Developing technologies: devising strategies for the adoption of e-Infrastructure
By Zack Kertcher, University of Chicago; Ralph Schroeder, University of Oxford; Franz Barjak, University of Applied Sciences Northwestern Switzerland; Simon Robinson & Tobias Hüsing, Empirica, Germany
Hundreds of projects from dozens of countries are pursuing the development of e-Infrastructures, yet these technologies are still not as simple as using a cell phone or surfing the Net. As part of the European Commission funded eResearch2020, we researched e-Infrastructure adoption barriers and strategies, aiming to identify ways to enhance technology adoption and facilitate the creation of global virtual research communities. This article reports our key findings.
The Diverse technological terrain
e-Infrastructures span continents, scientific and professional practices, and technologies. To effectively analyse such a heterogeneous population, we selected 18 case studies, based on their development scope, geographic range, maturity, and number of participating organizations. Using interviews and archival analysis, we identified common themes and, furnished with results of a qualitative cross-case comparison, designed a survey. We first surveyed relevant users and developers, then distributed the survey further to persons who had participated or shown interest in e-Infrastructurerelated activities. Over 400 individuals completed the online questionnaire. This multi-stage design afforded a wide-ranging perspective on the development and adoption challenges of e-Infrastructure technologies.
Development and adoption challenges
Pointing to cultural differences between developers and users, our respondents reported some challenges at the early development stages. These difficulties manifest when developers have little understanding of specialized user practices, resulting in communication problems and the pursuit of divergent objectives (such as when developers aim to work on cutting-edge technologies, while users seek basic and robust services). Respondents also noted barriers in reaching out to new users. Frequently mentioned challenges were users’ negative attitude toward computer-enhanced research environments and a reluctance to spend the time and resources required to learn the new technology. Other adoption barriers were that e-Infrastructures had insufficient technical capabilities, as well as low usability and organizational barriers. Despite these challenges, which cut across our sample, respondents generally lacked knowledge of adoption barriers, or detailed information about their users. One reason for this is that some infrastructures do not distinguish between individuals and organizations. Users also often register with an infrastructure, but some infrastructures only monitor access to their website, wiki or portal, but not actual use.
Strategies that work
e-Infrastructure projects commonly accommodate cultural differences between developers and users by improving communication channels: conducting routine meetings and telephone conferences to aid the development of common ground. The cases we studied employed a variety of strategies to enhance user adoption. We distinguished between passive strategies, involving limited user engagement, and active strategies, which focused on ongoing interaction with users. All cases recruited users by disseminating information to potentially interested organizations and individuals, and by presenting at conferences of potentially interested user communities. These more passive user recruitment methods were limited in scope. Several projects had also ventured into more active recruitment. An effective vehicle was the use of “engagement teams” to work with leading users in diverse communities.
“Brokering,” the use of key individuals and organizations specializing in e-Infrastructure and a user domain, was another successful method for canvassing adopters. This had been used by a number of European and US projects. For example, the US TeraGrid ran a program in which “campus champions” served as institutional mediators for recruiting users and as local technical experts. The European D4Science and the US-based Open Science Grid used third-party organizations offering e-Infrastructure technology, along with other services, to dedicated user communities. Respondents suggested that more active user recruitment methods build a communication channel between e-Infrastructure stakeholders, sensitizing developers to users’ needs and to potential adopters, and helping adopters derive more benefit for their research. Another strategy for enhancing adoption involved reducing the cost of learning the new technology. Passive approaches shared by the infrastructures were refining documentation, using wikis and Web 2.0 mechanisms, and running training workshops.
Like with user recruitment, active cost reduction strategies appeared to be advantageous. Relying on brokers, some projects had achieved good results by obtaining knowledge about targeted adopter communities and then designing virtual environments that simulate their users’ typical computational environment, for example, through domain-specific portals. Another approach involved masking e-Infrastructure complexity by relying on third parties to provide specialized virtual technologies supported at the back end. A small minority of e-Infrastructures was pursuing this type of brokerage, but it may offer considerable benefits in the long run. After all, mostly due to its complexity, only a small number of early adopters used the Web prior to the introduction of a broker called a “Web browser.” For further information about e-Research 2020, including its findings, please see http://www.eresearch2020.eu/
