General Approach Questions
LTI App Catalog:
Approach questions:
What’s the approach for storing and using content from alternate formats like LTI or H5P?
The current approach is to use the XBlock API as an adapter layer for any such external formats. There is an H5P XBlock (third party) and an LTI Consumer XBlock (built-in).
How are we thinking of the differences between raw blocks and configured blocks? Where is configuration stored? When is it applied?
Known contingencies: Completion of global LTI configuration
Adaptive Engines/Services
Approach questions:
Is there a “gold standard” adaptive engine we can aim to support as an example? Market demand for particular engines or use cases?
Possibly two categories we want to support. One would be developed by tinkerers in a lab at a university, Alosi Labs. One would be highly proprietary and defensive of their IP, Newton. How we support this should be completely open via APIs and, probably, message oriented integration patterns. We should stay arms, or greater length, away from anyone's IP. Using our data for training is a separate question.
What does it mean to have dynamic control over content, and how can we go beyond the small “window of adaptivity” (only an individual component changes) currently possible?
Known contingencies:
Completion of descriptive metadata bit of Phase 2