Two Socio-Technical Gaps of Cyberinfrastructure Development and Implementation for Data-Intensive Collaboration and Computational Simulation in Early e- Science Projects in the U.S

This position paper explores two socio-technical gaps that hamper effective synergy between human and machine intelligence as well as data-intensive and cognitively complex sense making and decision making in early U.S. escience projects. Based on qualitative methods guided by grounded theory and 70 interviews, we identify limiting constraints at the levels of scientists, technologists, science funders, and technology quality, under the participatory and bespoke natures of cyberinfrastructure. More specifically, scientists cannot fully envision the new computational tools they need for their science (i.e., human intelligence in domain science), and technologists have limited knowledge of the domain science they are building cyberinfrastructure/computational tools for (i.e., representing the machine intelligence built into computational tools). Furthermore, science funders offer limited direct funding for cyberinfrastructure development, and this development is time consuming and it produces unstable software (i.e., thus affecting the quality and development of machine intelligence). A closer examination of these key limitations led to the identification of two gaps: the specialization-synergy gap between scientists and technologists, as well as the science investment-technology quality gap between funders and development. We argue that these two gaps hamper the effective cognitively-complex sense making and decision making in data-intensive e-science projects.