There's a decent amount of redundancy in the raws, and there are a number of ways of dealing with this. The way I had considered before (see #8 and #19) was to add more complexity to the data structure and to the application.
A better way might be to create a different source format for the raws. We might use separate files for each of the lists, and then preprocess them with Jsonnet to generate the JSON output that is ultimately consumed by the application.
I don't know Jsonnet, so I should do some research and ponder how this might be used to:
improve the breadth and depth of spawns
improve the design of the application (i.e. removing complexity that was added to reduce the burden of hand-editing raws)
anything else
Acceptance Criteria
[ ] I've pondered rewriting the raws and preprocessing them as Jsonnet
[ ] A followup issue has been opened to address my findings
Story
There's a decent amount of redundancy in the raws, and there are a number of ways of dealing with this. The way I had considered before (see #8 and #19) was to add more complexity to the data structure and to the application.
A better way might be to create a different source format for the raws. We might use separate files for each of the lists, and then preprocess them with Jsonnet to generate the JSON output that is ultimately consumed by the application.
I don't know Jsonnet, so I should do some research and ponder how this might be used to:
Acceptance Criteria
Implementation Notes