# Public vs Private Networks

## Contents

## Public Networks Are Strictly Dominant Strategies in Infinitely Long Games

This is a proof demonstrating that there are no private/restricted networks that can't be made a strict topological subset of a public/decentralized one. Such that, when a network is defined as a set, S_{i}, in a shared topological space on and "above" a spheroid (or in other words, bounded by a finite, earth shaped, 3-dimensional hole) with restricted networks denoted, S_{p}, and unrestricted networks denoted, S_{d}, there are no infinitely long games in which S_{p} ⊂ S_{d} does not occur and remain in perpetuity. Making S_{d} a strictly dominant strategy in infinitely long games.

### Caveat

This isn't to say that private Grassland networks are bad, necessarily. But as we'll show, they only make sense in circumstances where S_{i} contains its limit points, like inside a building with walls on all sides and a roof. But if S_{i} does not contain its limit points, there are no feasible economies of scale for an S_{i} upon and above the surface of a spheroid (e.g. the earth).

As well, private networks should only be used on a limited scale since the act of creating a large, "dark" Grassland "stack", a set of pointers to 4D (space+time) coordinates on a "heap" used by a separate Grassland stack, would cause the sequestering of information from those parties, lead to uncoordinated multithreading causing race conditions and remove the mutual knowledge required to maintain peaceful Nash Equilibrium. And could lead to unfavourable scenarios where other parties have no choice but to deallocate offending resources on the heap through some form of externalized garbage collection.

### Proof As Embeddings In Topological Spaces:

It is sensible to define Grassland networks as sets in a topological space (Topological space). That is to say, a set of points along with a set of neighbourhoods (Topological Neighbourhood) for each point.

1) We'll define each network as a set, S, in a shared, 3-dimensional topological space that's only bounded in the same finite region by an earth/spheroid shaped, 3-dimensional hole.

2) The spheroid is opaque and it is not possible in any way to place cameras at its center pointing outwards to a surface (precluding any economies of scale).

3) All the points of S are made up of 1) each camera's raycasted principal point and 2) the neighbourhood of raycasted points around it within the frame's field-of-view. The raycasted principal point (RPP) being the initial point at which an imaginary line (the "ray" in "raycasted") drawn outwards from the center (principal point) of the camera's aperture intersects the external, real-world object (an unprojection from the camera to the world).

4) We'll denote any private (restricted) network as S_{p}, and any decentralized (public) network as S_{d}.

5) And we'll define the "domination" of one set, S, by another set, S', to be whenever the points in S can be approximated (substituted) to a higher degree by points in S' than the other way around. To put it another way, the probability that any given x ∈ S is arbitrarily close to some y ∈ S' is higher than the probability that any given y ∈ S' is arbitrarily close to some x ∈ S. Therefore, the number of points in S' is greater than the number of points in S (even if it's just one more camera/node)

Since both S_{p} and S_{d} are on the surface of the earth, a spheroid having no edges, no set, S, can contain ALL its limit points (Limit Point) to the effect of keeping points of another set, S', from approximating points of S. And it goes without saying, that this is true along the vertical axis. So even continent's aren't topologically compact. They don't contain their limit points in regards to camera placement since along vertical axis the "sky's the limit". Which means there's always some way for points in an unrestricted Grassland network, S_{d}, to approximate more points in S_{p}. Topologically speaking, a greater number of points in S_{p} can be shown to be merely a point in a neighbourhood of some point in S_{d}.

### Proof As Embeddings In Infinite Sets:

We can also demonstrate this by finding an embedding of S into the infinite sets. And showing any differences in cardinality between them. In the sense that some infinities are larger than others.

The points of a private Grassland network would find an embedding in the set of all Integers. It can theoretically stretch infinitely in both directions, [-infinity, ..., -3, -2, -1, 0, 1, 2, 3, ..., infinity], with an infinite amount of points from a countably infinite amount of cameras. But it's still bound to certain values; S_{p} has more restrictions than S_{d} by definition. And there exists points in S_{d} which fall "outside" and can't be defined using the operation(s) that govern the elements of S_{p}. Since the domain and range of this operation(s) are, *by design*, restricted to the elements of S_{p}.

Therefore, because a decentralized (public) Grassland network doesn't have those restrictions it finds a complementary embedding into the Reals, [-infinity, ..., -Pi, ..., -1.001, ..., -0.001, ..., 0, ..., 0.001, ..., 1.001, ..., Pi, ..., infinity]. It's an infinity too but it's a much larger, uncountably infinite, infinity because it can theoretically stretch to infinity even between the Integers.

And it therefore approximates the Integers. S_{p}'s point at "1" can be approximated by an infinite number of points in S_{d}. Moreover, "1" in S_{p} can be approximated to an arbitrary degree of accuracy by points in S_{d} from both sides, "0.9999...9" and from "1.0000...1". This is because y ∈ S_{d} has the freedom to approach any x ∈ S_{p} along every dimension of the space they're embedded into.

### Conclusion:

So since more points of S_{p} can be approximated to an arbitrary degree of accuracy by points in S_{d} than the other way around, and since S_{d} therefore contains the limit points of S_{p}, we can say that the set, S_{pn}, comprised of the neighbourhoods of all points in S_{p} is merely a smaller subset of the set, S_{dn}, a set comprised of the neighbourhoods of all points in S_{d}. And therefore S_{p} is dominated by S_{d} according to point 5 above.