I immediately recognised many of the commands, both their names and byte values, as exactly the same as those of the Philips PCF8833 from 2 decades ago, an extremely common LCD controller for the tiny displays on mobile phones of the time.
The displays have 320 x 320px square addressable pixels, but only the circular portion is displayed - that is to say you can draw pixels as though they were there in the lower corner, past the radius of the circle, but nothing gets drawn.
Always thought round LCDs (and rounded corners on displays and now GUI windows) were stupid, and this explains exactly why. What would otherwise be perfectly usable pixels are missing, and the panel itself is still square.
Is there some sort of ideological imperative to use every single pixel? If your product design doesn’t call for it because of low information density, who cares.
> Always thought round LCDs (and rounded corners on displays and now GUI windows) were stupid
If you've ever used a Nest thermostat before, you'd understand why a square display would be stupider.
It would either:
1. force a larger interface so the square display could have its diagonal fully enclosed by the diameter of the radial control (i.e., oafishly large thermostat too big for a human hand to easily manipulate)
2. force a smaller LCD to fit inside a normal hand-sized radial control, making it less readable to all but the spriteliest of youths
3. make a radial control that is a spinning rhombus, which is pretty ugly
Or we could just do what the Nest does and the only person who "suffers" is the original designer, one time, when they write the code
But Nest is using a square display. They are just hiding the corners behind plastic with round hole, which according to you would mean either 1 or 2.
The article says the display is physically round.
The pictures tell a different story.
The display itself is round, although the module is a square with trapezoidal corners.
This is odd because you get the intended shape but not the benefit of the shape. There are plenty of displays that are actually round, such as for smart watches.
I'm curious if Google did this for cost and had decided on a larger bezel from the start.
My guess: it's much easier to cut straight lines. If you don't need it, why go through the more complicated step to cut it circularly.
I don't think they're stupid, especially if the application isn't particularly information dense.
Like... this is a thing meant to be operated at arm's length. Some extra pixels aren't going to meaningfully improve how you view the current temperature.