Do we need course-specific schemas?

[chris-steele]

Feature description

Is there a compelling argument for compiling schemas on a per-course basis? If we dropped this behaviour we could use the schema cache much more easily: we would only need to recompile schemas when plugins were added or removed from the tool. This would provide a significant performance boost as currently there is a large amount of redundant schema compilation.

Can you work on this feature?

• I can contribute

[taylortom]

Not sure I understand the angle of this issue - are you suggesting adding a course-specific schema cache? - if so, I think this is a good idea.

Currently we have no schema caching for any content schemas. The reason for this being that each course has its own custom content schemas which vary based on the Adapt plugins installed on that course.

[chris-steele]

Yeah I was thinking about a per-course cache initially...

I went on to wonder whether we could scrap the plugin toggle for courses and just enable all plugins for each course and therefore have a single set of compiled schemas. You'd control which plugins were enabled in a given course via the usual _isEnabled property. Bit of a wild departure, I know.

As an aside I really wish we could do something about the JSON being bloated with defaults. Sometimes a single block can be 200+ lines :/