WEEK 6

User Testing

What each session produced

Nine participants wrote the full alphabet using the interface. Each session was documented through:

How participants described it

Reinterpretation of movement, not handwriting mimicry

Across responses, participants consistently described the glyphs as reinterpretations of their movement rather than reconstructions of their handwriting. Several expressed surprise and curiosity. The strongest moment was recognition: the sense that the system was reflecting something personal.

• “It is not trying to copy my handwriting exactly. Instead, it takes parts of my movement and turns them into a new geometric form.”
  
• “My handwriting was being abstracted into data and then rebuilt into another visual system based on a set of rules.”
  

This confirms that the system is being read as a translation of embodied action rather than as stylised typography. This aligns precisely with the system’s intention: behavioural abstraction governed by authored rules.

A hierarchy emerged

User feedback revealed a hierarchy of signal clarity.

• Most legible:
  Triangles (directional pivots)
• Moderately legible:
  Rectangle breaks (hesitation)
• Weakest:
  Circles (pressure peaks)

Participants clearly recognised triangles at sharp turns. These moments were consistently identified as linked to specific directional changes. Rectangle breaks were generally understood as moments of interruption or slowdown, but were sometimes visually subtle. Circles, however, occasionally felt disconnected from the backbone structure.

• “The circles. It seems disconnected from the rectangles giving me the impression that my movement ended abruptly.”
  
• This indicates a scaling and integration issue rather than a conceptual flaw. Pressure detection works, but its geometric translation needs refinement.

Same letter, different signatures

Nine versions of “g”

To evaluate consistency, I compared nine versions of the same letter “g” across participants. Although the letterform was identical in name, the behavioural glyphs varied significantly: different pivot points, different hesitation locations, and different pressure anchors. The variation reflected individual writing tendencies. This comparison visually validates the system’s ability to differentiate behavioural signatures across writers.

Participants wanted to test the system on purpose

Several participants expressed curiosity about how the glyph would change if they altered their writing style. Slower strokes, sharper turns, or heavier pressure were expected to produce visible differences. The presence of export functionality and the desire to experiment indicate that the system is functioning as a tool, not just a visual artefact. It invites iteration.

When the backbone starts to look like stylised typography

The most significant issue revealed in Week 6 was scaling imbalance. Rectangular backbones sometimes became visually dominant, causing glyphs to resemble stylised alphabets rather than behavioural reconstructions. When this happened, behavioural signals were less legible. This highlights a need to:

• Recalibrate base
  width proportions
• Refine circles
  integration further
• Maintain behavioural priority
  over letter resemblance