Stop 4 of 9 · How many bits fit on the boundary, and what does full look like?
Bits on the boundary — and what full looks like
Every square on the walls of a region is one bit: the step across it is open or closed. No way of filling the region holds more than the walls allow, S = 2·dy·dz·log 2 on T³. It is reached in exactly one way: when every pattern of the walls is equally likely. Raise the chance of each bit and watch the count reach the bound at one half.
The walls of a region
The region is the slab on the left of the plane. On T³ every axis closes on itself, so the slab has two walls:
the plane you choose, and the seam where the axis closes. Each wall has dy × dz squares.
Orange: open (1). Empty: closed (0). Move the depth and the plane: the count of bits does not change. Drag to rotate.
With each square of area 4ℓ_P² log 2 (one bit of area, as on the open cube) the capacity is the Bekenstein–Hawking value
A / 4ℓ_P² of the picture "bits on the boundary": area, not volume.
Fill it up
Now let every bit be 1 with the same chance. With a small chance the walls are almost all empty and hold little information; the information held grows as the chance goes to one half, and comes back down after. The bar reaches 100 % only at one half: every pattern equally likely. That is the only state that fills the region (Lean: the uniform distribution and no other).
The region turns dark only when it is full: that is Nava's postulate, declared in Lean as a bridge (HSaturationDark), not derived.
Below the top the postulate says nothing, so the page does not shade the region in between.
What is exact here
- Lean Every distribution of the walls has
H ≤ S(the Bekenstein bound); the uniform one reaches it. - Lean
H = Sonly for the uniform distribution: full means perfectly mixed. - Lean Each configuration of the walls is a different field operator.
- Lean Each bit of the walls is a yes/no question, and all of them can be asked together: they commute, so the walls obey ordinary probability (Kolmogorov). A single bit varies at most
¼, a fair coin. Questions that do not commute, like position and transport, obey Robertson instead. - Declared bridge A full region is dark: a postulate declared in Lean, with its consequences proved and its consistency shown.
- Browser The page draws independent bits with one common chance; at one half that is the uniform distribution.
Theorem trail
| Statement | Name |
|---|---|
H ≤ S; the uniform distribution reaches it | TorusBekenstein.entropy_le_cutEntropy, TorusBekenstein.entropy_uniform, TorusBekenstein.exists_saturating |
H = S iff uniform | TorusSaturation.saturated_iff_uniform |
| the bits are field operators | TorusBoundaryField.configField_injective, TorusBoundaryField.card_fieldConfigs |
| full ⇒ dark (declared) | TorusSaturation.HSaturationDark, TorusSaturation.hSaturationDark_satisfiable |
| the wall bits are commuting events | EventsTopology.prob_openEvent, EventsTopology.openEvent_comm |
a bit varies at most ¼; commuting bits have a joint law | QuantumEvents.variance_le, QuantumEvents.joint_sum |
| non-commuting questions: Robertson for bits | QuantumEvents.robertson_events |