Clarence Ma

Between the Columns

Organising a Constrained Event Space

Design Phase 01

Plan of the warehouse hall, from the problem to The Lock A 30 by 18 metre hall with twelve columns on a 6.0 by 4.5 metre grid. The columns hide a long-wall stage from 59 percent of the floor. Three questions are read on it under one set of rules: the stage on the long wall (4.4 metres mean distance) rather than the short wall (8.6 metres); bars at the edges (68 percent clear) rather than an island bar (72 percent, but its queue stands in the room); and the arrival, by the side door or through the loading door and a lock, left open. The Lock is then drafted on it: a 4.20 metre arrival street behind column row C entered through a lock on the stage axis, an 8 by 4 metre stage across the central bay, bars and art in the column shadows, one platform for wheelchair viewing and the mix desk (27.94 square metres complete, up from 19.05 as a simple block), and hooded lamps pooling light over the bar queues. W DOOR E DOOR LOADING DOOR 1 2 3 4 A B C SERVICE ROOM TOILETS W DOOR E DOOR LOADING DOOR 1 2 3 4 A B C SERVICE ROOM TOILETS IF THE STAGE GOES HERE 1 SHORT WALL 8.6 m LONG WALL 4.4 m 2 ISLAND BAR 72% clear QUEUE IN THE ROOM EDGE BAR EDGE BAR 68% clear 3 SIDE DOOR OR THROUGH A LOCK tested next ART 5 · LOCK STREET EXIT EXIT STAGE +0.60 A2 · A3 ON STAGE BACKSTAGE LED SCREEN MAIN BAR SATELLITE BAR ART 2 ART 3 · WALL ART 1 AS A DIAGRAM 19.05 m² SEATED 3 WS + 3 C +0.30 FOH +0.60 RAMP 1:12 KEEP CLEAR STEP-FREE ENTRY BUILT 27.94 m² L1 L1 OPEN EDGE OPEN EDGE PERFORMANCE VIEWING CIRCULATION SUPPORT 1 2 3 4 5 1 2 3 4 The Lock with the simulated guests' paths drawn over it Plan of The Lock, the 30 by 18 metre hall with its stage on the north wall and the lock inside the loading door on the south wall. Over it, the crowd simulation's guest paths, from seed 1. In the opening surge, 180 arrivals queue in the street outside the loading door, pass through the lock, split east and west along the arrival street behind the crowd and fan out to their viewing spots in front of the stage. Among them, three wheelchair users, each with a companion, come through the lock's step-free wicket and take the railed lane up to the wheelchair row on the platform; 3 of 3 are in their space at the end. In the counterflow test, 45 guests arrive through the lock while 45 leave by the west side door, so the two streams use different doors. W DOOR E DOOR LOADING DOOR 1 2 3 4 A B C SERVICE ROOM TOILETS ART 5 · LOCK STREET EXIT EXIT STAGE +0.60 A2 · A3 ON STAGE BACKSTAGE LED SCREEN MAIN BAR SATELLITE BAR ART 2 ART 3 · WALL ART 1 SEATED 3 WS + 3 C +0.30 FOH +0.60 RAMP 1:12 KEEP CLEAR STEP-FREE ENTRY L1 L1 OPEN EDGE OPEN EDGE PERFORMANCE VIEWING CIRCULATION SUPPORT KEYarrivingleavingwatching, at the endstanding still on the waywheelchair user and companionBodies at true scale · seed 1 of 5 · step-free wicket

Intro

Twelve columns in the way

Between the Columns

Notes

A design study, not a delivered event: a music and art night for 180 guests in a hypothetical 30 × 18 m warehouse. I drew the venue, set the scoring rules and designed the entrance, then tested it in a crowd simulation I specified and directed (built with an AI coding assistant). Its twelve fixed columns hide a long-wall stage from 59% of the floor, so every layout has to work around them.

Outcome: The Lock, chosen for its arrival; similar spatial performance to the control on the plan, about 1.4 s at the rush in the model. Case Study Deck (opens in a new tab)

Working

3,722 ÷ 6,336 = 59% of the floor in shadow

Step 1 of 3

Stage, bar and entry

Between the Columns

Notes

Which wall for the stage? The long wall: 4.4 m mean distance against 8.6 m. Where do people drink? At the edges: the island bar sees a little more (72% against 68% clear) but puts its queue in the room. How do guests arrive? The side door, or the loading door and a lock: the question tested next.

Working

mean distance: long wall 4.4 m · short wall 8.6 m

Step 2 of 3

One change to the control

Between the Columns

Notes

The Lock keeps the control's long-wall stage and makes the loading door the arrival: guests come in on the stage axis through a light-and-sound lock, which gives a partial view of the hall before it opens, then split along a street behind the crowd. Bars and art take the floor the columns already hide.

Working

street 4.20 m · bars + art 86% in shadow

Step 3 of 3

The turning point: access at full size

Between the Columns

Notes

The early plans held access as a 19.05 m² block. Drawn at full size to the 2010 ADA Standards (three wheelchair spaces with a companion seat beside each, a 1:12 ramp), it needs 27.94 m², so every layout was rescored. The mix desk, first sketched on the viewing riser where it took a wheelchair space, moved to its own pod behind the row. The cost is floor: the ramp takes space the art had used. Three spaces is a study assumption, not approved provision: ADA's count (five for 151–300) is written for fixed seats; how it applies to a standing crowd of 180 is for an access reviewer to settle. Lamps on the bar-side columns edge the queues with light, not barriers.

Working

19.05 → 27.94 m² once complete

Summary

Similar spatial performance, a different arrival

Between the Columns

Five moves

  1. Arrivals through the lock. A 4.20 m street behind the crowd.
  2. Stage on the long wall. 4.4 m mean distance.
  3. Bars and art in the shadow. 86% of their floor.
  4. Access at full size. 27.94 m², every layout rescored.
  5. Light, not barriers. The column gaps stay open.

Three plan measures, The Lock against the control: audience floor (standing area left after the stage, bars, art and platforms) 108 m² against 107 m²; share of it with a clear view of the stage past the columns and art 67% against 68%; mean distance to the stage 4.4 m in both. So the entry was chosen for the arrival it creates, not for a better number. In the crowd simulation, the lock adds about 1.4 s at the opening rush, keeps walkers arriving and leaving at separate doors and gets all three wheelchair users to their spaces. Still open: the bar queues and the number of wheelchair spaces.

Step 1 of 2

The opening rush through the lock

Between the Columns

Notes

180 simulated guests arrive at once through the loading door, three wheelchair users and their companions among them. Each line is one simulated guest's path, from the street, through the lock, to the spot they watch from. Red marks a guest standing still on the way in. The wheelchair users, in black, take the railed lane to their spaces on the platform.

Working

9 in 10 guests are out of the 5 × 5 m square at the door within 4.2 s of crossing it · control 2.8 s

Step 2 of 2

Arrivals and departures kept apart

Between the Columns

Notes

A stress test: 45 simulated guests come in through the lock while 45 leave, all starting gathered by the west side door, the worst case for it. Solid lines arrive, dotted lines leave, and the two streams never share a door. At the control both use the west door, and its railed lane closes the way in beside column C1: in the model 12–18 of 45 entrants are sent back round the building (real guests would most likely wait). With leavers spread to both side doors, its entrants are in within 2.9 s.

Working

9 in 10 entrants are out of the 5 × 5 m square at the door within 3.1 s, while 45 leave

Summary

Why I chose The Lock

Between the Columns

What I tested

What delay does the lock add in the model, and which access questions stay open?

  1. Every simulated arrival got in. All 75 runs with arrivals.
  2. About 1.4 s at the rush. 4.2 s against 2.8 s to clear the door's 5 × 5 m square, 9 in 10 guests.
  3. Directions kept apart. Walkers never share a door. Wheelchair users pass leavers only if they use the west door.
  4. All three wheelchair users placed. Each with a view over standing heads, in all 90 runs.

Similar on the plan, so I chose The Lock for its arrival: a partial view through the lock before the hall opens.

Simulated guests under one assumed movement routine, not calibrated against real crowds: 90 runs, seed 1 drawn; door times are 90th percentiles, the median of 5 runs. A comparison in a model; it certifies nothing. Access questions still open: the wicket or the west door, the bar queue by the accessible WC, and whether three wheelchair spaces is enough. Before delivery Case Study Deck (opens in a new tab)

  1. Design
  2. Problem
  3. Options
  4. The Lock
  5. Access
  6. Result
  7. Simulation
  8. Surge
  9. Counterflow
  10. Result

Open items

Before delivery

What has to be resolved before this could be delivered. The roles are proposed; nobody is appointed.

Open questionWho would take it onWhat would settle it, or the fallback
Wheelchair entrance: the wicket or the west doorEvent coordinator, with the venue team and a qualified access reviewerConfirm whether the wicket is step-free. Wicket: no clash with leavers, but walkers behind a wheelchair wait about 1.4 s. West door: the slowest wheelchair user reaches their space in 46 s against 53 s, but passes leavers there (13 close passes a run in the stress test, 1 when leavers spread to both doors).
Number of wheelchair spacesEvent coordinator, with a qualified access reviewer and the designerConfirm how the count applies to a standing venue. If it is five, the platform most likely grows and every layout is rescored again.
Bar queue by the accessible WCBar team lead and floor host; the event coordinator decides on extra serviceRehearse a surge with the satellite bar closed and check the queue stays inside its belt barriers and off the WC route. Fallback: pause new joins, and open the satellite bar early if staff allow.
The light-and-sound lockProduction or technical lead; the door host on the nightScope, cost and how it runs on the night. Fallback: a simpler arrival through the same door.
Turning space at the lane mouthRamp steward, with the floor hostMark and keep a clear apron, about 1.8 m across, in front of the wheelchair lane; the steward redirects guests who gather at the ramp approach.

Back to the case study · Case Study Deck (opens in a new tab)