “resurrected” when a former PI lands a new grant. Goalposts also change, such as when a small product built for one article or class now needs to be deployed as a larger “bulletproof” service.
Our participants will also discuss the personal considerations of doing software engineering for the academy. Participants' training and backgrounds span a wide spectrum from industry and freelancing to humanities departments and self-taught programmers. How does this diversity shape DH development? As DH developer becomes a more established position, how does the academy treat software engineering time differently from researcher or teacher time? And what are the opportunities for professional advancement in software engineering in DH development?
We hope that this roundtable will be of particular interest to software developers in the DH community, broadly defined. We also believe this roundtable will be valuable for non-programmers who want to better understand the mindsets of their colleagues.