📢 Finding Common Ground: Let's Collaborate on Our Data Science Environment

Common Ground.md

📢 Finding Common Ground

Let's Collaborate on Our Data Science Environment

Dear Zone Friends,

I want to thank everyone for the passionate discussion about our Kubeflow environment. The diverse perspectives shared have highlighted important considerations and helped us refine our approach.

Acknowledging Different Perspectives

We've heard valuable feedback about:

• The importance of users learning to manage their own environments
• The technical elegance of minimal base images with only system dependencies
• The long-term vision of dynamic package sourcing
• The value of teaching environment management as a professional skill
• The need to reduce barriers to entry for new users
• The reality that some packages are genuinely difficult to install without system privileges

These are all valid considerations that we're taking into account as we move forward.

Our Collaborative Approach Forward

We believe we can find a middle ground that addresses both technical ideals and practical needs:

1. We're committed to removing packages too!
   Just as we're considering what to add, we want to streamline by removing underutilized packages. We'll be creating a poll to identify which packages we can safely remove without disrupting workflows.

2. Exploring a dual-image approach
   We're working on creating two distinct JupyterLab images to serve different needs:

   • A slim image with minimal packages (no SAS, just the essentials) for users who prefer to build their own customized environment
   • A fully featured image with SAS and popular packages pre-installed to help those who are less familiar with package management

3. Focusing on what users truly cannot install themselves
   Both images will include critical system dependencies (Ubuntu packages) that require sudo access, but will differ in their pre-installed R/Python packages.

What We're Proposing

Based on all feedback, we're planning to:

1. Install critical system dependencies that users cannot install themselves (Cairo, Pango, XML libraries, etc.) in both images

2. Create a poll for package removal to ensure we don't eliminate anything important to your work

3. Develop a more than one image approach:

   • Slim image: Minimal packages, no SAS, for advanced users
   • Fully featured image: Popular packages and SAS included, for easier onboarding

4. Invest in documentation and training to help users build environment management skills

The Zone Team