Preface

Article navigationArticle index

00Preface011. The Original IdeaNot yet published022. Why the Painting Did Not Become the 3D SceneNot yet published033. Organising the Memory Before ModellingNot yet published044. From a Relationship Map to a Metre-Based PlanNot yet published055. Grayboxing: Walking Through the MemoryNot yet published066. How Sketches Mediated the CollaborationNot yet published077. From Reference to Concept Art to GLBNot yet published088. Why the Banyan Tree Needed So Many VersionsNot yet published099. Establishing the Art DirectionNot yet published1010. Making the Wall and Ground MaterialsNot yet published1111. Materials for Standalone ModelsNot yet published1212. Bringing Textbooks, Exercise Books, and Posters into the SchoolNot yet published1313. From a Place to Look At to a Place to PlayNot yet published1414. Blackboard and Chalk: Completing a Small InteractionNot yet published1515. Finding the Performance ProblemsNot yet published1616. Optimisations That Did Not Become Production FeaturesNot yet published1717. Why the Loading Screen Had to Leave LastNot yet published1818. Testing, Archiving, and RollbackNot yet published1919. How the Human–Agent Collaboration WorkedNot yet published2020. Lessons from the ProcessNot yet published2121. Turning a Basketball Prop into an Open-Ended InteractionNot yet published2222. Turning Six Old Tables into Table-Tennis Practice and a MatchNot yet published2323. How Phone Testing Changed Both Interaction and Performance WorkNot yet published2424. How First-Hand Scene Knowledge Led to Classroom-Level Detail ManagementNot yet published2525. How an Old Handheld Became a Memory Machine the Human and Agent Could Disassemble TogetherNot yet published2626. Where the Project Stood After Roughly One WeekNot yet published2727. How Five Childhood Games Grew Out of the CampusNot yet published2828. How Two Old Slingshots Became a Natural Play CornerNot yet published2929. Why the Second Handheld Could Not Simply Copy the FirstNot yet published3030. Why Flag Raising Did Not Become a Scoring GameNot yet published3131. How an Exercise Book Remembers Activities Scattered Across the CampusNot yet published3232. Where the Project Stood After Roughly Two WeeksNot yet published33ConclusionNot yet published

Preface

This project did not begin with a complete set of architectural drawings or a finished game design document. It began with a personal memory of an old school.

That memory held a front gate, a courtyard, classroom buildings, a playground, an older classroom block, a toilet building, teachers’ quarters, and a great banyan tree. It retained the colours of the buildings, the rough sequence of doors and windows, the turn inside the toilet entrance, the way a raised terrace met the dormitory and the tree roots, and the desks, blackboards, chalk, and old textbooks inside the classrooms. But none of this arrived as a single, complete account. It was scattered across a memory painting, sketches drawn over time, a small number of historical photographs, dimensional estimates, and details that resurfaced during development.

From the outset, then, this was never a conventional exercise in “modelling from a picture.” It was closer to a form of spatial research conducted by a person and an AI agent together. The person supplied memory, judgement, and feeling. The agent organised fragmented evidence into material that could be discussed, translated uncertain relationships into drawings and parameters, implemented confirmed decisions as a 3D scene, and returned the browser experience to the person for another round of looking and correction.

The campus gradually took shape through that exchange.