Transitive Credit as a Means to Address Social and Technological Concerns Stemming from Citation and Attribution of Digital Products

Assessing and tracing the outcomes and impact of research infrastructures

“…Whereas many metrics have been developed for measuring the reliability and performance of software applications (e.g., Fenton & Bieman, 2014), relatively few have attempted to measure the impact of software used in science. This is problematic because the sustainable development and maintenance of this type of research infrastructure depends, in part, on having reliable documentation about how software is used and having ways to measure and clearly communicate that use (Ahalt et al, 2015;Katz, 2014;Katz et al, 2014). Much like research data, the infrastructure and cultural norms that support traditional metric-based approaches to impact, such as citation analysis, are emergent and not well established (Jackson, 2012).…”

“…What Metrics Can Be Commonly or Practically Used, Have Validity, and Can Be Commonly Reported for Research Infrastructures? Can New Measuring Schemes (such as "Transitive Credit" Presented by Katz, 2014) Lead to More-Nuanced Views of the Impact of Individuals and Particular Research Infrastructures?…”

Summary of the First Workshop on Sustainable Software for Science: Practice and Experiences (WSSSPE1)

“…A recurring theme of the panel discussion was that software work in science is inadequately visible within the reputation system underlying science; in other words it often doesn't "count". In his paper for this workshop, Katz placed software work along with other "activities that facilitate science but are not currently rewarded or recognized" [49]. Priem and Piwowar argued for the need to "support all researchers in presenting meaningful impact evidence in tenure, promotion, and funding applications."…”

“…It was noted that there appear to be no widely accepted standards on how the use of software towards a paper ought to be mentioned, and that journals, citation style guides and other guides to scientific conduct are vague about how to describe software. To address this, papers advocated the need for a fixed identifier for software, either directly through a mechanism such as a Digital Object Identifier [49,51] or via a published paper written to document the software (and perhaps its creation), a "software paper" [55]. However, as was pointed out during the panel discussion, one of the problems with papers as the cited product is that their author list is fixed in time, which discourages potential contributors who are not on the original author list from designing incremental improvements as integration work rather than separate (and hence possibly rewritten) software products [52].…”

“…Katz approaches this issue by proposing a system of "transitive credit," recording dependencies and relative contributions outside particular papers, relieving authors from the responsibility of acknowledging each and every dependency. Instead authors would acknowledge the percentage contribution of the software they used directly and an external system would then be able to recursively allocate that credit to those who had provided dependencies [49]. Finally Priem and Piwowar argued that machine learning techniques could examine the body of published literature and extract mentions of software, coping with the multitude of informal ways in which authors mention software they used [50].…”