MemoryMachine

system

Breakdex::Identity::MemoryMachine

The scientific side — FSRS scheduling, retention modeling, honest mastery tracking

Breakdex as a memory machine treats breakdancing moves as knowledge to be retained over time. It models human forgetting curves, schedules reviews at optimal intervals, and tracks mastery through three honest states (NEW → LEARNING → MASTERY). No vanity metrics — if you forgot a move, the system says so.

Relationships

→ Depends On 3

FsrsSchedulerFlashcardReviewProgressDashboard

← Reads From 4

fsrs_cards_tablereviews_tableachievements_tablemilestones_table

⚙ Configured By 2

fsrs_configreview_session_provider

Data Access

\u2190 Reads 4

fsrs_cards_tablereviews_tableachievements_tablemilestones_table

Answers

States

StateBehavior
all Mastered
message: All moves at mastery — maintenance reviews will keep them there
shows: celebration
long Absence
message: Cards have decayed — FSRS recalculates intervals based on actual elapsed time
shows: overdue_cards
no Reviews
message: Start reviewing to see your progress — the memory model needs data
shows: empty_progress

Invariants

Mastery states must be honest — derived from FSRS, never manually inflated

Enforced by: fsrs_service

Review history is immutable — snapshots entity identity at review time

Enforced by: reviews_table_snapshots

Every reviewable entity gets an FSRS card on first review, not on creation

Enforced by: fsrs_cards_dao_ensure_card