766 B

Raw Blame History

FAQ

Why add course ingestion?

Because many open or user-supplied courses already encode:

topic sequencing
learning objectives
exercises
project prompts
terminology

That makes them strong starting material for draft domain packs.

Why not just embed all course text?

Because Didactopus needs structured artifacts:

concepts
prerequisites
projects
rubrics
mastery cues

A flat embedding store is not enough for mastery planning.

Why avoid PyKE or another heavy rule engine here?

Dependency stability matters. The current rule-policy adapter keeps rules simple, transparent, and dependency-light.

Can the rule layer be replaced later?

Yes. The adapter is designed so a future engine can be plugged in behind the same interface.