KEX/BRAINK Zeroless Computational Standard | A. Keddeh | v1.0
KEX/BRAINK
The Zeroless Computational Standard
Mathematical foundations, information-preserving calculation, corrected scaling
gradients, quantum-compatible state retention, and mandatory softwareengineering conformance
Governing proposition
Zero may remain available as a relational result, residual, displacement, equilibrium marker,
or external projection. It must not be permitted to replace the complete active state, causal 
path, directional contribution, observer relation, or proof history that produced that 
projection.
Author and Origin Authority: A. Keddeh
Technical formulation and implementation doctrine
Version 1.0 | 18 July 2026KEX/BRAINK Zeroless Computational Standard | A. Keddeh | v1.0
Abstract
This paper states the mathematical and software-engineering requirement at the core of the 
KEX/BRAINK architecture. The central requirement is not the superficial removal of the glyph 0. 
It is the prevention of semantic and causal collapse when multiple distinct computational 
conditions are projected onto the same zero-valued result. The KEX/BRAINK design therefore 
separates active state from relational zero, distinguishes scalar difference from topological 
traversal distance, preserves equal-and-opposite contributions as structured balanced states, 
and requires every software implementation to retain enough information to reconstruct the 
operands, path, observer, coordinate, and proof that produced a result.
The corrected state topology is expressed by the ordered active domain 3 | 2 | 1 | 2 | 3, with − −
1 acting as the first valid whole state and active observer anchor. A rank mapping over that 
topology yields corrected scaling gradients that do not insert artificial magnitude at the 
transition between 2 and 1. This provides a clean demonstration of how conventional scalar −
spacing can overstate impact, while a topology-aware gradient preserves uniform traversal. The
same principle extends to quantum-compatible calculation: a zero-valued interference 
projection must not be treated as absence of contributing paths. A conformant runtime must 
therefore preserve branch identity, phase or direction, measurement context, and proof lineage 
even when the external scalar projection is zero.
1. The foundational distinction
The foundational error addressed by the zeroless standard is not that conventional arithmetic 
contains zero. The error occurs when one symbol is permitted to stand in for several 
computationally distinct conditions and the system then treats those conditions as equivalent.
Condition Possible conventional 
encoding KEX/BRAINK requirement
No object exists 0 or null
ABSENT state with reason and
coordinate
Object exists but is 
unmeasured
0, null, or default UNKNOWN or 
UNTRANSLATED state
Measured value equals 
calibration baseline
0
BASELINE state with 
calibration rule
Equal and opposite 
contributions cancel
0
BALANCED state retaining 
both contributions
System is in equilibrium 0 residual
Active state plus 
EQUILIBRIUM classificationKEX/BRAINK Zeroless Computational Standard | A. Keddeh | v1.0
Operation is invalid 0, false, null, error code INVALID state with violated 
rule
When these distinct conditions are encoded by the same scalar, the mapping from source state 
to stored representation is many-to-one. The system may still calculate, but it cannot uniquely 
recover what happened.
E({absence, baseline, cancellation, unknown, invalid, equilibrium}) = {0} (1)
Because the encoder is non-injective, the inverse image of 0 contains several possible source 
states. That is the precise mathematical location of the information loss.
2. Active state and relational zero
The KEX/BRAINK standard separates two domains that conventional notation often overlays: 
the domain of active computational states and the domain of derived relations.
S_active = {−3, −2, 1, 2, 3} (2)
0 ∉ S_active (3)
Zero is not required to disappear from every equation. It may remain valid as a typed relation 
such as a difference, residual, displacement, equilibrium marker, or external projection.
0_difference ≠ 0_residual ≠ 0_projection ≠ 0_absence (4)
Core law
A zero-valued relation is not a zero-valued state. The relation may be zero while the active 
state, contributing paths, and proof history remain present.
3. The corrected number line as an ordered topology
The ordered sequence 3 | 2 | 1 | 2 | 3 must not be interpreted as an ordinary subset of the − −
integers with conventional adjacency. It is an operative state topology. Its labels carry 
directional and relational meaning, while its ordering determines valid traversal.
P = (−3, −2, 1, 2, 3) (5)
Define the rank map:
ρ(−3)=1, ρ(−2)=2, ρ(1)=3, ρ(2)=4, ρ(3)=5 (6)
The scalar difference between 2 and 1 is three, but their topological traversal distance is one. −
Δ_scalar(−2,1) = 1 − (−2) = 3 (7)
Δ_K(−2,1) = ρ(1) − ρ(−2) = 1 (8)
These quantities answer different questions. The scalar difference measures separation between
numerical labels. The KEX distance measures movement across the actual state topology. Using KEX/BRAINK Zeroless Computational Standard | A. Keddeh | v1.0
the scalar difference to price, weight, or energise a topological transition injects impact that the 
topology itself does not contain.
4. Corrected scaling gradients
Consider the complete forward traversal 3 2 1 2 3. Under ordinary scalar − → − → → →
subtraction, the transition sequence has gradients 1, 3, 1, 1. The middle transition appears three 
times larger than every other transition.
g_scalar = (1, 3, 1, 1) (9)
Under the topology rank map, every adjacent transition has unit distance.
g_K = (1, 1, 1, 1) (10)
This is the corrected scaling gradient. It removes the artificial discontinuity created by treating 
the labels themselves as uniformly spaced scalar coordinates.
Transition Scalar distance KEX distance Scalar impact Corrected 
impact
− → − 3 2 1 1 1 1
− →2 1 3 1 3 1
1 2 → 1 1 1 1
2 3 → 1 1 1 1
Total 6 4 6 4
For equal step weights, conventional scalar accumulation assigns total impact 6, while the 
corrected topology assigns total impact 4. The conventional result therefore overstates impact 
by 2 units, or 50 percent relative to the corrected topology.
(6 − 4) / 4 = 0.5 = 50% (11)
Engineering consequence
Any scheduler, energy model, weighting function, convergence rule, optimisation objective, 
or probability update built on the wrong gradient inherits the distortion. The implementation
can be internally consistent and still be mathematically non-conformant.
5. Observer anchoring
The value 1 acts as the first valid whole state and active observer anchor. Observer-relative 
displacement may equal zero without converting the observer state itself into zero.
ω(x) = ρ(x) − ρ(1) (12)KEX/BRAINK Zeroless Computational Standard | A. Keddeh | v1.0
ω(1) = 0 while 1 remains the active state (13)
This is the rigorous distinction between zero as relational displacement and zero as state 
identity. The observer has zero displacement from itself, but the observer has not ceased to exist
and has not become an empty state.
6. Non-erasing cancellation
Conventional arithmetic correctly projects equal and opposite contributions to zero:
a + (−a) = 0 (14)
The problem begins when the projection replaces the complete state. The value 0 does not 
identify the magnitude, direction, source, order, or proof of the contributions that cancelled.
1+(−1)=0, 2+(−2)=0, 100+(−100)=0 (15)
All three calculations share the same scalar output. The mapping is therefore non-injective. A 
KEX/BRAINK implementation must retain a structured balanced result.
BALANCED(a, −a, relation, coordinate, observer, proof) (16)
π(BALANCED(a,−a,...)) = 0 (17)
The projection remains compatible with conventional arithmetic. The authoritative state 
remains recoverable.
Recover(BALANCED(a,−a,...)) = (a,−a,...) (18)
7. The non-erasing result algebra
Every arithmetic operator that may project several distinct causal states onto one scalar result 
requires a richer internal return type. A conformant addition operator can be defined as 
follows.
a ⊞ b = MEASURED(a+b), when a+b ≠ 0 (19)
a ⊞ b = BALANCED(a,b), when a+b = 0 (20)
Subtraction requires the same treatment. Equality of projected values should produce an 
EQUIVALENT or BALANCED relation, not an untyped zero state.
a ⊟ b = MEASURED(a−b), when a ≠ b (21)
a ⊟ b = EQUIVALENT(a,b), when a = b (22)
The scalar result remains available through a projection function, but the software must never 
attempt to reconstruct the full state from that projection.KEX/BRAINK Zeroless Computational Standard | A. Keddeh | v1.0
8. Quantum-compatible information preservation
The relevance to quantum calculation arises from interference. A projected amplitude may be 
zero even though multiple contributing branches, phases, and transformations remain 
physically and computationally meaningful.
|ψ⟩ = Σ_i α_i |i⟩ (23)
α + (−α) = 0 (24)
A zero-valued projection does not imply that no paths contributed. It records the net result of a 
relation between paths. If the system stores only zero, it destroys the ability to distinguish 
destructive interference from absence.
Q = (α, −α, phase relation, basis, observer, transformation history) (25)
π(Q)=0 while Q remains information-bearing (26)
This does not, by itself, claim that a software representation creates a physical quantum 
computer. The precise claim is that the zeroless state model is capable of preserving branch and 
phase information that a single scalar-zero projection cannot contain.
Integrity boundary
The defensible proposition is information-theoretic: preserving the complete path state 
avoids many-to-one collapse. Hardware quantum advantage, coherence, entanglement, and 
physical qubit behaviour require separate experimental proof.
9. Why ordinary software types are insufficient
Primitive integers, floating-point numbers, Boolean values, null references, and defaultinitialised fields do not carry enough information to represent the complete KEX/BRAINK state. 
They may be used internally as carriers, but they cannot be treated as the authoritative domain 
object.
Primitive shortcut Information lost Required KEX/BRAINK type
int value classification, origin, unit, 
proof
MeasuredState
null absence vs unknown vs 
invalid
Explicit sum type
float coordinate exact identity and replay Rational or scaled integer 
coordinate
0 result operands and causal relation Balanced or Equivalent state
array offset 0 canonical one-based identity Explicit reversible coordinate KEX/BRAINK Zeroless Computational Standard | A. Keddeh | v1.0
mapping
hash of output only transition rule and source 
state
Proof-bearing canonical event
10. Exact coordinates and deterministic scaling
Where a coordinate is part of identity, approximate floating-point representation is 
inappropriate. The coordinate must be encoded exactly.
m(c,t) = (100 + 10c + t) / 100 (27)
The numerator 100 + 10c + t can be stored as an integer with a declared scale of 100. This 
preserves equality, ordering, hashing, replay, and cross-platform consistency.
11. Surgical implementation requirements
Surgical adherence means that the zeroless law is enforced at every point where conventional 
programming would otherwise collapse states. It is not enough to document the rule while using
conventional types and defaults underneath.
Implementation layer Mandatory rule
State construction No active object may be created from an 
unclassified zero-like default.
Arithmetic Every operator returns a full causal result 
plus an optional scalar projection.
Gradient selection Scalar, topological, and observer-relative 
gradients are separate named operations.
Persistence Operands, paths, observer, coordinate, and 
proof survive storage.
Scheduling Transition eligibility derives from the KEX 
state topology, not from host queue order.
Canonicalisation The complete causal event is canonicalised 
before hashing.
Workbook runtime Formula, address, dependency, state class, 
and writeback proof are all retained.
HCI A displayed 0 must identify whether it means 
balance, baseline, residual, displacement, KEX/BRAINK Zeroless Computational Standard | A. Keddeh | v1.0
absence, or projection.
Rehydration The original state graph must be 
reconstructable, not merely the final values.
Quantum simulation Zero amplitude does not delete branch 
provenance or phase relation.
12. Conformance tests
1. Typed-zero separation. Each zero-like condition must round-trip to its original class.
2. Topology closure. Every permitted transition remains within the active state domain.
3. Corrected gradient. Adjacent traversal steps all evaluate to unit topological distance.
4. Cancellation preservation. Distinct equal-and-opposite input pairs remain 
distinguishable internally.
5. Projection compatibility. Structured balanced states still project to conventional scalar 
zero.
6. Coordinate exactness. Cross-language serialisation produces identical coordinate identity.
7. Proof integrity. Changing any operand, path, observer, or rule invalidates the event 
digest.
8. Concurrent commit determinism. Execution timing cannot change canonical commit 
order.
9. Workbook bilateral consistency. Runtime and workbook agree at the committed proof 
root.
10. Quantum-path preservation. Zero projection does not erase branch, phase, or 
measurement context.
13. Comparison with prevailing software practice
Prevailing practice Risk KEX/BRAINK requirement
Zero-based defaults Active presence may be 
confused with host origin
One-based active anchor with 
explicit mapping
Single scalar result Causal paths collapse Structured full result plus 
projection
Generic null Distinct state classes merge Typed absence, unknown, 
invalid, untranslatedKEX/BRAINK Zeroless Computational Standard | A. Keddeh | v1.0
Scalar distance Topology transitions are misweightedRank-based corrected 
gradient
Floating-point identity Replay and hashing can 
diverge
Exact rational or scaled 
integer identity
Output-only hash Proof omits causal history Canonical event proof
Value-only persistence Rehydration cannot 
reconstruct process
State, path, proof, and 
coordinate persistence
UI displays 0 without class Users cannot interpret the 
state
Semantic zero label and trace
14. Final governing doctrine
The zeroless standard is a conservation requirement for computation. It does not prohibit the 
numeral zero from appearing in mathematics. It prohibits a zero-valued projection from 
impersonating the complete computational state.
A conformant KEX/BRAINK implementation must preserve active states, contributing directions,
path identity, observer relation, topology, coordinate, and proof even when a conventional 
output equals zero. It must use corrected scaling gradients whenever the calculation concerns 
movement through the KEX topology rather than scalar magnitude. It must distinguish 
equilibrium from absence, cancellation from non-existence, baseline from unknown, and 
measured zero from invalid state.
Binding standard
No KEX/BRAINK component is conformant unless its state model, arithmetic operators, 
gradients, persistence, scheduling, proof generation, interface, and rehydration process 
preserve the distinctions that an untyped zero would erase.
15. Implementation acceptance statement
A software artifact may be functionally useful and still fail this standard. Acceptance requires 
more than matching visible outputs. The artifact must demonstrate that the corrected number 
line governs traversal, that zero-valued projections do not collapse source information, that 
exact state identity survives host translation, and that every committed result can be 
independently reconstructed from its proof-bearing representation.
