Production Brief — WO-0005: Floor Plan Optimizer

Production Brief — WO-0005: Floor Plan Optimizer

Stage: 1 — Client Intake Owner: Camille (Client Intake Manager) Date: 2026-03-28

Client Profile

Field Value
Client Bruno Bock (CEO, internal — dogfooding)
Segment Personal residential architecture
Domain Residential spatial planning
Methodology Custom residential metamodel (metamodel-maison.md)
Language Bilingual FR/EN (UI English, room labels French)
AI platform N/A — standalone browser tool, no AI runtime
Users 1 (client only)
Documentation platform Local filesystem + Architecture Maison project
Naming conventions PRJ-NNNN, KDD-NNN per Architecture Maison standards

Discovery Summary

Business Context

Bruno is designing a new residential construction project (PRJ-0001-programme-spatial). He has existing PDF floor plans (pages A11 = ground floor, A12 = 2nd floor) and wants to explore mathematically how rooms can be rearranged within the fixed building envelope. Current workflow is manual iteration in Draw.io — slow and doesn't systematically explore all options.

Pain Points

  1. Manual iteration is slow — Draw.io requires redrawing each layout variation by hand
  2. Combinatorial blindness — no way to see ALL valid arrangements, only the ones manually imagined
  3. No comparison view — can't easily compare multiple layout options side-by-side with metrics

Current State

  • Floor plans exist as PDF scans (permit drawings A11, A12)
  • Draw.io selected as primary 2D plan tool (KDD-001)
  • Sweet Home 3D used for 3D visualization (.sh3d files exist)
  • Residential metamodel defines 7 object types: Actor, Life Function, Space/Room, Infrastructure, Constraint, Spatial Relationship, Realization Phase
  • Room types catalogued: Main, Secondary, Service, Circulation, Transition, Covered/Open exterior, Technical
  • Design conventions defined: scale (20px = 1 foot), wall colors, room color palette by type

Product Definition

Product Type

Interactive single-page HTML tool (standalone, no server, no AI dependency). Runs 100% client-side in browser.

Core Capabilities

# Capability Description Priority
1 Floor plan canvas Upload 2D floor plan image (PNG/JPG from PDF export), display as scalable interactive background Must
2 Room zone editor Draw rectangular room zones on the plan with name, function type, dimensions (W × L in meters) Must
3 Building envelope Define fixed exterior walls as the immovable constraint boundary Must
4 Constraint engine Set min/max room sizes, adjacency requirements (kitchen↔dining, bedrooms↔bath), fixed elements (stairs, entries) Must
5 Permutation generator Mathematically generate all valid room-assignment combinations within constraints Must
6 Layout comparator Side-by-side visual comparison of generated layouts with area utilization % Must
7 Room color coding Apply Architecture Maison color palette by room type (yellow=living, blue=bedroom, cyan=bath, etc.) Should
8 Export/save Save room definitions and generated layouts as JSON for reload Should
9 PRJ-0001 presets Pre-load room data from the existing programme-spatial project Nice

Use Cases

  1. Explore RDC (ground floor) layouts — Upload A11 plan, define rooms, generate all valid arrangements for kitchen/living/dining/entry zones
  2. Explore 2nd floor layouts — Upload A12 plan, generate bedroom/bathroom/office arrangement options with stair position fixed
  3. Compare side-by-side — View 3-4 layout options simultaneously to identify the best spatial flow
  4. What-if analysis — Change one room's min size and instantly see how it constrains the other arrangements

Scope

  • In: Image upload, room drawing, constraint definition, permutation math, visual comparison, area metrics, JSON save/load
  • Out: 3D rendering, structural engineering, cost estimation, ML image recognition, Sweet Home 3D integration, Draw.io import/export
  • Compliance: No external API calls. No data leaves the browser. 100% offline-capable.

Feasibility Assessment

Risk: Low. This is a combinatorics + canvas rendering problem. No AI, no integrations, no regulatory concerns.

Dimension Assessment
Technical complexity Medium — canvas drawing + constraint-based permutation algorithm
Methodology compatibility Full — residential metamodel provides room types and constraint categories
Tool integration None needed — standalone HTML
Blockers None identified

Macroscope Decision

Not needed. Single user, personal use, no integrations, no regulatory environment, no enterprise scope.

Success Criteria

  • 30-day: Bruno has used the tool to explore 3+ layout variations for his ground floor (A11)
  • 90-day: Tool listed in factory portfolio; production pipeline lessons captured in retrospective

Gate 1 Checklist

  • Product type identified — Interactive single-page HTML tool
  • Client domain and methodology identified — Residential spatial planning, metamodel-maison.md
  • Working language confirmed — Bilingual FR/EN
  • Target tools/platforms identified — Desktop browser, standalone
  • At least 3 use cases described — 4 use cases defined
  • Scope boundaries defined — In/Out/Compliance documented

Gate 1 verdict: PASS — Ready for Stage 2 (Design)