2026-03-22
Git 6cd2c76
First meaningful development
Flashcards, Hive persistence and text-to-speech established the first learning loop.
Engineering case study · Evidence pack
An evidence-backed case study of a Flutter app for learning Facility Management terminology — from its local-first learning core to its cloud and speech boundaries.
Verified timeline
2026-03-22
Git 6cd2c76
Flashcards, Hive persistence and text-to-speech established the first learning loop.
2026-05-19
Git a6b69ed
The local Whisper plan moved to online speech APIs behind the proxy boundary.
2026-05-31
Git tag v1.0.0
The repository recorded its first explicit application release tag.
2026-06-05
Git tag v1.1.0
Global group leaderboard and UI fixes were included in the latest app tag.
2026-09-06
Phase 2 baseline f59332b
Canonicalization and cleanup were completed; Phase 3 adds documentation only.
Actual role
The role covers problem definition, requirements, UX/product decisions, directing AI agents, manual testing, acceptance and release decisions. The repository is not used to infer that the maintainer personally typed or independently designed every line.
Architecture exploration
Refactoring
Automated tests
Debugging & audits
Architecture
Hive, SharedPreferences and bundled JSON serve core learning paths.
Auth, Firestore progress sync and RTDB social data have separate boundaries.
Authenticated proxy, rate limiting and provider fallback for speech/content requests.
Engineering areas
Bundled dictionary and tests are imported into Hive. Lookup, saved words, quizzes, tests and much of progress handling have local paths.
Updates validate IDs, stage a backup, replace the Hive box and restore the previous dictionary if a step fails.
Firebase Auth, Firestore progress sync and RTDB social data serve different paths; the documented boundaries are not presented as a fully authoritative social backend.
Temporary WAV audio goes through the authenticated proxy, while the returned transcript is scored locally by a rule-based scorer.
Claim boundary
FM Dictionary