Abstract-Binder is an open source web service that lets users create sharable, interactive, reproducible environments in the cloud. It is powered by other core projects in the open source ecosystem, including JupyterHub and Kubernetes for managing cloud resources. Binder works with pre-existing workflows in the analytics community, aiming to create interactive versions of repositories that exist on sites like GitHub with minimal extra effort needed. This paper details several of the design decisions and goals that went into the development of the current generation of Binder.Index Terms-cloud computing, reproducibility, binder, mybinder.org, shared computing, accessibility, kubernetes, dev ops, jupyter, jupyterhub, jupyter notebooks, github, publishing, interactivityBinder is a free, open source, and massively publicly available tool for easily creating sharable, interactive, reproducible environments in the cloud.The scientific community is increasingly unified around reproducibility. A survey in 2016 of 1,576 researchers reported that 90% of respondents believed there exists a reproducibility crisis in the scientific community. A majority of respondents also reported difficulty reproducing the work of colleagues [Bak16]. Similar results have been reported in the cell biology community [The] and the machine learning community [Pin17]. Making research reproducible requires pursuing two sub-goals, both of which are difficult to achieve: as well as the "data heavy" approach many fields are adopting, these problems become more complex yet more tractable than ever before.Fortunately, as the problem has grown more complex, the open source community has risen to meet the challenge. Tools for packaging analytics environments into "containers" allow others to re-create the computational environments needed to run analyses and evaluate results. Online communities make it easier to share and discover scientific results. A myriad of open source tools are freely available for doing analytics in open and transparent ways. New paradigms for writing code and displaying results in rich, engaging formats allow results to live next to the prose that explains their purpose.However, manual implementation of this processes is complex, and reproducing the full stack of another person's work is too labor intensive and error-prone for day-to-day use. A recent study of scientific repositories found that citation of "both visualization tools as well as common software packages (such as MATLAB) was a widespread failure" [SSM18]. As a result, the technical barriers limit practical reproducibility. To lower the technical barriers of sharing computational work, we introduce Binder 2.0, a tool that we believe makes reproducibility more practically possible.
