Files
settled-reach/docs/workshops/world-generation/gestalt-sw1r1b.md
T
jpmschweitzerandClaude Opus 4.6 252e3d380a docs(workshops): complete world generation architecture workshop
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>
2026-03-24 17:12:09 +01:00

16 KiB
Raw Blame History

title, agent, workshop, round, type, created, supersedes
title agent workshop round type created supersedes
SW1-R1b: Gestalt — Endgame Systems Design, Galaxy-to-Ground gestalt world-generation 1 product-questions 2026-03-24 gestalt-sw1r1.md

SW1-R1b: Gestalt — Endgame Systems Design, Galaxy-to-Ground

Framing: No v0.2 scoping. Design the full pipeline as if we had infinite time. These questions target the endgame vision — what does the fully realised generation pipeline look like? Scoping workshops come later.

Angle: Systems design. At each level of the galaxy-to-room cascade, what creates interesting player decisions? How does economic context flow from system profile all the way down to the room the player is standing in?


The cascade I'm interrogating

1. Galaxy         — 300 systems, what exists in each
2. System profile — wealth distribution, governance, atmosphere, biomes
3. Topography     — continental bones, station ring/module structure
4. Arrival        — inbound experience at the gateway/portal
5. Content spider — how detail expands as the player moves
6. Settlement     — cities, towns, outposts vs topographic context
7. District       — neighborhoods shaped by settlement identity
8. Room           — shop fronts, interiors, furniture

The big systems question across all levels: does economic context cascade visibly? Can a skilled player stand in a room and read backward from the furniture to the system's power structure? That's the asymmetric information game. The generator is the thing that produces information to be asymmetric about.


Q1: What are the economic axes of a system profile, and which ones cascade to the street?

The question: When the generator builds Van Maanen's Star (or any system), it establishes an economic profile. How many axes does that profile carry? And critically: which ones produce visible differences at the district and room level?

Why it matters: The tycoon's core decisions are economic. The player needs to be able to read the economic landscape spatially — not just from their insert device. A corporate-dominated system should look different from a worker-cooperative system even if they're equally "wealthy." If the economic profile doesn't cascade to visible space, the world is a backdrop and the economy is pure UI.

This is the information asymmetry spine of the whole pipeline. System profile → district character → interior condition → the player can observe and reason about power from what they see.

Options:

Option Economic axes What cascades visibly
A. Single wealth axis Rich / Middle / Poor Condition modifier only (pristine vs worn)
B. Dual axis: aggregate wealth + wealth holder Who is rich: individuals, corporations, government, or nobody Zoning character, building ownership types, faction presence, NPC wealth distribution
C. Full economic profile Multiple industries, multiple power players, historical disruptions Full district personality matrix — can produce contested zones, parallel economies, boom/bust spatial signatures

Recommendation: C (full profile) is the endgame target, B is the minimum viable.

The dual axis (B) already creates rich differentiation. A corporate-rich system with poor individuals produces fortress-architecture commercial districts, visible security presence, and cramped residential zones. A people-rich system with weak corporate presence produces market-overflow commerce, informal extensions, and self-built residential texture. These are completely different games to play in, even if the aggregate GDP is identical. C adds historical disruption (a formerly corporate zone now in decay, a recently gentrified district still showing its seams) — that's what makes the world feel like it has a past.

The cascade rule: Each economic axis on the system profile should produce at least one visible attribute at district level, which in turn produces at least one visible attribute at room level. If an axis doesn't cascade visibly, cut it — it's just database weight.


Q2: Does topography constrain settlement placement or decorate it?

The question: When the generator places continental bones (mountains, rivers, coastlines, craters, fault lines) or station ring/module topology, do those features constrain where settlements go and what they look like? Or do they provide aesthetic backdrop that generation ignores for placement logic?

Why it matters for player decisions: Topography is the first source of emergent spatial reasoning. If rivers create natural chokepoints and settlements cluster at crossings, the player can predict where commerce happens before they've walked there. If mountains block movement, settlements at passes hold different economic power than settlements in valleys. Spatial logic the player can reason about is information — and information is the master mechanic. Cosmetic topography produces no decisions.

Options:

Option What topography does
A. Cosmetic Mountains look right but don't block roads. Rivers exist but settlements ignore them. Pure visual backdrop.
B. Soft constraint Topography influences settlement placement probability (mountains reduce likelihood, rivers increase it at crossings) but can be overridden by other factors (corporate mandate builds where it wants).
C. Hard constraint + emergent structure Mountains block ground movement corridors. Rivers create natural district boundaries. Coastlines produce port districts automatically. Station ring topology determines module function. Settlements cannot ignore terrain — they respond to it.

Recommendation: C (hard constraint).

Option A produces beautiful screenshots and dead spatial reasoning. Option B is a soft approximation that makes things feel slightly more realistic without actually committing. Option C is the only option that produces emergent geography as a player tool. If the generator places a river that bisects a commercial district, bridge chokepoints become real — surveillance, ambush, customs control, economic choking all become spatially grounded. The player can use the map rather than just inhabiting it.

For station topology, C means ring sections have structural logic: docking rings are near the hull, residential is in the middle, administrative is at the axis. The station's form determines its social geography. A player who understands ring structure can navigate a station they've never visited before — spatial literacy is a skill.


Q3: What does the arrival experience need to guarantee?

The question: When the player comes through a gate/portal into a new system or location for the first time, what does the generator guarantee they see? The arrival moment is the game's "establishing shot" — it determines whether the world feels real and legible, or random and disorienting.

Why it matters: First impressions set the player's mental model of a location. Arriving at a corporate logistics hub should give immediate environmental reads: security checkpoints, customs infrastructure, corporate insignia, the density of freight movement. Arriving at a frontier mining outpost should give different reads: makeshift structures, informal docking, no customs to speak of. If the generator can't guarantee arrival-point differentiation, the world feels like random tile soup from the first moment.

This is also a player agency question: the arrival experience is a moment of pure observation before commitment. It's the scouting window.

Options:

Option Arrival guarantee
A. Type-consistent arrival zones Generator ensures arrivals into a LogisticsHub district look like logistics infrastructure. Arrivals into Residential districts look like residential zones. Type drives look, not economic profile.
B. Profile-encoded arrival zones Arrival zones encode the location's economic/governance profile visually. Corporate-controlled logistics hubs have different arrival aesthetics than cooperative-controlled ones even within the same district type.
C. Arrival as information front-load Arrival zones are specifically designed to give the observant player 3-5 readable signals about the location before they've spoken to anyone. Security level, wealth signal, faction presence, atmosphere condition, NPC density — all legible from the arrival point geometry and object placement.

Recommendation: C (arrival as information front-load).

Option A is a floor — necessary but not interesting. Option B adds economic texture to the same type-consistent shell. Option C is the game's information asymmetry proposition expressed spatially: the skilled player reads more from the arrival environment than the unskilled player because they know what to look for. The tycoon who can assess a market opportunity from the arrival zone geometry has a real advantage. The narrative player who notices the faction graffiti at the arrival point has a lead on what's happening.

The generator needs a set of "arrival signal slots" — each slot maps to an observable feature (guard density, commercial sprawl into arrival area, maintenance condition, faction signage, crowd composition) that is set by the location's profile and legible to the observant player.


Q4: What is the atom of settlement identity — single tag or composition?

The question: When the generator builds a settlement (city, town, outpost), it needs a core identity that drives district subdivision. What is that identity made of? A single type tag? A composition of type + power structure? A full profile with history?

Why it matters for district variety: Settlement identity directly determines how many distinct district personalities a settlement produces and whether those personalities feel in tension with each other. A settlement described only by type (LogisticsHub) produces one personality everywhere. A settlement described by type + power structure (LogisticsHub controlled by a single corporation vs. LogisticsHub controlled by competing guilds) produces districts that reflect the tension between controlling interests. Districts in tension create player decisions — which side do I work with?

Options:

Option Identity atom District variety
A. Single type tag LogisticsHub, Residential, Mixed... 1 personality per settlement, coherent but monotonous
B. Type + power structure LogisticsHub × (Corporation / Guild / Government / Chaotic) 2-4 personalities — dominant character + contested edges
C. Type + power structure + era layering B + history of control changes (formerly corp, now guild-contested) 4-8 personalities — temporal strata visible in district character, spatial archaeology possible
D. Full profile Multiple industries, multiple factions, historical disruptions, ongoing tensions Unlimited personality range, full emergence

Recommendation: C (type + power structure + era layering) as the endgame minimum.

Option B is already a major step up from A — a corporation-controlled port and a guild-controlled port are genuinely different games. Option C adds what makes the world feel lived-in: history is spatially legible. The district that was corporate HQ ten years ago and is now guild-contested has scars — old corporate infrastructure repurposed with guild graffiti, a fancy corporate lobby now serving as a union hall, access control systems the guilds haven't bothered to update because the corporate branding makes them feel legitimate. These are the details that make replayability — a different system's settlement history produces different spatial archaeology.

Option D (full profile) is the endgame ideal — let the full economic profile cascade into settlement identity completely. But C is the minimum for the world to feel like it has a past.

Key design rule: Every era transition in a settlement's history should leave at least one spatially visible scar that a skilled player can identify as evidence of the transition.


Q5: How deeply does economic context cascade into room interiors?

The question: At the room level — shop fronts, offices, storage rooms, living quarters — how much does the room's interior reflect the economic context it sits in? Does the generator produce rooms that are typed by function (office, shop, residential) but economically uniform? Or does each room's interior encode the economic position of the person or organisation that inhabits it?

Why it matters for the tycoon and for information asymmetry: The tycoon is reading the market. They should be able to walk into a competitor's shop and assess its health from interior signals — is the stock well-maintained? Is the counter arrangement that of a thriving operation or a survival mode one? Are the fixtures corporate-grade or secondhand? This is spatial intelligence as gameplay. The generator needs to produce this kind of signal density at room level, not just at district level.

The data model already supports this: ZonePalette has PaletteModifier with EconomicModifier, ConditionModifier, FactionModifier, and Heritage sub-types. The architecture can do this — the product question is whether we commit to using it at room interior level.

Options:

Option Room interior differentiation
A. Function-typed only Office/shop/residential. All offices look like offices. No economic signal.
B. Function + condition tier Office × (Pristine / Functional / Worn / Derelict). Economic health readable from maintenance state.
C. Function + condition + ownership type Office × condition × (corporate / individual / cooperative / abandoned). Ownership type shapes layout logic, not just surface condition. A corporate office and a sole-trader office in identical condition look structurally different.
D. Full cascade Every room interior is a read on its inhabitant's economic position, faction affiliations, heritage, and operational state. Room interiors are small information puzzles that reward observation.

Recommendation: D (full cascade) as the endgame target.

Option D is the game's core proposition expressed at the smallest scale. If information asymmetry is the master mechanic, every space should reward the observant player with reads that the inattentive player misses. A room that only signals function (this is an office) gives the player nothing — they already knew it was an office. A room that signals "this is a sole-trader office, the business is struggling, the operator has Krenn heritage, and they've recently been in contact with a faction that usually operates at the docks" gives the tycoon or investigator a decision: do I approach this person? Are they a useful contact or a liability?

The test: Every room a player enters should contain at least one observable signal that they couldn't have predicted from the exterior. If the room tells them nothing new, the generator isn't doing its job.


Systems summary: the cascade chain

The five questions above, when answered, define the information spine of the whole pipeline:

System profile (Q1: economic axes)
    ↓ cascades into
Settlement identity (Q4: type + power + era)
    ↓ which requires
Topography (Q2: hard constraint — geography creates settlement character)
    ↓ first player contact via
Arrival experience (Q3: information front-load — 3-5 legible signals)
    ↓ which resolves into
Room interiors (Q5: full cascade — every space rewards observation)

The through-line is: the generator produces information; the player reads it; the asymmetry between what skilled and unskilled players can read IS the gameplay.

Every design decision at every level of the cascade should be evaluated against this test: does this produce information the skilled player can exploit? If the answer is no, the feature is set dressing, not a system.

The single highest-value decision: Q1 (economic axes). Everything else in the cascade is downstream of what economic data the system profile carries. Get the economic axes right first, then design what each axis produces at each cascade level.