Nindova v0.3.0 — Shahi Mound

Nindova v0.3.0 — Shahi Mound

Nindova v0.3.0 gives Masala Mound a richer identity and a structurally tighter Deeper opening while keeping the Session bounded, local, and free of performance judgment.

Added

• The paired-diamond Phulkari lattice, drawn Nindova wordmark, Shahi Raat palette, original utensil silhouettes, PWA icons, social card, and complete CC0-1.0 asset provenance.
• A qualitative end reflection: Gentle names its woven layers; Deeper names the opened triple crown. Neither estimates intelligence, speed, ability, or improvement.
• One optional deterministic three-note chime and a paired brass-and-peacock match bloom, with sound off by default and an opacity-only reduced-motion alternative.

Improved

• Deeper now opens through three staged crown matches. Each step presents unmatched free decoys around one legal pair; exhaustive verification covers 510 reachable Deeper states, one terminal state, zero dead states, and no losing legal choice. Gentle retains its proven 382-state geometry.
• The Session cards, board frame, layer treatment, landing page, README, documentation, install icons, and public imagery now share the same royal-night system.
• The standalone file remains self-contained. The installable PWA precaches only its local static shell and icon set.

Evidence boundary

• Gentle and Deeper are equal tonight-only preferences. There is no IQ estimate, brain-speed judgment, score, rank, timer, streak, grade, achievement, history, or randomized reward.
• Sensory feedback is described by its observable behavior. Nindova does not claim it creates dopamine, ASMR, cognitive gains, or better sleep.
• Browser automation covers the standalone and PWA independently, the complete Deeper path, keyboard operation, reduced motion, 200% zoom, responsive phone and desktop surfaces, offline closure, same-origin requests, and the real production ceiling.
• The crown-search proxy and exhaustive solver prove board structure, not a human difficulty rating or a typical fifteen-minute play time. Those remain untested.
• Human Punjabi cultural-authenticity review, real-device VoiceOver/TalkBack acceptance, broader installed Safari/Android proof, and a physical-device QR scan remain unverified.

Nindova v0.2.0 — Quiet Depth

Nindova v0.2.0 — Quiet Depth

Nindova v0.2.0 adds a little more thinking without turning bedtime into a competition. The nightly Session stays bounded, local, and genuinely willing to end.

Added

• Gentle and Deeper Masala Mound profiles: two equal, tonight-only 36-Tile choices with real three- and four-layer occlusion.
• One deterministic brass-bloom response for legal pairs, with optional sound and an opacity-only reduced-motion alternative.
• Primary Rest plus optional Rasoi Image Drift, a response-free three-object handoff with no route back into play; closing and deliberately reopening can begin another separately bounded same-night Session.
• Apache License 2.0, redistributed font notices, and a direct non-tracking QR artifact for the canonical /play/ route.

Improved

• Exhaustive board verification covers 382 Gentle and 517 Deeper reachable states, one terminal state each, and zero dead states.
• Tile semantics distinguish covered-from-above, blocked-on-both-sides, free, selected, and settled states.
• Recipe version 4 identifies each board profile. Long-lived Dawn remains version 3; version-4 same-tab records restore a profile and reject stale or unknown-profile active state without creating performance history.
• Public documentation and rendered media now match the layered game and its evidence boundary.

Evidence boundary

• Masala Mound uses recognition, visual search, and spatial planning. It is not proven to improve general memory.
• Nindova is not proven to make people sleepy, improve sleep, or produce a useful dopamine response.
• No weekly rank, score, streak, personal best, history, sleep grade, memory grade, or improvement claim is included. Optional sleep education is paraphrased, attributed, population-level, and outside the nightly Session.
• A pinned GitHub Pages workflow publishes the static build from main. The canonical route, rendered QR destination, service-worker control, same-origin requests, and offline reload are browser-verified; a physical-device QR scan remains unverified.

Nindova v0.1.0 — Rasoi Pairs

Nindova v0.1.0 — Rasoi Pairs

Nindova’s first GitHub release replaces the original experimental arc with Rasoi Pairs, a calm, bounded pair-removal game made from everyday Indian kitchen forms.

Highlights

• 36 tiles, nine motifs, three shallow racks, and one readable free-edge rule.
• Exhaustive verification of all 343 reachable states: one ending, zero dead states.
• No score, count, timer, streak, grade, collection, randomized reward, or promoted replay.
• Native tile buttons, visible names, keyboard operation, 44px targets, 200% zoom, and reduced motion.
• Automatic settle at twelve minutes and a hard fifteen-minute ceiling.
• Same-night deterministic replay and strictly validated same-tab resume without long-lived interaction history.
• First-light Rasoi Dawn with local still and optional silent loop.
• Standalone HTML plus a composed website/docs/offline-PWA bundle.

Downloads

• nindova-v0.1.0.html — self-contained Session; open directly in a current browser.
• nindova-web-v0.1.0.zip — landing page, documentation, /play/ PWA, and standalone file; serve over HTTP for installation/offline behavior.
• SHA256SUMS.txt — SHA-256 digests for both artifacts.

Evidence and limits

Chromium automation verifies the exhaustive engine, full browser arc, responsive layouts, continuous keyboard completion at 200% zoom, strict active-record recovery, reload-during-settlement closure, reduced motion, local migration, Dawn, same-origin PWA/offline behavior, and standalone independence. The real wall-clock production run is recorded separately in docs/testing/m6-release-hardening.md.

This is a behavioral design study, not a sleep tracker or treatment. Rasoi Pairs has not been clinically shown to make people sleepy, produce a specific dopamine response, or improve sleep. Human Punjabi cultural-authenticity review, real-device VoiceOver/TalkBack acceptance, broader installed Safari/Android proof, and public website deployment remain open limitations. The iOS Wall is not included.