jpmschweitzerandClaude Opus 4.6
1014680ebb
Standardized YAML frontmatter on all 26 docs/discussions/ files with title, description, type, status, round number, and created date. All marked as archived. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
65 lines
3.8 KiB
Markdown
65 lines
3.8 KiB
Markdown
---
|
|
title: "Round 12: Top-Down Final"
|
|
description: "Final gut-check confirming top-down over first-person 3D; systems-interaction visibility wins over immersion trade-offs; decided D-019"
|
|
type: discussion
|
|
status: archived
|
|
round: 12
|
|
created: 2026-02-09
|
|
---
|
|
|
|
# Round 12: "Final gut-check on top-down vs first-person"
|
|
|
|
**Team Leader (Jeroen):** One last challenge before we lock it in. What are we missing out on fun/immersion-wise if we don't go first-person? I'm set on the choice, just have a nagging doubt.
|
|
|
|
### OZZIE (Player Experience)
|
|
|
|
Honest answer: first-person would be more immersive in the moments that matter most. Walking through a wormhole. Sitting across from a liar. Peeking around a corner in a firefight. The terror of hearing footsteps behind you. These are visceral in first-person. In top-down they're abstracted. We're losing something real.
|
|
|
|
### PAULA (Narrative & Political Depth)
|
|
|
|
Conversation scenes. Disco Elysium works partly because you SEE the person. Body language, the way someone shifts when pushed on a lie. Top-down, it's two sprites and a text box.
|
|
|
|
### MIRI (Lore Expert)
|
|
|
|
Walking through the wormhole. In the books it's a sensory experience - light changes, air smells different, gravity shifts. Top-down, it's a loading screen between maps.
|
|
|
|
### GORE (Themes & Endgame)
|
|
|
|
Ascension moments. Going Higher should feel like senses expanding. First-person could literally render new visual spectrums. Top-down, it's a stat change.
|
|
|
|
### TYRE (Technical Architect)
|
|
|
|
Now the cost column: 3D environments = months before testing a single mechanic. Tactical awareness dies in first-person. Political/strategic layer fights with the camera. NPC simulation becomes invisible. Time-to-prototype multiplies 5-10x.
|
|
|
|
### GESTALT (Systems Design)
|
|
|
|
The fun of this game is SYSTEMS INTERACTING. See character enter building → check insert → thermal scan → monologue fires → mental note → later exploit camera hack. That sequence is BETTER top-down because you observe multiple layers simultaneously. First-person shows you a building facade and nothing else.
|
|
|
|
### PAULA (Narrative & Political Depth)
|
|
|
|
Actually... the conversation thing? Monologue solves it differently. First-person shows a face. Monologue shows your character's INTERPRETATION. *"She's lying. The micro-expressions are wrong - I've seen this pattern, centuries ago."* That's richer. And more faithful to Hamilton - he writes what characters THINK about faces, not what faces look like.
|
|
|
|
### TYRE (Technical Architect)
|
|
|
|
Architecture supports a first-person client later. Client-server separation means the renderer is swappable. Top-down now ≠ top-down forever.
|
|
|
|
**Team Leader (Jeroen):** Tyre makes a good point about different clients. There's a great middle ground: we can switch to 3D for cutscenes. I'm sold.
|
|
|
|
### OZZIE (Player Experience)
|
|
|
|
There it is. Top-down for detective work, 3D cutscene for the wormhole walk-through. That's my moment back. The Dyson barrier opening. First contact. Moments that SHOULD feel different because the world just changed.
|
|
|
|
### TYRE (Technical Architect)
|
|
|
|
3D cutscenes are decoupled from gameplay engineering. Can be pre-rendered, can be simple, can be added AFTER core game works. They're polish, not architecture.
|
|
|
|
### GESTALT (Systems Design)
|
|
|
|
Camera change to 3D IS the dramatic signal. Player knows: something significant is happening.
|
|
|
|
### SCRIBE (Documenter)
|
|
|
|
Confirmed as D-019 (top-down final, 3D cutscenes for key moments). Camera debate closed. Full honest tradeoff analysis preserved in the record.
|
|
|
|
*Engine selection. NOW. Hard requirements: client-server (D-010), deterministic sim (D-010), chunk-based maps (D-012), LOS shadowcasting with vision cone (D-011/D-015), sound propagation (D-018), observer-based perception queries (D-017), 2D top-down primary with potential 3D cutscenes (D-019). Q-001 is next.*
|