Schema transition

Technical elements of transition

The Github repo has been transitioned to Travis CI for creation of the web pages.

• The Github repo needs a Census-centric home
• Travis CI needs to be enabled for the Census repo
• The current repo needs to be cleaned up
  • "formats" -> schema naming 101
  • clean separation of content and formatting where possible 101
  • clean separation of documentation and processing thereof where possible (in particular the shell scripts) 100

Production inputs

Production inputs that are important pre-processes to the Schema

• should be identified
• should be separated from the workflow if needed
• should be identified inputs to the Schema

[production geo] -> [schema] -> [rest of production]

Processing of schema

The creation of the schema should remain a discrete component. It probably does not make sense to try and shoehorn it into the production process.

Use by Production

• The creation of the schema and the use by the production system should be cleanly separated.
• The parts of the production system that are currently inputs to the schema should not feed directly into any other production components (functional separation)
• The production process should always use the latest schema.
• There might be a query to the Schema maintainers if any changes are to be published.

Medium term activities

See Issues

• PSEO
• LODES
• Agg level

Governance of the schema

Who decides on iteration, numbering, etc.?

• Versioning is defined by VERSIONING.md. It only needs to be interpreted.
• Iteration is defined on-demand - availability of geo, upstream changes to definitions.