The margin of error on this system, running on mobile devices with inconsistent and unpredictable OS and network latency/overhead must limit its usefulness.

It's cool though, and nice to see another .space in the wild.

It's also very easy to fake/attack. Sybils in this system would not be difficult to spin up.

How? nodes validate each other in the mesh.

Just browsed it real quick. They claim you can locate something in 3D with 4 sensors, but as configured they can't differentiate between any positive Z distance and any negative Z distance.

You actually can even if all the sources / transmitters are on the same plane. It's pretty easy to conceptualize as well.

With one measuring point and it's distance to a fixed target, you know that the target lies somewhere on the surface of a sphere with a radius equal to the distance.

With a second measuring point and it's distance to the same fixed target you can overlap those two spheres. Two spheres that overlap do so as a circle. You now know the target lies somewhere on the circumference of the circle.

With a third measuring point and it's distance, it's sphere will intersect the circle at precisely two points. This is the minimum required for GPS as it can usually discard one of the points using other knowledge (your receiver is not in space).

With a fourth measuring location the ambiguity is removed and you're left with a single location.

The orientation and location of the spheres don't matter as long as you know the precise locations of the measuring points, all measurements are taken while the target is fixed (which is why time synchronization matters here), and all the spheres overlap (in the real world noise, and mulitpath reflections make this quite difficult).

Source: I worked on wireless trilateration for several years for a commercial project.

GPS works the same way - it locates you at one of two points, one in space and one on the surface of the planet. The GPS receiver then assumes you're not in space.

I think the assumption that the 4 points are on the plane was only to make showing it easy. In reality, I would imagine there is more z axis variance just due to the nature of reality.

This video clearly explains the root of the problem:

https://www.youtube.com/watch?v=wteiuxyqtoM

More on FOAM's use of Time Synchronization for dynamic proof of location here: https://github.com/f-o-a-m/public-research

Product white paper: https://foam.space/publicAssets/FOAM_Whitepaper.pdf

I learned a lot from this, including that the Eiffel tower was scheduled for destruction. Nowadays, you can't legally take a picture of it!

You can take a picture of the tower but not its lights. http://c4sif.org/2010/12/copyright-and-the-eiffel-tower/

This is very cool. I have been thinking a lot about ideas in this space for a few years.