Skip to content

Deep dive into governance models #9

@diegoripley

Description

@diegoripley

Establishing a Governance Model for Data for Canada

Background

As the Data for Canada project expands, it is important to formalize a governance structure that ensures long-term sustainability and high-quality output. This discussion is inspired by the Evolving Zarr’s Governance update.

Governance Philosophy: The Middle Ground

The intent for this project is to avoid a Benevolent Dictator for Life (BDFL) model. However, maintaining high standards is paramount to prevent "sloppy code" or poor design choices from entering the disseminated data packages.

There is significant value in the model of ultimate authority seen in the Linux kernel—where strong technical opinions ensure the integrity of the ecosystem. We are looking for a middle ground: a structure that encourages community innovation while maintaining a strict "quality gate" managed by project maintainers.

Proposed Contribution Workflow

I am envisioning a "Curation-First" model where the community drives experimentation and maintainers ensure professional-grade execution.

  1. Ideation & Experimentation: Contributors are encouraged to explore new ideas, experiment with data structures, and propose additions.
  2. Maintainer Review & Refinement: Once a concept is proven, project maintainers will take the lead on "cleaning up" the contribution. This ensures that every package brought into the main project meets our rigorous standards.
  3. Standardization: Maintainers will be responsible for ensuring that all data fits into the "4th dimension" and adheres to the project's requirements for modern file formats and metadata.

Discussion Points

  • How can we best define the boundary between community experimentation and official maintainer refinement?
  • What specific standards should be mandatory before a data package is considered "disseminated"?
  • How do we ensure the maintainer role remains a facilitator of quality rather than a bottleneck for innovation?

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

Milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions