What computational scientists need to know about intellectual property law: A primer

Data and code are becoming as important to research dissemination as the traditional manuscript. For computational science, the evidence is clear: it is typically impossible to verify scientific claims without access to the code and data that generated published findings. Gentleman and Lang [1] introduced the notion of the “research compendium” as the unit of scholarly communication, a triple including the explanatory narrative, the code, and the data used in deriving the results. One of the reasons for including the code and data is to facilitate the production of really reproducible research, a phrase coined by Jon Claerbout in 1991† to mean research results that can be regenerated from the available code and data. Claerbout's approach was paraphrased by Donoho and Buckheit [2] as follows: Enabling computational replication typically means supplying the data, software, and scripts, including all parameter settings, which produced the results [3,4]. This approach runs headlong and unavoidably into current intellectual property law, which creates a stumbling block rather than an impassable barrier to the dissemination of really reproducible research. In this chapter, I describe these intellectual property stumbling blocks to the open sharing of computational scientific knowledge and present solutions that coincide with long-standing scientific norms. In Section 12.2, I motivate scientific communication as a narrative with a twofold purpose: to communicate the importance of the findings within the larger scientific context and to provide sufficient information that the results may be verified by others in the field. Sections 12.3 and 12.4 then discuss intellectual property barriers and solutions that enable code and data sharing, respectively. Each of these three research outputs, the research article, the code, and the data, requires different legal analyses and action in the scientific context as described in the following. The final section discusses citation for digital scholarly output, focusing on code and data.