Skip to content
ulamsulamsulams docs

0062. Adaptive interface as a presentation profile over catalogue components

Generated from docs/decisions/0062-adaptive-interface-profile.md

  • Status: Proposed
  • Date: 2026-10-09
  • Plan: docs/plans/phase-4.md (section 11.4)

Context and problem statement

Section titled “Context and problem statement”

Phase 2.7 asks for interface-level personalisation (density, chunk size, navigation, visible hints) driven by Learner Insights. It must stay within catalogue components and theme constraints and be measured. This was moved to Phase 4 (pending #59).

Considered options

Section titled “Considered options”
  1. A rule-derived presentation profile per learner and course, applied only through catalogue props. The learner can override it. Behind a flag, measured by an A/B experiment.
  2. Model-generated layouts per learner.

Decision

Section titled “Decision”

Option 1:

  • Profile fields. density, chunking, navigation and hints, derived from statuses by rules and stored in learner_presentation_profiles.
  • Application. Prose is split into Steps, plus initial hint counts and the outline mode. No model call and no model-written UI.
  • Flag. learner_insights.adaptive_interface, default off.
  • Measurement. An experiment with delayed retention, SUS/UEQ-S and NASA-TLX.

Consequences

Section titled “Consequences”
  • Good: predictable, accessible, free to run and easy to explain.
  • Bad: less expressive than generated layouts; generated remediations cover the content side.