Think about "R at work"

[hadley]

e.g.

Workbench ultimately solves three core problems:

• Problems of Scale — the most boring problem.

  • The desktop is not powerful enough compared to a server.

  • Workbench provides simple access to large server or elastic compute via Kubernetes/Slurm and other HPC tooling.

• Problems of Security/Access — R at work problems.

  • There may be sensitive data that is not accessible via desktop IDEs — so they need to use a server to access.

  • Workbench provides a secure environment with SSO and managed credentials for seamless access to Cloud data without PATs.

  • The authentication and authorization problems for full Identity Access Management.

• Problems of Collaboration — humans are messy.

  • Workbench provides a standardized way to update versions of R/Python across all users and standardizes more of the environment between users and production destinations.

  • Can also include Docker-backed sessions in Kubernetes/Slurm for very specific but flexible environment management.

  • Environment management on desktop is too messy, and they’re having issues operationalizing or productionizing their work and want to centralize/standardize environments.