The Align's screen is 128 pixels square. Small enough that it is tempting to treat it as a constraint to work around. In practice it turned out to be the most useful design brief we had, because at that size you cannot show anything you have not justified.
The budget
128 × 128 is 16,384 pixels. A modern phone has roughly two hundred times as many. Text at a legible size on a screen this small consumes an enormous fraction of it: a single line of readable characters can easily take a tenth of the total area.
So the question is never "what would be nice to show". It is "what can be removed", asked repeatedly until nothing else can go.
What survived
Three things, in this order of priority:
- State. Whether you are inside or outside your chosen limit. This is the question the device exists to answer, and it has to be readable at a glance, from an angle, in a bright room.
- Magnitude. How far. Useful, but secondary — you can act correctly knowing only the state.
- Direction. Which way you have drifted, so the correction is obvious rather than trial and error.
Everything else that was proposed — session timers, battery percentage, signal-quality graphs, history — did not survive the question "would you look at this during a procedure?"
Colour does more work than text
The screen is colour, and on a display this small colour is not decoration. It is the most information-dense channel available.
A number requires reading, which requires focus, which requires looking away from the work. A colour is apprehended without focus. At 128 pixels a colour field can occupy enough of the screen to register in peripheral vision, which no text at that size can.
So state is carried by colour and reinforced by the number, rather than the other way round. Someone who never reads the digits still gets the answer.
Why not more resolution
A bigger screen was available. We did not take it, for reasons that had little to do with cost.
The device mounts on an instrument held in a hand during precise work. Screen area is physical volume and physical mass in exactly the place where both are most expensive. A larger display would have made the device worse at the thing it is for, in exchange for showing information we had already decided not to show.
There is a rendering cost too. Every pixel has to be pushed over a serial bus, and a frame that takes longer to draw is a frame during which other work is not happening. Four times the pixels is four times the transfer, for a screen nobody is staring at.
Partial updates matter more than frame rate
The cheapest frame is the one you do not draw.
Most of what is on screen does not change between frames. Redrawing all of it to update a number that moved by a degree is waste, and the waste shows up as latency elsewhere in the system. Redrawing only the region that actually changed keeps the display responsive without the rest of the device paying for it.
This is unglamorous work and it is where most of the display effort went. The visible result is not "the screen looks good" — it is "the screen keeps up, and the sensor is never kept waiting".
The transferable part
A small screen forces a question a large one lets you avoid: what does the person actually need to know, right now, in the situation they are in?
On a phone you can show the answer, and the context, and the history, and the settings — and never find out which of them mattered. At 128 pixels you find out immediately, because there is only room for the thing that matters.
We would not have arrived at a three-element display by starting with a large screen and simplifying. We arrived at it because there was no alternative, and the device is better for it.