Facet, a cursor you hold, not one you push
Product and Cognitive Ergonomics · Industrial design · NID Bengaluru · October–November 2025
Intro
A handheld pointing device sized from the 5th percentile female hand upward, built to prove a simple position: most input hardware is dimensioned for the largest user and then asked to work for everyone else. This one starts at the other end of the distribution.
Controller Design was an industrial design exercise at NID Bengaluru: design a handheld device that moves a cursor and selects on a GUI, using anthropometric data and Henry Dreyfuss’s Humanscale to set its dimensions. I treated it as a sizing problem before a styling problem. Every surface on the final object is traceable to a percentile decision, and the form that resulted is faceted rather than sculpted because the hand needed landing zones more than it needed curves.
Concept film30 sec
Impact
What the project produced
Four outcomes, written as what the project produced rather than as metrics.
5th–95th
percentile range accommodated, with one rule per parameter
40 mm
palm centre to ball, inside the 5th percentile thumb reach
86 × 65 × 30
millimetre shell, every dimension out of the percentile table
10 mm
flank buttons, 1.5 to 2.0 mm travel
A dimensioned specification that accommodates the 5th to 95th percentile range
Each parameter is assigned its own accommodation rule rather than a single blanket percentile.
A trackball placed at 40 mm from palm centre
Inside the 36 to 42 mm thumb reach envelope of the 5th percentile hand, so the smallest user in the accommodation range reaches the primary control without wrist extension.
A faceted shell that self indexes in the hand
The flats tell the fingers where to sit, which removes the orientation guesswork a symmetrical blob creates, and the planar geometry holds draft angles for moulding where compound curvature would not.
A complete design record from divergent sketching through physical form study to a dimensioned 3D model and render
Which is the part of the process the brief was actually testing.
Discovery
The brief carried a hidden constraint
The stated brief was open: a handheld controller that moves a cursor and clicks. The constraint sat in the method.
Dimensions had to come from anthropometric data and Dreyfuss’s Humanscale, which meant the form could not be decided first and justified afterward. The body had to set the envelope and the form had to live inside it.
That reversal is the whole project. In most student product work the sketch leads and the ergonomics are a retrofit. Here the percentile table was the brief.
Most student product work
- Sketch
- Form decided
- Ergonomics retrofitted
This project
- Percentile table
- Envelope
- Form inside it
Divergence
Divergence, then a deliberate pullback
I opened with wide divergent sketching from the block form developed in class: spheres with cut channels, wedge bodies with protruding stalks, cube masses with a trackball on a riser. The ideas got more interesting on paper and worse in the hand. Stalks and risers lengthened the moment arm on the wrist. Compound curvature and undercut channels meant tooling I could not justify. Several concepts required a precision grip for an action the hand wants to perform from a relaxed power grip.
I stopped and named the failure rather than designing through it: the sketches had drifted from the form language into fabrication fantasy. The pullback is a decision I would make faster now, and it is the reason the rest of the project moved quickly.
Three rounded concepts rendered to test the first direction. Interesting as objects, worse in the hand. 
A rectangular body with a d-pad around the ball, one of the riser ideas. 
Mapping movement and clicks onto the rectangular body.
The sketches were getting better as drawings and worse as objects. That gap is the thing to watch for.
Form study
Form study in the hand, not on the page
Before resolving any geometry I built a crude physical study: a wood block cut to the rough mass with a foam sphere standing in for the trackball, mounted on a stand so I could look at the silhouette and hold the mass at once. A block in the hand answers questions a three quarter view cannot: where the thenar eminence actually lands, whether the thumb falls on the ball without recruiting the wrist, how much the fingers want to wrap. The physical study is what killed the riser mounted ball and flattened the body down toward 30 mm.
The second round of ink sketching after the form study is visibly tamer than the first, and that is the point. Convergence is a result, not a style.
The form study: wood for the mass, foam for the ball, a stand to see the silhouette. 
An early rounded form model in the hand. Where the thumb lands is the question.
The second round of ink, after the block: lower, flatter, tamer.
Accommodation
Accommodation strategy
The common mistake in anthropometric work is picking one percentile and applying it to everything. Different parameters fail in different directions, so each one needs its own rule.
| Parameter | Percentile used | Why |
|---|---|---|
| Grip comfort | 5th percentile female hand | Small hands can hold it comfortably; large hands still adapt |
| Clearance and spacing | 95th percentile | Avoids cramping and contact for large hands |
| Button reach and thumb placement | Mixed 5th–95th range | Reachable by the smallest, not cramped for the largest |
Grip is a fit to the smallest. Clearance is a fit to the largest. Reach has to satisfy both ends at once, so it is the parameter that gets designed last and constrains the others.
The numbers
The percentile table
| Parameter | 5th percentile (small) | 95th percentile (large) |
|---|---|---|
| Palm breadth, excluding thumb | 65 mm | 86 mm |
| Across flats, hexagonal section | 65 mm | 86–90 mm |
| Length | 65–68 mm | 86–92 mm |
| Thickness | 22–28 mm | 30–38 mm |
| Primary button size | 10 mm | 12–14 mm |
| Primary button travel | 1.5–2.0 mm | 1.8–2.5 mm |
| Primary reach from palm centre | 36–42 mm | 47–56 mm |
| Secondary button size | 8–10 mm | 10–12 mm |
| Scroll size | 8–10 mm | 10–12 mm |
| Fillet radius | 6 mm | 8–12 mm |
| Palm crown, convex | 3 mm | 4–6 mm |
Solution
The object
An 86 × 65 × 30 mm faceted shell, held rather than pushed, with a 12 mm trackball on the upper plane and two 10 mm buttons on the near flank under the index and middle finger.
Decision 1
A trackball instead of a tracked surface
A conventional mouse converts cursor travel into arm travel. The shoulder and the wrist do the work, the device needs a desk, and the repeated lateral sweep is where ulnar deviation accumulates. Moving the sensing surface onto the object transfers that work to the thumb, fixes the forearm in a neutral position, and removes the desk from the requirement list entirely. The cost is a lower ceiling on fine positioning speed, which is the trade I accepted for the brief’s priority on comfort.
Decision 2
40 mm to the ball
The trackball centre sits 40 mm from palm centre. The 5th percentile thumb reach envelope is 36 to 42 mm, so the smallest hand in the accommodation range lands on the ball inside its comfortable arc rather than at the limit of it. The 95th percentile envelope extends to 56 mm, so a large hand reaches it early in the arc, which reads as a short throw rather than a stretch. One number, both ends of the distribution satisfied.
Decision 3
Facets over curvature
The shell is a faceted solid with a 6 mm fillet on every edge. Three reasons, in the order they mattered.
Indexing
Flats give the fingers a discrete place to land, so the device communicates its orientation through touch. A fully radiused body has to be looked at before it can be gripped correctly.
Pressure distribution
A 3 mm convex palm crown spreads load across the palm instead of concentrating it on a single contact ridge, while the flats below keep the finger pads on plane.
Manufacture
Planar faces carry draft angles and simple parting lines. The compound curvature of the early concepts would have meant tooling disproportionate to the object.
Decision 4
Buttons on the flank, not the top
Putting the primary buttons on the near flank keeps the top plane clear for the ball and puts click actuation in finger flexion, which is a stronger and lower effort movement than the extension a top mounted button requires while the thumb is already loaded. Sizing them at 10 mm with 1.5 to 2.0 mm travel places them at the 5th percentile button size and the lower end of the travel band, so actuation force stays low for smaller hands without losing the tactile detent.
Reflection
What the exercise taught me
Anthropometric data is not a constraint applied to a form, it is a form generator.
Once each parameter had its own accommodation rule, the envelope was mostly decided and my job narrowed to resolving the surface between the numbers. That is a better way to start than sketching and justifying.
What I would change
I would run the physical form study before the divergent sketching, not after. The block model answered in ten minutes what two rounds of sketching could not, and running it first would have saved the detour into forms I could not fabricate.
The next step
The specification is ready for validation: a 3D printed shell tested across a spread of hand sizes, measuring time on target, grip pressure at the palm crown, and self reported fatigue over a sustained session. I would weight the study toward the lower half of the distribution, the half the design argues for and the half where it has the most to prove.
The position the project holds
Input hardware defaults to the large hand and treats everyone below it as a tolerance. Starting the dimension chain at the 5th percentile and designing upward is a small change in method that moves who the object is for.
Next project
What an autonomous cab says when the driver's face is gone
AURAA · Multimodal interaction · 2026
