A / TEST
Fragmented test
Staggered partitions and denser decision load.
- Decision points
- 10
- Ambiguity zones
- 07
ALTN VISION
2026
Spatial
Critic.
An architectural wayfinding experiment that turns a plan into a testable, walkable and evidence-conscious research system.
- Discipline
- Spatial research
- Role
- Concept, system, interface
- Format
- Browser-based prototype
- Status
- Pilot ready · 2026
01 / THE BRIEF
Can a plan critique its own wayfinding before it is built?
Architectural plans are usually reviewed through drawings, precedent and expert judgement. Spatial Critic asks whether an early plan can also become a transparent navigation experiment—one that exposes where a route becomes longer, less direct or more ambiguous.
The project deliberately avoids claiming that a model understands human comfort. It begins with inspectable geometry, authored assumptions and measurable behaviour, then prepares those assumptions for comparison with real participants.
02 / CONTROLLED COMPARISON
One envelope.
Two spatial logics.
The comparison holds boundary, scale, walkable area, entry and destinations constant. Only partition arrangement and authored decision load change.
B / CONTROL
Legible control
Aligned circulation spines and lower decision load.
- Decision points
- 05
- Ambiguity zones
- 04
16 × 12 grid
1.5 m per cell
112 walkable cells
3 shared destinations
03 / THE SYSTEM
From plan geometry
to behavioural evidence.
- 01Define
Translate walls, destinations and decision zones into an inspectable grid.
- 02Simulate
Compare shortest and risk-aware route baselines without hiding their costs.
- 03Walk
Let participants navigate the same geometry in a neutral first-person view.
- 04Measure
Record anonymous traces, pauses, backtracks, contact and recovery.
- 05Compare
Separate cohorts by task and condition before drawing any conclusion.
04 / EVIDENCE FRAME
Measure behaviour.
Measure behaviour.
Do not invent emotion.
Time spent leaving the first authored decision cell.
Effective trace gaps of at least 1.5 seconds.
Distance travelled while re-entering previously visited space.
Continuous collision contact counted as a single event.
Whether a backtrack eventually reaches an unvisited cell.
A required participant rating after each completed task.
05 / INTERFACE ETHICS
One system.
Two protected views.
Participants never see spatial-condition names, authored risk or performance metrics. Researchers never enter analysis while a study is in progress.
06 / CURRENT STATUS
A working prototype,
not a finished claim.
- Two controlled, reachable spatial conditions
- Transparent route baselines and behavioural metrics
- Anonymous, counterbalanced pilot workflow
- Participant and researcher role separation
- No claim about human comfort or emotion
- Authored ambiguity remains an explicit assumption
- Real participant evidence has not yet been collected
- Next step: a small consent-based feasibility pilot