settled-reach/docs/workshops/content-gap-analysis_v0_1/round1-araminta.md

Round 1 Workshop Response: Araminta (Visual Design)

Agent: Araminta (Visual Designer) Date: 2026-02-11 Scope: All 9 layers from visual design perspective

---

Layer 1: Art Direction & Visual Identity

Q1: Can I define a v0.1 visual identity without resolving Q-003 fully?

Yes. And we should do exactly that.

Q-003 asks about long-term art direction. That's a question about final fidelity, sprite style, animation approach, lighting systems. We don't need to answer any of that to build v0.1. What we need is a v0.1 visual grammar - a set of rules for how information is communicated visually, regardless of art fidelity.

Proposal: Resolve Q-003 partially with a scoped decision.

D-XXX: v0.1 visual approach - "readable information design, atmosphere through text and spatial composition."

• Entities are colored rectangles with text labels. No sprites in v0.1.
• Color communicates RELATIONSHIP TO PLAYER, not object type.
• Spatial layout communicates function - the bar FEELS different from the logistics hub through room shape, furniture density, corridor width.
• Atmosphere is carried by monologue text, dialogue content, and sound indicators - NOT by art.
• All visual rules must scale: a color that means "danger" in boxes-with-labels must still mean "danger" when we add sprites.

This is not "we'll figure out art later." This is "v0.1's art IS information design." The boxes aren't placeholder art waiting for real art. The boxes ARE the visual language at this fidelity level.

Long-term Q-003 (sprite style, animation, lighting) stays open. It doesn't block anything.

Q2: What's the minimum art direction document needed for consistency?

A one-page visual grammar covering:

1. Color palette (see below)
2. Entity sizing rules (NPCs, furniture, walls, doors - relative proportions)
3. Label conventions (what text shows on entities, when, at what zoom)
4. UI layer separation (game world vs. insert overlay vs. monologue text vs. dialogue)
5. Fog/visibility rendering rules (visible, fog-edge, hidden - how they look)

That's it. No style guide for art we haven't made. No font selection for UI that doesn't exist yet. Just the rules for communicating game state visually.

Proposed v0.1 Color Palette

The Commonwealth is sleek, advanced, subtle. Even at boxes-with-labels fidelity, we can communicate mood.

Background/Environment:

• #1a1a2e - Deep navy. Base floor/ground. The station is dimly lit by default.
• #2d2d44 - Slightly lighter. Walls, barriers, solid obstacles.
• #0f0f1a - Near-black. Fog/hidden areas. Darkness should feel HEAVY.

Entity Colors (relationship-to-player, NOT object type):

• #4a9ebb - Cool teal. Neutral/unknown NPCs. Default state.
• #6bc9a6 - Soft green. Friendly/trusted NPCs. People your character knows and likes.
• #e8c547 - Warm amber. NPCs of interest. Your monologue has flagged something about them.
• #d45d5d - Muted red. Hostile/dangerous. ONLY when the character KNOWS there's danger. Not omniscient threat display.
• #8b8ba0 - Grey. Furniture, static objects, environmental elements.

UI Overlay Colors:

• #3d7aaf with 40% opacity - Insert/minimap overlay tint. The neural implant's "screen."
• #ffffff at 85% opacity - Primary UI text (labels, insert readouts).
• #a0a0b8 at 70% opacity - Secondary UI text (non-critical info).

Monologue Text:

• #c8d0e0 - Internal monologue. Slightly cool, slightly ethereal. This is thought, not speech.
• Italicized. Positioned bottom-center or left panel. Fades in/out, doesn't snap.

Dialogue Text:

• #ffffff - NPC speech. Clean white. Grounded, real.
• NPC name in their relationship color above the dialogue line.

Key principle: Red means danger TO YOUR CHARACTER, not danger in the abstract. If the smuggler's friend is secretly an informant, that NPC is GREEN until the smuggler discovers the betrayal. The detective might see that same NPC as AMBER (person of interest) from the start. Same NPC, different color, because asymmetric information drives the palette.

This is D-011 (fog of perception) applied to the color system. The player sees what the character knows.

Fog Rendering Specification

Three visual states for tiles:

State	Appearance	When
Visible	Full color, full detail, entities shown	Currently in LOS
Fog-edge	Desaturated 50%, entities hidden, sound indicators shown	Adjacent to LOS, within medium sound range
Hidden	Near-black (#0f0f1a), no entities, no indicators	Outside perception range
Remembered	Desaturated 70%, static snapshot of last-seen state, no entities	Previously visited, now outside LOS

The "remembered" state is important. D-011 says fog returns on departure, but the player should see a ghost of what they last saw - furniture stays, NPCs vanish. This communicates "I was here, but things may have changed." It costs almost nothing to implement and sells the information-decay concept visually.

---

Layer 2: Narrative Framework & Character Design

Visual implications of character divergence

The two characters need distinct visual treatments even at boxes-with-labels fidelity:

• Smuggler's player entity: Slightly different shade or shape indicator. Doesn't need to be dramatic - a subtle color difference in the player marker is enough.
• Detective's player entity: Same principle, different accent.

More importantly: the insert overlay should feel different per character. The smuggler's insert is consumer-grade, maybe slightly less polished. The detective's is institutional - cleaner layout, more structured information display. This is a v0.2+ concern but worth noting now so the insert UI is designed with parameterization in mind.

For v0.1: a small character icon or color accent on the HUD that says "you are the smuggler" vs "you are the detective." Minimal but present.

---

Layer 3: Station District Setting

Visual layout principles for the three social sites

Even with boxes, spatial design communicates EVERYTHING about a location. Here's what each site should feel like through layout alone:

Workplace (Logistics Hub):

• Regular grid layout. Rows of workstations. Wide corridors for cargo movement.
• A few enclosed offices (walls block LOS - management has privacy).
• One loading bay area (open, high visibility, the "public" face).
• One back area (restricted, fewer sightlines - where secrets live).
• Visual feel: Ordered, institutional, functional. The regularity IS the aesthetic.

Social Venue (Bar):

• Irregular, organic layout. Booths create pockets of privacy.
• Central bar area with high visibility (everyone can see who's ordering).
• Corner tables with partial LOS blocks (intimate conversations).
• A back room or alley exit (the sketchy escape route).
• Visual feel: Warm-toned compared to the hub. Furniture density higher. More "walls" from booths and partitions creating complex sightlines.

Smuggling Ring (Hidden within/adjacent to other spaces):

• NOT a labeled "smuggling HQ." It's a maintenance corridor, a storage room, a section of the cargo bay that's officially off-limits.
• Tight spaces, poor sightlines, few entrances/exits.
• Visually indistinguishable from other utility spaces unless you know what to look for.
• Visual feel: Cramped. The contrast with the open logistics hub IS the atmosphere.

Cross-site visual language:

• Public spaces: wider corridors, more open tiles, lighter floor colors.
• Private spaces: narrower, enclosed, slightly darker floor tones.
• Transition zones: doorways, corridors between sites. These are where "wrong place, wrong time" moments happen. They need enough space for the player to see someone passing through.

No art needed. Room shapes and tile colors do the work. A bar booth is a 3x2 rectangle with half-height wall blocks. An office is a 4x4 enclosed room with a door entity. The spatial relationships create the gameplay affordances.

---

Layer 4: Content Pipeline & Authoring Format

Visual implications of the authoring format

Not my primary domain, but one critical visual note:

Every dialogue line needs a display context tag. The renderer needs to know:

• Is this spoken dialogue? (Display in dialogue UI panel, NPC name header)
• Is this overheard conversation? (Display as fog-edge text snippet, lower opacity, partial)
• Is this a monologue reaction? (Display in monologue area, italic, character voice color)
• Is this environmental text? (Display as world-space label on the object/sign/terminal)

These are four different rendering pipelines. The authoring format must tag which pipeline each line targets. Mellanie and Paula need this in the format spec so they don't write a line that falls between categories.

---

Layer 5: Template Content (3 Social Sites)

Visual deliverables per template

Each social site template needs a spatial layout document alongside the content pack. This is my deliverable. For each template:

1. Floor plan - tile-level layout showing walls, doors, furniture, sightlines
2. Sightline diagram - from key observation positions, what can you see? Where are the blind spots?
3. Sound propagation notes - which walls are thin (sound bleeds through), which are thick
4. NPC position markers - where do NPCs stand/sit during their routine phases?
5. Player approach vectors - how does the player enter? What do they see first?

This is the spatial design that makes observation gameplay work. A bar where you can see everyone from the entrance is a different game than a bar where you have to walk past three booths before you see the back table.

Proposed ticket: spatial layout design per social site template (3 layouts, high priority). This should be parallel with the content packs, not after them. Layout affects what content makes sense.

---

Layer 6: NPC Profiles & Dialogue

Visual expression of NPCs at boxes-with-labels fidelity

At v0.1 fidelity, NPC differentiation is:

1. Color (relationship state, as defined in the palette above)
2. Label text (name when known, role when partially known, "?" when unknown)
3. Position and movement (are they at their expected location? Moving with purpose or loitering?)

That's it. No character portraits. No sprite variation. No clothing differences.

But this is enough. The detective walks into the logistics hub. They see: a cluster of teal rectangles labeled "Worker," "Worker," "Worker," one amber rectangle labeled "Jax (shift supervisor)." Amber because the detective's case file flagged Jax. The smuggler walks in and sees: green rectangles labeled "Tomas," "Kira," "Jax" (all known, all friendly). Same room, different visual state, different information.

The color system IS the visual expression of asymmetric information. It doesn't need sprites to work.

Tell system visual language (#251)

D-024 mentions a tell system - observable behaviors mapping to internal states. At boxes-with-labels fidelity, tells are NOT visual. They're:

• Monologue: "Jax is tapping his foot. He does that when he's nervous."
• Dialogue: NPC speech patterns change based on mood (Mellanie's domain)
• Behavioral: NPC is in an unexpected location or moving unusually

The monologue is the tell renderer for v0.1. Visual tells (sweating, fidgeting, avoiding eye contact) are sprite-fidelity features. Text carries the load for now.

When we get to sprites: each NPC should have 2-3 behavioral animation states (idle, nervous, angry) that map to mood axes. But that's later.

---

Layer 7: Monologue & Observation Text

Monologue presentation design

This is one of my key deliverables. The monologue is the player's primary information channel in v0.1.

Presentation rules:

Property	Specification	Rationale
Position	Bottom of screen, left-aligned, 60% width max	Doesn't obscure the game world. Left-align because we read left-to-right; it's the natural "start reading" position.
Font	Monospace or clean sans-serif. Slightly smaller than dialogue.	Monologue is internal thought - quieter than spoken words. Monospace gives a "data overlay" feel consistent with the insert aesthetic.
Color	#c8d0e0 (cool off-white)	Distinct from dialogue (#ffffff) and UI text. Slightly ethereal.
Style	Italic	Universal convention for internal thought. Instantly readable as "this isn't spoken."
Animation	Fade in over 0.3s, hold for read-time + 1s, fade out over 0.5s	Not jarring. Doesn't demand attention like a popup. Respects the player's pace.
Stacking	Max 3 lines visible. New lines push old lines up and fade them.	Prevents monologue flooding during high-event moments.
Priority	Observation > atmosphere > tutorial hint	If multiple monologue triggers fire simultaneously, observation wins. The player needs to know "that NPC is acting weird" more than "nice evening."

Character voice differentiation (visual):

The smuggler and detective should have subtly different monologue presentation:

• Smuggler: Slightly warmer text color (#d0d0c0), slightly more casual phrasing (Mellanie's domain, but the visual warmth supports it).
• Detective: Slightly cooler text color (#c0c8e0), more structured phrasing.

This is a minor touch that compounds over 30 minutes. The player's subconscious registers "this character thinks differently" through the text color before they consciously parse the words.

Observation text vs. monologue vs. dialogue - visual hierarchy

The player needs to instantly distinguish these three information channels:

Channel	Position	Style	Color	Example
Dialogue	Center-bottom panel, NPC name header	Regular weight	#ffffff on dark panel	Kira: "Quiet night at the hub."
Monologue	Bottom-left, no panel/frame	Italic	#c8d0e0	She says that, but she won't look at me.
Observation	Same as monologue but with icon marker	Italic + small eye icon	#c8d0e0	[eye] Jax just walked past the restricted door.
Environmental	World-space, attached to object	Small caps, subtle	#8b8ba0	CARGO BAY 7 - AUTHORIZED PERSONNEL ONLY

The observation marker (small eye icon) is important. It tells the player "your character noticed something." It's the visual cue for the detective mechanic. Over time, players learn: eye icon = pay attention.

---

Layer 8: UI Text & Environmental Content

Q1: What text appears on screen that isn't dialogue or monologue?

Complete inventory:

Insert/HUD Layer (diegetic - it's the neural implant):

• Time display (D-031: game clock, day phase indicator)
• Minimap with POI dots and directional arrows (D-013)
• POI labels when close enough
• Alert notifications ("Message from [contact name]")
• Insert-mode indicators if perception modes are active (v0.1: just natural vision, but the slot should exist)

World-Space Labels:

• Location names (room/area labels: "Logistics Hub - Bay 7")
• Door labels ("Office," "Storage," "Restricted")
• Sign text (business names, warnings, directions)
• Terminal screens (if examine mechanic exists: cargo manifests, schedules, case files)
• NPC name/role labels (floating above entity rectangle)

System UI (non-diegetic, minimal):

• Pause indicator
• Character name/role reminder (top corner, very subtle)
• Save indicator (if save/load exists)
• Interaction prompt (when near an interactable: "E: Examine" or similar)

NOT on screen (deliberately):

• Health bars, stamina bars, XP indicators - none of these exist in v0.1 and may never exist in traditional form
• Quest log/objective marker - the monologue IS the quest hint system
• Mini-tutorial popups - monologue handles this diegetically (D-016)

Q2: How does the neural implant UI look?

It's a literal implant overlay, not an abstracted game UI. This is non-negotiable given D-013.

The insert is what the character sees when they activate their neural implant's navigation overlay. In v0.1, this means:

Visual treatment:

• Semi-transparent overlay with a subtle blue-tinted border or vignette (#3d7aaf at 15-20% opacity)
• The minimap is NOT a separate UI window. It's an overlay that tints the top-right (or bottom-right) corner of the game view.
• POI markers are simple geometric shapes: circles for locations, diamonds for contacts, squares for objectives.
• Text on the insert uses a slightly different font weight or a monospace font - it's DIGITAL text rendered by the implant, distinct from the player's natural perception.

Behavioral rules:

• The insert is always "on" in v0.1 (the character always has their implant active). Toggle is a v0.2+ feature.
• Insert information updates based on character knowledge (D-013: POIs only appear when learned).
• The overlay should feel like looking at a heads-up display, not a game map. Subtle scan lines or a very light grid pattern would reinforce this. Even at v0.1, a 1-pixel grid at 5% opacity costs nothing and communicates "technology."

What the insert does NOT do:

• Show NPC positions beyond what the character can sense (no omniscient radar)
• Show information the character hasn't learned
• Provide objective markers for things the character doesn't know about

The insert is BOUNDED by the same information rules as everything else (D-010 principle 2). It's a tool the character uses, not a game abstraction.

Q3 (Gestalt asked this, but visual input): Onboarding through monologue

The monologue IS the tutorial. But the insert could display a brief "new overlay detected" or "insert calibrating" sequence on first load to teach the player that the blue-tinted area is their implant. One line of insert text: "Neural insert active. Navigation overlay online." Then the monologue takes over: "Let's see what's on the schedule today." Diegetic from the first second.

---

Layer 9: Audio Direction

Visual design implications of sound

Since v0.1 may have limited or no audio, sound information needs VISUAL representation:

Sound indicators at fog edge (D-018 medium range):

• Small animated ripple or pulse icon at the fog boundary, in the direction of the sound.
• Color-coded by sound type:
  • #c8d0e0 (neutral) - footsteps, movement, ambient activity
  • #e8c547 (amber) - conversation, voices (something to investigate)
  • #d45d5d (red) - aggressive sounds, alarm, conflict
• Size/intensity pulses with sound volume.
• The monologue fires simultaneously to interpret: "Voices. Coming from the back office."

Close-range sound (within LOS):

• No special visual indicator needed. You can SEE the source.
• Dialogue or conversation text renders normally.

Long-range insert alerts:

• Display on the insert overlay as a notification blip.
• Text: "[Contact name] sent a location pin" or "Disturbance reported in [area]."
• These are spoofable (D-018) - the visual design should NOT distinguish real from fake alerts. The player has to figure that out.

Even if we have zero audio files in v0.1, the sound model works through these visual indicators plus monologue text. Sound-as-information doesn't require sound-as-audio. The three-range model can be proven with visual indicators and text alone.

Audio direction for later: I'd defer actual audio content entirely from v0.1. The visual indicator system proves the sound model mechanically. Adding audio is additive polish, not structural. When we do add audio:

• Ambient: low drone, station hum, ventilation. The Commonwealth is a machine. It breathes.
• Close range: footsteps on metal, door mechanisms, muffled conversation bleed.
• No music in gameplay. Music in menus/loading only. The station's ambient sound IS the soundtrack.

---

Summary: Visual Design Deliverables for v0.1

Documents to produce

Document	Priority	Description
v0.1 Visual Grammar (1 page)	Critical	Color palette, entity sizing, label rules, UI layer separation, fog rendering spec. Everything above in Layer 1, codified.
Spatial Layout: Logistics Hub	High	Floor plan, sightlines, NPC positions, approach vectors
Spatial Layout: Bar	High	Floor plan, sightlines, NPC positions, approach vectors
Spatial Layout: Smuggling spaces	High	Floor plan, sightlines, NPC positions, approach vectors
Insert/HUD Wireframe	High	Minimap position, text areas, notification zones, monologue display area
Monologue Presentation Spec	High	Typography, animation, stacking, priority rules
Sound Indicator Visual Spec	Medium	Fog-edge indicators, color coding, pulse behavior

Proposed new tickets

Title	Priority	Rationale
v0.1 Visual Grammar document	Critical	Everything visual depends on this. Must exist before any renderer work begins.
Spatial layout design - 3 social site templates	High	Sightlines and room layout are the physical foundation of the observation mechanic. Parallel with content packs.
Insert/HUD wireframe and spec	High	The insert is the player's primary UI. Needs to be designed before implemented.
Monologue display system spec	High	Monologue is the primary content delivery mechanism. Its rendering rules affect every other visual element.
Sound indicator visual design	Medium	Required for the three-range model to function without audio files.
Entity color system implementation	High	The relationship-to-player color mapping is the primary visual expression of asymmetric information.

Dependencies

Visual Grammar ──> Spatial Layouts ──> Content Packs (layout affects what content makes sense)
Visual Grammar ──> Insert/HUD Wireframe ──> Insert implementation
Visual Grammar ──> Entity Color System ──> Fog rendering, NPC rendering
Visual Grammar ──> Monologue Display Spec ──> Monologue system UI

The Visual Grammar document is the root dependency for all visual work. I can produce it quickly once this workshop concludes and we have consensus on the color-as-relationship-state approach.

One strong opinion

The entity color system (color = relationship to player, not object type) is the single most important visual decision for v0.1. It's how asymmetric information becomes visible. It's how two characters in the same room see different things. It's the visual proof of concept. If we get nothing else right visually, get this right. A teal rectangle labeled "Worker" for the detective and a green rectangle labeled "Jax" for the smuggler - that IS the game, rendered at minimum fidelity.

Everything else is refinement. This is the foundation.

---

Written by Araminta. Keep it clean, we can layer detail later.