Sharing code that supports research articles: is it time to align goals and publisher policies?

This BoF aims to understand if there is community interest in the code sharing approaches taken by different publishers (and possibly funders) and if there is a desire and compelling reason to look towards aligning or standardising these at the present time. Code sharing, i.e. openly sharing code that underpins a research article, is related to data sharing in that they are both methods that support reproducibility of research, but is not approached in a consistent manner by publishers or funders, who may or may not include it in data sharing policies. In addition, publishers may also implement code sharing tools, such as code presented in executable forms like a Jupyter notebook or compute capsule. Code sharing can be achieved in a number of different ways and with different goals, for example availability, openness, reproducibility or preservation, which creates differing expectations of the research community.

As both technological and cultural (policy) approaches to code sharing are already being used in the research workflow and scholarly publishing, this BoF session would aim to bring those approaches together to look for common areas of interest that might form the basis of a future RDA group that aims to advance efforts in code sharing.

This session also will explore if there is consensus on where the community should focus efforts to increase code sharing, or if the community - including researchers, publishers, funders and infrastructure providers - is even ready to decide this is an appropriate goal. We would seek thoughts from the community on where efforts should be focused or what evidence is needed to progress work towards community (publishers, funders, researchers who produce and use code, code infrastructure providers, and data managers) alignment when it comes to code sharing. For this session we will discuss source code in a broad sense to encompass the different types of code that can be produced during the research process, e.g. analysis scripts, models, so we are not prescriptive in which areas the community might want to discuss nor are we unduly influencing the outcome of the session. 

The topic of code sharing goals and approaches is not currently served by any RDA group but is closely related to data sharing in certain research fields, if reuse and replication of research are to be fully achieved. The closest related group is the Software Source Code Interest Group, which deals with the mechanics of software sharing (e.g., metadata, archiving, discovery) rather how code is approached from a publication or policy point of view (e.g. should sharing be mandatory, how can different stakeholders facilitate code sharing). Other RDA groups have already looked towards policy issues for research data (e.g., Data policy standardisation and implementation IG and Research Funders and Stakeholders on Open Research and Data Management Policies and Practices IG) and could serve as a model framework for later approaches if there is enough interest in code sharing.