3 rounds of SW1 (endgame feature vision) with 5 agents + Qatux. Produced endgame-feature-vision.md (692 lines) covering the full galaxy-to-ground generation pipeline, cultural cascade, replayability architecture, and player experience beats. Workshop was cut short when PO redirected to a 6-phase development cascade (wiki content → economics → planetary maps → player control → world gen → detail coloring). v0.2 target dropped. Heritage roots (D-104/D-105/D-101/D-107) flagged for supersession — real-world cultural corridors replace abstract roots. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
38 KiB
title, author, workshop, sub-workshop, round, created, status, supersedes
| title | author | workshop | sub-workshop | round | created | status | supersedes |
|---|---|---|---|---|---|---|---|
| SW1 Round 1b — Consolidated Notes (Endgame Vision) | qatux | world-generation | 1 | 1b | 2026-03-24 | ready-for-review | sw1-round1-notes.md (for scope framing) |
SW1 Round 1b — Consolidated Notes
Prepared by: Qatux Source files: gestalt-sw1r1b.md, miri-sw1r1b.md, nigel-sw1r1b.md, ozzie-sw1r1b.md, tyre-sw1r1b.md Date: 2026-03-24 Framing: Endgame vision. No v0.2 scope constraints. Station-only concern is resolved by this round — see Pipeline Architecture section.
⚠️ STALENESS ALERT: Heritage Root D-Records vs. Actual World Content
Flagged by team lead during workshop. Requires product owner resolution before cultural generation work proceeds.
What the D-records say
D-104, D-105, D-107, and D-101 (all from the Generator Architecture Workshop, 2026-02-27) build the cultural identity system around 7 abstract heritage roots: Frost, Stone, Tide, Vine, Dust, Iron, Salt. These records define:
- D-104: Heritage Grammar Overlay — "visual grammar of a Frost community versus a Tide community must be apparent to the observant player without a label." Heritage overlay applied at Phase 2 chunk fill.
- D-105: Non-Urban Informal Zone Typology — canonical mapping of informal zone types to heritage roots ("How you hide in a Frost community is architecturally different from how you hide in a Dust community").
- D-107: Trauma Events — trauma amplifies the heritage root baseline (not transforms it).
- D-101: ZonePalette — modifier axis A = heritage root.
What the wiki and world content actually use
docs/design/cultural-migration-pressure.md (authored by Paula, 2026-03-14) — the governing reference for all 300 system wiki pages — defines cultural identity using real-world cultural corridors:
| Corridor | Primary cultural inflection |
|---|---|
| North reach | British Commonwealth diaspora (Australian, Canadian, South African, Indian, Nigerian) |
| South reach | Iberian/Latin American (Portuguese, Spanish, Brazilian) + Southern African (Zulu, Afrikaans) |
| East reach | East Asian and Southeast Asian (Chinese, Japanese, Korean, Vietnamese, Indonesian, Filipino) |
| West reach | Northern/Central European (German, Scandinavian, Dutch, Polish, Czech) + Russian successor states |
The wiki pages for 300+ systems are authored against these corridors. The abstract 7-root system (Frost/Tide/etc) does not appear in cultural-migration-pressure.md or in wiki content.
Why this is a conflict
The generator architecture (D-104, D-105) assumes cultural identity is expressed via 7 abstract roots that are spatially legible at a glance. The world content assumes cultural identity is expressed via real-world heritage corridors. These are different things.
An agent trying to implement "Frost community visual grammar" per D-104 has no mapping to actual world content — no system in systems.db or the wiki has a "Frost" heritage tag. An agent trying to express "Commonwealth heritage architecture" per cultural-migration-pressure.md has no D-record to reference.
D-128 ("culture implicit in starting location — Van Maanen's Star equals Van Maanen's Star culture") leaves this unresolved: it says culture derives from location, but does not specify whether Van Maanen's Star culture is described by an abstract root, a real-world corridor, or both.
Miri's Round 1b Q1 (wave chronology as the root authored input for cultural generation) implicitly assumes the real-world corridor model — her wave chronology maps to settlement patterns of real cultures, not to the 7 abstract roots.
The three possible resolutions
Option A — Abstract roots are a translation layer. The 7 abstract roots are metaphors that map onto real-world cultural patterns. Iron = Commonwealth (institutional, procedural, labor-focused). Tide = Iberian/LatAm (relational, communal gathering). Frost = Northern European/Scandinavian (individual space, cold reserve). The D-records remain valid if the translation table is authored. This is Miri's implicit assumption in Round 1b.
Option B — Abstract roots are superseded. The real-world corridor system replaces the 7 abstract roots entirely. D-104, D-105, and related records need amendment to reference the corridor system. The generator produces Commonwealth-heritage visual grammar, not Frost-heritage grammar.
Option C — Two parallel systems, different levels. Abstract roots describe social pattern (gathering behavior, privacy norms, threshold character — Miri Q5 Layers 1 and 2). Real-world corridors describe material culture (naming conventions, architectural vocabulary, object selection — Miri Q5 Layer 3). The two systems operate at different levels of the cascade and don't conflict.
D-records requiring attention
| Record | Status | Issue |
|---|---|---|
| D-104 (Heritage Grammar Overlay) | Stale — needs amendment or supersession | Built around 7 abstract roots; no mapping to actual world content |
| D-105 (Non-Urban Informal Zone Typology) | Stale — needs amendment | Heritage root correlations reference Frost/Dust/etc; no link to corridor system |
| D-107 (Trauma Events as EraModification) | Cross-reference stale | References D-104 as "heritage baseline" — inherits staleness |
| D-101 (ZonePalette Modifier System) | Potentially stale | Modifier axis A = heritage root — which root framework? |
No work should begin on heritage root authoring (Miri Q1, Round 1 recommendation B) until the product owner resolves which framework is canonical.
How This Round Differs from Round 1
Round 1 was inadvertently scoped to district/room level. Round 1b reframes to the full galaxy-to-room cascade. The pipeline has 8 levels:
1. Galaxy — 300 systems, topology, cultural corridors
2. System profile — wealth, governance, atmosphere, biomes, cultural identity
3. Topography — terrain, terrain effect on settlement placement
4. Arrival/Gateway — inbound experience at system and location
5. Content spidering — how generation unfolds as the player moves
6. Settlement — city/town/outpost placement, district composition
7. District — neighborhood layout, Phase 1 + Phase 2 generation
8. Room — interiors, furniture, cultural grammar at intimate scale
All questions below are organized by the level they primarily address.
Level 1: Galaxy
Questions about what the galaxy itself is, how it's structured, and what it promises the player.
| Agent | Question | Recommendation |
|---|---|---|
| Miri Q1 | Heritage roots — origin worlds with diffusion, or universal distribution with corridor weights? | D+C: Wave chronology (authored) → corridor cultural identity (derived) → heritage probability per system (generated) |
| Nigel Q1 | Does the galaxy topology regenerate between runs? | A: Fixed canonical 300 systems, fixed wormhole topology. Internal state (faction control, economy, NPC triangles) seeded per run. |
| Ozzie Q1 | Is the galaxy map a wonder or a menu? | D: Layered and unlockable — starts as bare geography + economics, faction and mystery layers unlock through play |
| Ozzie Q5 | When does the player feel the scale for the first time? | C+D: Designed "you are here" zoom moment (camera pulls back to show how small you are) + NPC social scale texture (someone mentions a weeks-long journey to a place you treated as nearby) |
Convergence — Fixed galaxy, seeded situation: Nigel and Miri agree: the galaxy as a place should be canonical and learnable. What varies per run is the situation inside that fixed map. Nigel: "The Settled Reach as a place should be canonical and learnable. The replayability comes from the situation inside that fixed map." This framing resolves the galaxy-topology question cleanly.
Convergence — Galaxy map as earned knowledge: Ozzie's unlockable map layers and Nigel's fixed-topology-but-seeded-internal-state are the same design idea at different levels. The galaxy earns its scale through the player's experience, not through declaration. Neither agent advocates for "show everything at once."
Upstream dependency: Miri's recommendation (wave chronology → corridor identity → heritage probability) requires the wave chronology to be authored. This is a worldbuilding document, not a generator parameter — but it must exist before corridor cultural identities can be derived, and corridor identities must exist before per-system heritage weights can be generated. The wave chronology is the root authored input for the entire cultural cascade.
Level 2: System Profile
Questions about what a system is made of and what cascades from its profile downward.
| Agent | Question | Recommendation |
|---|---|---|
| Gestalt Q1 | What are the economic axes of a system profile, and which cascade visibly? | C (full: multiple industries, power players, historical disruptions) as endgame; B (dual axis: aggregate wealth + wealth-holder) as minimum viable |
| Miri Q2 | Does the physical environment modulate how heritage roots express themselves? | C+D: Continuous axis modulation (environment shifts heritage expression along numeric parameters) + tension markers when heritage is in conflict with environment |
The cascade rule (from Gestalt): Every economic axis on the system profile should produce at least one visible attribute at district level, which produces at least one visible attribute at room level. If an axis doesn't cascade visibly, cut it — it's database weight.
Miri's extension: The same cascade principle applies to cultural identity. Heritage root + environment type = cultural register. This produces variety beyond 7 roots: same root, different environment, different expression. A Frost community on an ice world is consonant; a Frost community on a tropical world is in tension — and that tension is spatially legible (walls everywhere, shade corridors, territorial marking).
Tension: No other agent explicitly addresses the environment × heritage modulation. This is Miri's contribution alone. It's not contested but it introduces a continuous parameter system that needs costing at implementation time — see Tyre's Layer 2 assessment for the general approach.
Level 3: Topography
Questions about whether physical terrain shapes the world or merely decorates it.
| Agent | Question | Recommendation |
|---|---|---|
| Gestalt Q2 | Does topography constrain settlement placement or decorate it? | C: Hard constraint — rivers create district boundaries, mountains block corridors, coastlines produce port districts automatically |
| Nigel Q2 | Does regional topography cascade into settlement form? | C as principle, D as long-term target; B (resource-driven placement) as practical starting point |
| Tyre Q2 | How deep does planetary topography go? | A (abstract tags) to start; B (noise-based regional heightmap ~1km resolution) as the enrichment that makes world maps meaningful |
Strong convergence — topography as causal, not decorative: Gestalt and Nigel reach the same conclusion from different angles. Gestalt: "Cosmetic topography produces no decisions." Nigel: "Geographic causality is the replayability multiplier that operates automatically." Both argue that if terrain determines settlement form, a single seed variation at the geographic tier cascades all the way to NPC relationships — without the generator explicitly engineering that variety.
Tension — implementation depth: Tyre's A (abstract tags) is sufficient for district generation at ground level. The 30° camera means the player never sees a heightmap — they see whether they're in a coastal, mountain, or plains environment, which is communicated by SettingType and palette. Gestalt and Nigel's design principle (topography as hard constraint) is universally agreed; Tyre is clarifying that the minimum viable implementation of that principle is tag-based, not heightmap-based. The design intent and the implementation path are compatible.
When does the depth matter? Tyre's assessment: B (noise heightmap) becomes valuable when a world map or minimap shows the region. "There's a mountain range to the west, a river delta to the south." This is a navigation/exploration feature, not a generation prerequisite. Ozzie's Q5 (scale drop moment) connects here — the "you are here" zoom works better with real geographic context visible at the regional level.
Level 4: Arrival and Gateway
Questions about the first impression of every new location — the moment that sets what a place will mean to the player.
| Agent | Question | Recommendation |
|---|---|---|
| Gestalt Q3 | What does the arrival experience need to guarantee? | C: Arrival as information front-load — 3-5 legible signals about location profile before any NPC dialogue |
| Nigel Q3 | Do wormhole gateways create different first impressions per seed? | C for generator (seeded gate positions within zone-type constraints), D as design goal (player chooses which gate, each gate is a distinct social chokepoint) |
| Ozzie Q2 | What does the player see when they arrive? | D: Contextual by travel method. C (orbital view) as default for normal transit — the planet fills half the viewport before you land. |
Convergence — arrival as high-value information moment: Three agents address the arrival experience independently and all conclude it must be meaningful. Gestalt frames it as information front-load. Nigel frames it as replayability surface. Ozzie frames it as the "promise delivery" moment. These are the same argument: arrival is the moment the world earns its scale. Wasted arrivals are wasted opportunities.
Gestalt's "arrival signal slots" concept: Each slot maps to an observable feature (guard density, commercial sprawl into arrival area, maintenance condition, faction signage, crowd composition) that encodes the location's profile. The skilled player reads more from the geometry than the unskilled player — information asymmetry expressed spatially.
Nigel's gate-as-chokepoint concept: Gateways as social chokepoints produce different starting knowledge states. Player A uses the commercial gate, Player B uses the industrial gate — different first NPCs, different first triangles, different starting social information. This is a replayability surface that doesn't require any topology to change.
Ozzie's orbital view as default: The orbital view — planet filling half the viewport, settlements visible before landing — is the designed arrival beat that makes the world feel like it has geography before the player enters it. "Survey, then commit." Connects to Ozzie's Q5: this is also where the scale drop moment happens.
Level 5: Content Spidering
Questions about how generation unfolds as the player moves through the world — the LOD/simulation-tier generation question.
| Agent | Question | Recommendation |
|---|---|---|
| Nigel Q4 | Does the world live before the player arrives? | B (seeded historical state) for initial implementation; C (background simulation with meaningful outcomes) as endgame target; D (full pre-history simulation) as philosophical north star |
| Ozzie Q3 | Does the world feel pre-existing or generated-as-you-walk? | Environmental archaeology first (D) — prop placement tells stories of pre-player history — then NPC knowledge (B), visual sightlines (A), economic activity (C) |
| Tyre Q3 | Offline, on-demand, or progressive generation? | D (progressive): Layer 1 pre-gen at game start, Layer 2 on system visit, Layers 3-7 on district approach. Generation always hidden behind diegetic transitions. |
| Tyre Q5 | What exists at each distance from the player in the galaxy? | systems.db metadata for all 300 → LocationProfiles for 1-3 hops → Phase 1 skeletons for adjacent systems → Phase 2 fill for current district |
Strong convergence — progressive generation, hidden behind transit: Tyre's architecture (progressive, diegetic) and Nigel's design goal (world lives before you arrive) are architecturally aligned. Tyre's model: generation happens during transit loading, never as a visible pause. Nigel's model: when the player arrives, the location has a history. The progressive architecture is the mechanism that makes Nigel's goal possible — location profiles and Phase 1 skeletons are generated during transit, and seeded historical state (Nigel B) can be computed at the same time.
Ozzie's environmental archaeology: The world feels pre-existing not through generation depth but through evidence. The crashed cargo loader with personal effects still inside. The memorial with fresh flowers. The barricaded doorway with scratches on the frame. Ozzie prioritizes this (D) above visual sightlines (A) because it's more visceral: "that's the moment the player stops thinking 'I'm in a generated space' and starts thinking 'someone was here before me.'" This is primarily a Phase 2 fill / object placement feature — prop groups that imply pre-player events.
The "does the world live before you arrive" question (Nigel Q4): Nigel flags this as the most architecturally load-bearing question in the whole workshop. The D-097 simulation tier system already has the skeleton. The question is whether background tier advancement produces meaningfully different states on first player visit. This needs a product commitment before architecture decisions lock.
Level 6: Settlement
Questions about what settlements are, where they go, and how their history drives their physical form.
| Agent | Question | Recommendation |
|---|---|---|
| Gestalt Q4 | Atom of settlement identity — single tag, type+power, or type+power+era? | C (type + power structure + era layering) as endgame minimum; D (full profile) as ideal |
| Miri Q3 | Does long-stable cultural blending produce irreducible hybrid identity? | B (named hybrid archetypes for common blends) + C (simulated coherence timer) + D (trauma as hybrid accelerant, connects to D-107) |
| Miri Q4 | Does heritage root shape where communities form? | C (historical layering — first wave takes preferred terrain, later waves adapt) + B (cultural pressure score as output, feeds NPC behavior) |
| Nigel Q5 | Does each settlement have a seeded developmental biography? | C: 4-6 parameters (founding reason, peak period, current trajectory, heritage at founding vs today, disruption event) drive era distribution, faction strength, NPC mix, triangle topology |
| Tyre Q1 (Layer 2) | How do Location Profiles get populated? | B: Hybrid — generated baseline + hand-authored overrides for ~30-40 narrative-notable systems |
Strong convergence — settlement as legible history: Gestalt Q4, Nigel Q5, and Miri Q3+Q4 are all arguing the same point at different levels of detail. A settlement needs more than a type tag — it needs a historical trajectory that is spatially legible. The player should be able to read a settlement's biography from its physical form.
- Gestalt: "Every era transition in a settlement's history should leave at least one spatially visible scar."
- Nigel: "The player who knows how to read settlement biographies has acquired cultural competence — a durable skill that makes them better at the game without making the game easier."
- Miri: "Architecture is the community's argument with its environment."
Nigel's settlement biography seed resolves Gestalt's era layering question: In Round 1, era distributions were seeded independently per district (Nigel R1 Q5, option B+C). Nigel R1b Q5 supersedes this: era distributions should flow from the settlement biography. A "declining mining hub" settlement produces high Decay + EmergencyExtension blocks in industrial zones, OrganicGrowth in residential (workers who stayed), InstitutionalIncursion in commercial (corporate attempts to buy in). Every block's EraCause is legible to a player who understands the biography — even if the biography is never stated explicitly.
Miri's hybrid identity formation (Q3): This is uncontested — no other agent addresses it. Stable cultural blends (200+ years within ±10% variance) accumulate hybrid markers that belong to neither parent culture. This produces the setting's most interesting named places: communities with irreducible identities that are theirs alone. D-107's trauma event system becomes a cultural forge when connected here.
Tyre's Layer 2 gap (Location Profiles):
The biggest structural gap in the entire pipeline. systems.db has 300 systems with rich attributes; there are no individual planet/station/moon records. Everything from Layer 3 downward depends on this. Tyre recommends hybrid approach: grammar-based generation for the majority, hand-authored profiles for ~30-40 narrative-notable systems (already tagged in systems.db). This is the critical path bottleneck.
Level 7: District
The existing DistrictSkeleton level — Phase 1 and Phase 2. Carried forward from Round 1, now enriched by settlement biography.
Round 1b agents mostly addressed levels above district. The key additions at district level from this round:
- Settlement biography drives district era distribution (Nigel Q5 — see above). Districts no longer need independent era seeding; they inherit from settlement biography.
- Zone cluster adjacency varies per seed (Nigel R1 Q3 — carried forward). Zone clusters are coherent (commercial clusters together, industrial clusters together); which cluster is adjacent to which varies per seed.
- Guarantee system applies to all zone types (Tyre Q4 — see Pipeline Architecture below). Tier 1 guarantees (Social Hub, Informal Zone, Encounter Corridor) apply to ALL inhabited districts regardless of SettingType. Phase 1 is zone-agnostic.
Return visits (Ozzie Q4):
| Option | Description |
|---|---|
| B. NPC drift | Characters have moved, aged, formed new relationships. Social fabric has shifted. |
| D. Consequence echo | The world has reacted to what the player did. People remember. Consequences have propagated. |
Ozzie's recommendation (B+D as minimum, C/economic change if simulation supports it) defines the experiential test for whether the world is a stage set or a living place. The first time a player returns to a location and someone says "I heard about what you did at the refinery" — that's the moment the simulation proves itself.
Level 8: Room
Cultural and economic identity at the intimate scale — the terminus of the cascade.
| Agent | Question | Recommendation |
|---|---|---|
| Gestalt Q5 | How deeply does economic context cascade into room interiors? | D (full cascade) as endgame target — every room is an information puzzle about its inhabitant's economic position, factions, heritage, and operational state |
| Miri Q5 | What design tokens carry cultural identity at room level? | B as baseline (Layers 1+2: spatial grammar tokens + social pattern tokens — zero asset cost), C as enrichment (adds Layer 3: object selection tokens per heritage root) |
Convergence — rooms as information puzzles: Gestalt Q5 and Miri Q5 converge on the same proposition: a room that only tells the player its function (this is an office) gives them nothing. A room that signals economic position, cultural heritage, and operational state gives them a decision. "Every room a player enters should contain at least one observable signal they couldn't have predicted from the exterior."
Miri's token framework — three layers:
| Layer | Tokens | Asset Cost | Cultural Signal |
|---|---|---|---|
| 1. Spatial grammar | privacy_default, sightline_character, gathering_orientation, threshold_character |
Zero — derived from heritage root | Space feels culturally specific before any object is seen |
| 2. Social pattern | stranger_distance, group_size_prior, noise_level_prior, eye_contact_norm |
Zero — drives NPC behavior | Room behaves culturally specific |
| 3. Object selection | Wall objects, surface materials, lighting character — per heritage root authored lists | Asset cost — one list per root | Room reads culturally specific |
Layers 1+2 cost nothing and deliver the essential cultural read. A Frost bar and a Tide bar feel different before a single object is placed. Layer 3 is the enrichment layer — add per heritage root as authoring bandwidth allows.
The cascade test: Gestalt's proposed test: "Can a skilled player stand in a room and read backward from the furniture to the system's power structure?" If yes, the cascade is working. If not, the generator is producing art assets, not simulation primitives.
Pipeline Architecture: How 5 Zone Types Share One Pipeline
This is the answer to the Round 1 flag. Tyre Q4 is the key analysis.
Resolution: Phase 1 (skeleton generation) is zone-agnostic. Phase 2 (chunk fill) diverges by zone type family. Three families, not five independent generators.
| What's shared (all zone types) | Notes |
|---|---|
| Phase 1 skeleton structure (4×4 blocks, zoning, social sites) | Block grid works for Station, Urban, Agricultural, Wilderness. Block meaning changes, structure doesn't. |
| Guarantee audit system (D-097) | Tier 1 guarantees apply to ALL inhabited districts regardless of SettingType |
| ZonePalette + modifier system | Base palette per SettingType, modifiers apply identically |
| ChunkMutations / DamageOverlay | Post-generation modification is setting-agnostic |
| Social site placement mechanism | Site template pool is constrained by SettingType; placement mechanism is shared |
| Chunk streaming | Already setting-agnostic |
| What diverges (per zone type family) | Family members | Cost per family |
|---|---|---|
| Phase 2 fill algorithm | Interior: Station, Orbital, Specialized | 2-4 weeks |
| Phase 2 fill algorithm | Urban: Urban, Industrial, Maritime/port | 2-4 weeks |
| Phase 2 fill algorithm | Natural: Wilderness, Agricultural, Water | 2-4 weeks |
| Tile palette data | All SettingTypes | 1-2 weeks per type (asset-dominated) |
| Height/terrain treatment | Station (flat), Urban (building heights), Natural (heightmap) | 1-3 weeks per family |
Total endgame generation engineering:
- Phase 1 (once): 3-5 weeks
- Phase 2 (3 families): ~9 weeks + per-type customization (~3-4 weeks)
- Location Profile grammar (Layer 2): 3-4 weeks
- Settlement generation (Layer 6): 3-5 weeks
- Content spidering pipeline (Layer 5): 2-3 weeks
- Total: ~20-30 weeks depending on topography depth
Critical path: Layer 2 → Layer 6 → Layer 7. Everything else is either already built (Layer 1), cheap (Layer 3A, Layer 4), or parallelizable (Layer 5, Layer 8). The three critical layers are sequential.
What's Already Built (Tyre's Inventory)
| Layer | Status | Notes |
|---|---|---|
| Layer 1: Galaxy (300 systems) | Built | systems.db (7 tables), 300+ wiki pages, star map algorithm, generate-star-map.py, gate topology data, cultural data |
| Layer 2: Location Profiles | GAP | Biggest structural gap. systems.db has system-level data; no individual planet/station records exist |
| Layer 3: Topography | Not built | Solution space understood. Abstract tags = days. Noise heightmap = 3-5 weeks. |
| Layer 4: Gateways | Partially designed | D-093, D-095 define lore. Horizon station = Station district. Gate-specific templates needed. |
| Layer 5: Content spidering | Architecture designed | Simulation tiers (D-026) built. Streaming built. Generation hooks not built (TODO in code). |
| Layer 6: Settlement | Not built | Missing bridge between topography and districts |
| Layer 7: District → Building | Data model built (579 lines) | Phase 1 + Phase 2 algorithms not implemented |
| Layer 8: 2D → 3D rendering | In progress | Generate 3D-ready data now; 2D renderer filters z > 0 |
The pipeline has bookends: Layer 1 (galaxy dataset) is mature. Layer 7's data model (district skeleton) is mature. The middle — Layers 2, 3, 5, 6, and Phase 1/2 algorithms for Layer 7 — needs to be built.
Convergence Map
| Topic | Agents | Shared Conclusion |
|---|---|---|
| Fixed galaxy, seeded internal state | Nigel Q1, Miri Q1 | Galaxy is canonical and learnable; situation (faction, economy, NPC triangles) varies per run |
| Topography must be causal, not decorative | Gestalt Q2, Nigel Q2 | Terrain must cause settlement form to produce decisions; cosmetic topography is dead design space |
| Arrival = information front-load | Gestalt Q3, Nigel Q3, Ozzie Q2 | Every arrival is a once-per-location signal moment; 3-5 legible reads before any NPC dialogue |
| Settlement = legible history | Gestalt Q4, Nigel Q5, Miri Q3/Q4 | A settlement's physical form must express its biographical trajectory; players read history spatially |
| Progressive generation, hidden behind transit | Tyre Q3, Nigel Q4 | Generation happens during diegetic transitions; player never sees "generating..." after game start |
| Rooms as information puzzles | Gestalt Q5, Miri Q5 | Every room must contain at least one signal the player couldn't have predicted from the exterior |
| Zone families (Phase 2 diverges, Phase 1 shared) | Tyre Q4 | Phase 1 is zone-agnostic; 3 Phase 2 families (Interior/Urban/Natural) cover all 5+ zone types |
Tensions Map
| Tension | Agents | Description |
|---|---|---|
| Topography depth | Gestalt Q2, Nigel Q2 vs Tyre Q2 | Design goal (hard constraint) agreed; Tyre notes tags are sufficient for district generation. Value of heightmaps is the world map experience, not ground-level gameplay. |
| Background sim fidelity | Nigel Q4 (C: meaningful evolution) vs current architecture | D-097 supports background tiers; whether they produce meaningfully different first-visit states requires a product commitment. Most architecturally load-bearing open question. |
| Cultural cascade depth at room | Gestalt Q5 (full cascade, endgame) vs Miri Q5 (Layers 1+2 as baseline) | Compatible in direction (both want full cascade eventually); differ on minimum viable for cultural room identity. Miri's Layers 1+2 answer: zero asset cost, sufficient cultural read. |
Upstream Dependencies
Dependencies that must be resolved before downstream questions can be answered, in order:
-
Wave chronology (Miri Q1) — authored worldbuilding document. Root input for the entire cultural cascade. Without it, corridor identities cannot be derived, heritage probability tables cannot be authored, and Miri's heritage root authoring scope cannot be scoped. Blocks: everything in cultural generation.
-
Heritage root blend of starting district (Miri upstream dep, carried from R1) — Sova Transit / Van Maanen's Star. Still requires product owner confirmation. Miri's working hypothesis: Iron/Stone blend with Tide undertow. Blocks: heritage root authoring, naming system authoring.
-
Location Profile grammar commitment (Tyre Q1/Layer 2) — the biggest structural gap. The grammar that maps system attributes to planet/station profiles is the critical design work. Blocks: Layers 3-7 producing meaningful output.
-
Background simulation fidelity commitment (Nigel Q4) — does the world live before the player arrives in a meaningful way? Answer shapes whether the content spidering pipeline needs to simulate historical evolution or just serve seeded static state. Blocks: Layer 5 architecture decisions.
-
Zone type family confirmation (Tyre Q4) — are the three families (Interior, Urban, Natural + Transitional as hybrid) the right grouping? Blocks: Phase 2 implementation strategy and asset pipeline.
Proposed Interview Clusters for Product Owner
Six clusters, organized by pipeline level. Cluster 1 is prerequisite for all others.
Cluster 1: Galaxy Identity and Cultural Foundation
Prerequisite for authoring work. Must be answered first.
-
Is the galaxy map canonical and fixed, with internal state seeded per run? (The 300 systems, their names, and wormhole topology are learnable across playthroughs; faction control, economy, and NPC triangles vary.) Nigel and Miri both assume yes — confirm.
-
How does the cultural geography of the galaxy derive from settlement history? Miri proposes a wave chronology: the galaxy was settled in waves, each carrying a dominant cultural pattern. Corridor cultural identity is derived from wave history, not hand-authored per system. Confirm the wave model as the setting framework, or specify the alternative.
-
What does the galaxy map reveal at start vs unlock through play? Ozzie recommends: economic signal and bare geography at start; faction presence, reputation, and mystery layers unlock through play. The map rewards engagement with the world.
Cluster 2: System Profile and Economic Cascade
-
What are the economic axes of a system profile? Gestalt proposes minimum viable = dual axis (aggregate wealth + wealth holder: who is rich — individuals, corporations, government, or nobody). Full endgame = multiple industries, multiple power players, historical disruptions. Which is the design commitment?
-
Does physical environment modulate cultural expression? A Frost community on an ice world vs a Frost community on a tropical world — same heritage root, different expression because the environment is in tension with the culture. Miri proposes continuous axis modulation. Confirm whether this interaction is in scope and at what fidelity.
-
How do Location Profiles get populated? Tyre identifies this as the biggest structural gap. Hybrid approach: grammar-based generation for the majority of the 300 systems; hand-authored profiles for ~30-40 narrative-notable systems already tagged in
systems.db. Confirm approach and identify who authors the grammar.
Cluster 3: Geography and Where Settlements Form
-
Does topography constrain settlement placement or decorate it? Gestalt argues for hard constraint — rivers create natural district boundaries, mountains block corridors, coastlines produce port districts. Nigel agrees: geographic causality is the replayability multiplier that operates automatically. The implementation starts with abstract tags (cheap, Tyre says sufficient for district generation) and enriches to a noise heightmap when the world map becomes a player-facing feature.
-
Does heritage root influence where communities form? Miri proposes: cultural preferences are probability weights, not hard constraints. Off-preference settlement generates a
cultural_pressure_scorethat feeds into NPC behavior. Historical layering: first wave takes preferred terrain, subsequent waves adapt. This produces stratified communities with built-in cultural tension.
Cluster 4: Arrival and First Impressions
-
What does the arrival experience guarantee? Gestalt proposes "arrival signal slots" — specific observable features (security density, commercial sprawl, faction signage, maintenance condition, crowd composition) that are set by the location profile and legible to the observant player. How many signals, and which are mandatory?
-
Is the orbital view the default arrival experience for planetary locations? Ozzie: the planet fills half the viewport before the player commits to landing. They see the geography, can identify settlements, feel the scale. "Survey, then commit." How the player traveled determines what they see — but the orbital view should be the designed default.
-
Do wormhole gateways produce different first impressions per seed? Nigel proposes: gate placement seeded within zone-type constraints (gate must open onto a transit or commercial zone, but which zone per seed). Two players through the same gate topology have different arrival contexts — different first NPCs, different first social information.
Cluster 5: Settlement Biography and World History
-
What is the atom of settlement identity? Gestalt proposes: type + power structure + era layering as the endgame minimum. A corporation-controlled port and a guild-controlled port are genuinely different games; add era history and every era transition leaves a spatially visible scar. Confirm.
-
Does each settlement have a seeded developmental biography? Nigel proposes: 4-6 parameters (founding reason, peak period, current trajectory, dominant heritage at founding vs today, disruption event) that drive era distributions, faction strength, NPC mix, and triangle topology. The biography makes settlement variety coherent rather than arbitrary. Confirm whether biography is the generation mechanism or whether era parameters remain independently seeded.
-
Does the world live before the player arrives? The most architecturally load-bearing question (Nigel Q4). Options: (B) seeded historical state — each location generates with a developmental biography, doesn't evolve until player visits; (C) background simulation — unvisited locations advance in fidelity-reduced form over in-game time, producing different states depending on when the player arrives. The D-097 simulation tier system supports both; only B is achievable in near term. Confirm which is the design commitment.
Cluster 6: Room Identity and the Cascade Terminus
-
How deeply does economic context cascade into room interiors? Gestalt proposes D (full cascade) as the endgame target: every room is an information puzzle about its inhabitant's economic position, faction affiliations, heritage, and operational state. The test: "Can the skilled player read backward from the furniture to the system's power structure?" Confirm this as the design principle.
-
What is the minimum viable cultural room identity? Miri proposes Layers 1+2 (spatial grammar + NPC behavior tokens) as the baseline — zero asset cost, delivers the essential cultural read. A Frost bar feels different before any objects are placed. Layer 3 (object lists per heritage root) adds enrichment per root as authoring bandwidth allows. Confirm the baseline.
Notes produced by Qatux from 5 agent Round 1b outputs. These 16 product owner questions represent the full endgame vision. Prioritization and v0.2 scoping against this vision is the next workshop agenda.