[Question]: Handling replacement of geographies

[bhargav-lime]

In the geographies schema, I can see that a geography can replace previous geographies. Is that replacement effective immediately? The Spec verbiage is not clear on this one.

This can help me determine which geographies will be used to compute data and prepare the response.

[schnuerle]

I think it's effective immediately. Asking @marie-x for their opinion.

[marie-x]

As currently written, yes, it's effective immediately. Do you think we should have publish_date and a separate effective_date like Policy does?

EDIT:

If there's no effective_date given, then the Geography is effective as of the publish_date which would in practice mean "immediately." Otherwise the replacement isn't effective until effective_date.

[schnuerle]

As currently written, yes, it's effective immediately. Do you think we should have publish_date and a separate effective_date like Policy does?

Those two fields are in there now for Geographies.

Maybe all we need is a line saying geographies referencing others with prev_geographies immediately replace the previous ones.

[marie-x]

Updated my comment. I think maybe we can add a line of clarification to the spec.

[schnuerle]

If anyone has an idea of how to word that sentence and where to put it please share.

[schnuerle]

We added some clarification for this in a new "Updating or Ending Policies" section in Policy. Maybe we could do something similar for Geography. But our upcoming OpenAPI work should resolve the schema issue.

I added a clarifying line "geographies referencing others with prev_geographies immediately replace the previous ones" to the field definition.