Word Wall
An interactive poetry installation built from waterjet-cut aluminum, spring-mounted word buttons, LEDs, and custom wiring.
Word Wall is a large interactive poetry installation made from 128 physical word buttons. Each word is cut through a 6 foot by 6 foot aluminum front panel. Behind each word is a spring-mounted plexiglass button, custom LED lighting, and a switch that lets people press words to build glowing poetic sequences.
The piece grew out of my work with custom controllers, live visuals, and multi-button grid interfaces. I was interested in what would happen if a button grid became more than a control surface. Instead of triggering media clips or software actions, the buttons would become words. People could press them, arrange meaning through interaction, and watch the wall respond with light.
The interaction is intentionally simple. Visitors press words, build a sequence, and the wall plays those words back over time. The effect is somewhere between fridge magnet poetry, haiku, electronic signage, and a physical interface.
I conceived, designed, fabricated, wired, installed, and presented the project. My work included:
- Developing the concept and interaction model
- Creating the 128-word list and testing word combinations
- Designing the aluminum front panel layout
- Editing the stencil-style lettering for durability and readability
- Researching fabrication methods
- Sourcing springs, bolts, plexiglass, LEDs, switches, wire, and structural materials
- Designing and iterating the button mechanism
- Building and testing LED lighting assemblies
- Leading soldering and assembly sessions with volunteers
- Installing and presenting the piece at events
- Observing how people interacted with it and identifying future improvements
The challenge was turning a simple interaction idea into a working physical object at a large scale.
The piece needed to be inviting, durable, transportable, visually clear, and responsive enough that people would understand what to do without a long explanation. It also required learning unfamiliar fabrication and electronics processes while managing a large amount of manual labor.
Every word was a unique physical object. Different word lengths affected button size, mounting points, lighting coverage, wiring, and durability. The wall also needed to survive transport, installation, repeated pressing, and public interaction by both adults and children.
The project started with the word list. I wanted the words to suggest a Pacific Northwest feeling, but early lists pulled from common internet words felt too generic. I started testing sentence structures and word combinations, arranging printed words into short poems and phrases to see which combinations created the right kind of flexibility. That process shifted the project toward something closer to haiku or fridge poetry. The words needed to be specific enough to create mood, but open enough that visitors could create many different sequences.
I created the full panel layout in Adobe Illustrator and Inkscape, including the word placement, cut lettering, mounting holes, and button locations. Because the letters were cut through aluminum, the typeface needed to work as a stencil. Letters like O and A needed bridges so the interior shapes would not fall out. I edited the font to strengthen delicate areas, improve consistency, and make the cut letters durable enough for transport and public use.
The first button concept used bolts, springs, plastic sleeves, washers, and electrical contact points. It worked in theory, but it was too fiddly and unreliable over time. The better solution kept the springs and bolts but used them differently. Each plexiglass word panel was mounted so the lower connection points acted like a hinge. Pressing the top of the panel moved it forward into a standard lever switch behind the front panel. This created a simpler and more reliable mechanical button. Each button still required tuning for spring height, firmness, and alignment, but once adjusted, the mechanism became much more stable.
Early LED tests behind the white plexiglass looked promising, but the different word lengths made lighting complicated. A single LED assembly would not work for every word. With help from a friend who had surface-mount LED strips and loose LEDs, I tested different color and spacing combinations. A general pattern of two orange LEDs around one yellow LED created a warm glow that worked well behind the plexiglass. Each word used at least one three-LED section, while longer words used multiple sections. The goal was to keep the voltage requirements consistent while adapting the lighting coverage to each word.
The aluminum panel was cut by an industrial waterjet company in Seattle. I chose waterjet cutting after evaluating other fabrication options, including laser cutting. The fabrication process involved a significant amount of manual work: cutting and preparing materials, mounting switches and LEDs, assembling button panels, soldering LED sections, wiring buttons, tuning spring action, and building the supporting structure. Because the words were not arranged in a perfect horizontal grid, the wiring could not follow a simple row-and-column structure. This made the wiring messier and more labor-intensive than a standard button matrix, but it preserved the visual rhythm of the word layout.
The main interaction was direct and tactile. People did not use a screen or separate controller. The words themselves were the interface.
A strict grid would have made wiring easier, but it would have made the piece feel more mechanical. I chose a less uniform layout so the wall would feel more like language, rhythm, and composition.
The first button mechanism was unreliable, so I redesigned it around a hinged plexiglass panel and a standard lever switch. This made the interaction more durable and repeatable.
The yellow-orange lighting gave the wall a warmer, more inviting presence than a colder electronic glow. It helped the piece feel like an object in a gallery, not just a control panel.
The wall was large, so it was built in two pieces for transport and installation. When placed side by side, the halves created the full wall while keeping the piece more manageable to move and assemble.
Word Wall became a large-scale interactive sculpture with 128 illuminated word buttons. Visitors could press words, build sequences, and watch the wall play those words back, creating a changing poetic experience over time.
The piece was partially funded by Ignition Northwest and installed as a work-in-progress at Critical Massive. Additional support came from a Seattle artist brunch series. It was shown at multiple events and was especially popular with both children and adults, who were drawn to the glowing words and the simplicity of pressing language directly.
The project demonstrated a long-running thread in my work: interaction design beyond screens. It combined physical interface design, fabrication, electronics, language, public interaction, and observation of how people engage with technology in shared spaces.
Word Wall was one of the most complex physical projects I had taken on at the time. It required learning new fabrication methods, estimating unfamiliar labor, sourcing materials, coordinating volunteer help, and solving mechanical and electrical problems as they emerged. The biggest lessons were about scope, repeatability, and planning for the realities of physical work.
The amount of manual labor was much larger than expected. Future versions would need more realistic time estimates for soldering, assembly, tuning, wiring, and installation.
Access to a friend's shop was essential for early fabrication, but it had not been fully planned into the project. Future projects should budget for tool time, workspace, and fabrication support from the start.
The wall was split into two halves, but it was still difficult to move. Future versions should be designed around expected vehicles, installation constraints, and modular transport.
The first electronics approach was too fragile and specific. For future versions, I would use more standard, easy-to-replace electronics and plan the wiring system earlier in the design process.
Because so much of the project depended on custom fabrication and manual assembly, creating additional versions would be difficult. I started thinking about a future version with fewer mechanical parts, simpler wiring, and a more repeatable construction method.
Waterjet cutting produced softer edges than expected, which turned out to be safer for touch interaction. Still, future projects should test unfamiliar fabrication processes at a small scale before committing to a full-size final panel.