⚠️ Active development on this version was cut short on April 20, 2026 by a hardware failure on the dev machine, and most in-progress work was lost. What follows is the partial record that survived.
Highlights
- Restructure: the project converges from "a stack of transitional prototypes" into one formal release line, with version axes unified and experiment management reset.
- Innovation: replay is established as the real window into the world — once recorded, an experiment is immutable, self-contained, and replayable for good, not just a debugging tool.
- Improvement: docs are split into three layers (History / Current / Operations) with a machine-readable index, giving people and tools one trustworthy source of truth.
Restructure: from rolling prototype to a formal release line
V0.1 and V0.2 produced a lot of valuable prototype work, but the active codebase still mixed historical versions with current defaults, transitional compatibility layers with formal contracts, and scattered notes with real specs. V1.0's first job was to collapse that into a single formal release line, under one principle: history stays history; the active system gets one clear source of truth.
Innovation: replay as the window into the world
ALICE simulations are headless, so replay isn't a debugging aid — it's the primary way to observe the world. V1.0 makes replay a formal contract: an experiment is immutable once recorded and fully self-contained, with maps, labels, and character art frozen inside the bundle, and the launcher reads the bundle's own metadata instead of guessing compatibility. Replay becomes a durable window into past worlds.
Improvement: layered docs, fewer mixed assumptions
A lesson from the prototype era: mixed docs create mixed assumptions. V1.0 splits docs into historical records, current specs, and operations runbooks, plus a machine-readable index, so humans and agents resolve the right document directly instead of guessing from filenames.
Technical Notes
The baseline anchors all five version axes (world / system / cognition / replay-runtime / replay-bundle) at v1.0; the old prototype village is kept as a historical draft, and the official line moves to uva_village/v1.0. At this stage the map prioritizes structural correctness for cognition work over visual polish.
What's Next
Make the baseline genuinely usable: finish the formal map pipeline, keep cognition ahead of visual polish, and generate new experiments under the reset numbering system.