archivedcreated 2026-03-23· last activity 2026-08-15
New in this cycle: Three coexisting view modes — topology, group-by-tag, geography
Goal
Explore animated transitions when data changes at runtime — actors entering, exiting, and moving between positions as filters are applied.
What's New
SceneReconciler — pure diff engine that compares current scene against a new layout and produces a transition plan (enter/exit/move for actors, rugs, and flows)
TransitionOrchestrator — phased animation sequencer that executes transition plans in 4 phases: EXIT → RUG RESHAPE → MOVE → ENTER, with staggered animations
Transition scenarios — sequential filter snapshots simulating live data changes (progressive filter, scaling events, regrouping)
useTransitionPlayer — composable wiring scenarios to reconciler + orchestrator with play/pause/step controls
Incremental update over clear+redraw — filter changes typically add/remove a few actors, full rebuild would be jarring
Reconciliation pattern — layout engine stays pure, diff logic is a separate layer
Staggered animations — max 15 concurrent, 30ms stagger delay, capped per phase
BaseMesh tweens reused for scale/opacity (battle-tested), animateValue() for position (new)
Built From
C11: Grid Layout — the layout engine, entity system, and composable architecture carry forward unchanged.
Visual record
9 captures — heroes, page slices, and full-page renders. Click to view.
Topology mode (default) Geography mode — actors on the map Group mode at default tags Group mode from above Topology from above Region Migration dataset mid-transition Scale Storm dataset in geography mode Scale Storm grouped — 200 actors in tag rugs