Return to the full study

KEDDEH SYSTEMS · SPATIAL-RADIX · R1 · 18 October 2026

Spatial Radix Mapping: source-defined construction, finite tests, and engineering obligations

Study reference: K-SPATIAL-RADIX-R1
Edition: 1, 18 October 2026
Scope: A source-faithful topic study accompanying the Keddeh Systems research article. The displayed construction is examined in its own terms, with executable finite examples and separately identified engineering questions.

Reader-projection note: the private Library storage identifier in reference S1 is replaced by its logical source label. All article prose, equations, source hashes and claims remain complete. The exact earlier article remains retained; this reading projection has a different byte identity.

Abstract

The supplied conversation develops an array from one middle row and a copy of the second row placed above and below it. It requires a single central 1, an outward reading from that center, a relationship between numerical designation and physical presence, and a description of mirrored (-1+) wholes under diagonal division. A later author statement extends the construction through 9, with another array at whole 2, one step up, for chip truth tables. These are the organizing instructions for this study. They deserve to be understood before substitutions such as a height map, repeated-origin tiling, or an assumed conventional carry are introduced.

Exact analysis of the supplied fifteen cells establishes axial reflection, half-turn symmetry, a unique central label, and particular diagonal balance counts. It also clarifies what the picture leaves open: its full extension law, the physical meaning of presence, the native join, and the operational transition at whole 2. The earlier AI exposition repeatedly moved from a diagram to claims about speed, cryptographic authentication, and unlimited scaling. Those conclusions require additional definitions and evidence. The accompanying deterministic tests make the distinctions inspectable without replacing the construction with an invented native theory.

1. Read the construction before changing its meaning

Intended rule

The opening correction says to copy row 2 above and below row 1. The retained diagram is:

4  3  2  3  4
3  2  1  2  3
4  3  2  3  4

This is three rows by five columns. The author's later corrections require each number to be considered with its physical presence, the one to be centered within the box outward, and the one not to be repeated. The diagonal description retains the literal (-1+) whole notation. [S1: lines 2–9, 23, 37, 51, 77]

Mistaken interpretation

The earlier responses alternated between a valley, a pyramid, a mass distribution, an energy source, and a force-balancing structure. Those are different mappings from symbols to physical quantities. None was selected or calibrated by the author's statement that numbers have physical presence. A numerical minimum also does not force the intended origin to be a physical depression. The author expressly corrected that inversion. [S1: lines 25–49]

The response to “youre repeating the 1” then claimed that the middle row had repeated the one. The displayed middle row already contains it exactly once. The correction should therefore govern subsequent construction and repetition, rather than be used to rewrite what the row actually shows. [S1: lines 51–68]

Corrected reasoning

Preserve the construction first: a single displayed central reference, its surrounding labels, their relations, and the stated outward interpretation. A separate physical map must specify whether presence means occupied extent, cell position, volume, material state, or another authored quantity. Until then, sums of labels are sums of labels; they are not measured mass, stored energy, force, or mechanical stability.

The label CENTRAL1 used in this study refers to that supplied central 1. It is not silently substituted for the separate opaque token .1, nor is (-1+) normalized into subtraction, a negative integer, or a pair of conventional signs. Existing geometry documentation likewise distinguishes a central identity labelled 1 from adding one to every numerical coordinate. Existing BOX method control specifically cautions against reconstructing the native object from generic nesting. [S2; S3; S7; S8]

Worked example

The fifteen labels have counts 1:1, 2:4, 3:6, 4:4, and numerical sum 43. Every one of the 22 horizontal or vertical neighbor pairs inside this crop differs in magnitude by 1. These are reproducible statements about the displayed labels. They neither assign a unit to physical presence nor exhaust the author's intended geometry.

*Contextual facility: inspect the original cells and reset to the source fixture.*

2. Center-out reading is a relation; a generating law must be specified

Intended rule

The author asks for outward development from the single center, later extending the construction through 9. The present crop is an example to be retained when investigating that extension. [S1: lines 37, 51, 230]

Mistaken interpretation

The prior explanation states a vertical path 1 → 2 → 4 and a diagonal path 1 → 3 → 4. The three-row array has only one outward vertical cell on either side of the center. A unit diagonal step reaches a 3, after which a further unit diagonal step leaves the displayed crop. A route from the center through that 3 to a corner 4 turns; it is not a straight diagonal. [S1: lines 69–73]

Another response supplies 1 + max(|x|,|y|) + Φ(x,y) but never defines Φ. Naming a correction term does not specify a computable field or prove that it generates the diagram. [S1: lines 296–304]

Corrected reasoning

For a transparent analysis adapter only, assign equal-spaced cell centers coordinates x = column − 2, y = 1 − row, using zero-based row and column indices. These conventional coordinates are not the native origin notation. On the fifteen supplied positions, the candidate rule

f(x,y) = 1 + |x| + |y|

fits every label. This is a useful finite observation, not a claim that the author has defined Manhattan distance as the complete native law. The uncorrected Chebyshev expression 1 + max(|x|,|y|) disagrees with eight cells. Setting Φ to min(|x|,|y|) would make that expression equal f, but that choice would be an analyst's completion of the missing term.

The sample does not uniquely determine its continuation. For example,

g(x,y) = f(x,y) + max(0, |x| − 2)

also fits all fifteen cells and preserves axial symmetry and a unique minimum on the integer plane, but f(3,0)=4 while g(3,0)=5. More properties might exclude g; the finite diagram by itself does not.

Worked example

The verified outward horizontal path is 1 → 2 → 3. The verified outward vertical path is 1 → 2. The verified one-step lattice-diagonal path is 1 → 3. The line joining opposite corners passes through cell centers labelled 4, 1, 4, with gaps between those sampled centers. Naming the path makes each observation precise.

*Contextual facility: compare candidate rules on the fixed sample, then inspect a labelled outside-sample disagreement. No candidate becomes the default native expansion.*

3. What the diagonal 50/50 statement establishes

Intended rule

The author's diagonal statement concerns mirrored (-1+) wholes and conformance between symbolic designation and physical presence. Keep that relational statement intact while identifying the exact partition used by a numerical test. [S1: line 77]

Mistaken interpretation

The earlier exposition translated this into conventional positive and negative vectors, zero net gradient, physical force cancellation, structural stability, and a universal diagonal-reflection theorem. Those are additional statements. Their operators, units, boundary terms, and interpretation were not supplied. [S1: lines 80–100, 311–318]

Corrected reasoning

Under the equal-spaced analysis coordinates, the geometric corner-to-corner lines are 2y=x and 2y=−x. For either line, six cell centers lie on each open side and three lie on the line. The open-side sums are 17 and 17; the on-line sum is 9. Pairing (x,y) with (−x,−y) proves this balance: the half-turn preserves the label and exchanges the two open sides.

A rule for the diagonal cells is still needed if “half” is meant to include all fifteen cells. Giving half of the on-line weight to each side yields 21.5 on each side. This is a shared-boundary accounting convention, not fifteen indivisible cells partitioned into two integer-sized equal sets.

For this equally spaced non-square rectangle, geometric reflection in a corner-to-corner diagonal is not itself a grid symmetry. Reflection of (2,−1) in y=x/2 is (2/5,11/5), outside the grid. The verified symmetry is half-turn pairing, together with horizontal and vertical reflection. This distinction does not settle whether the author intends a different physical embedding or a native whole-level mirror operation; that operation must be recovered rather than assumed to be Euclidean reflection of this adapter.

Worked example

Both diagonals produce the same three-part signature: 6 cells / sum 17; 3 cells / sum 9; 6 cells / sum 17. The facility shows the on-line cells separately so that no center or boundary contribution disappears during the demonstration.

*Contextual facility: select a diagonal and inspect the partition, with the coordinate and boundary conventions visible.*

4. Consistency checks, content integrity, and actor authentication

Intended rule

The source suggests using structural relations as inspectable consistency conditions. That is a meaningful direction for a constrained data representation: an invalid change can violate a declared relation. The stronger self-authentication claims appear in the supplied AI exposition. [S1: lines 115–118, 316–318, 377–380]

Mistaken interpretation

Symmetry was described as a cryptographic hash that detects any corruption or malicious change instantly and removes validation work. A public balance condition alone does not identify a unique message, establish who produced it, or specify a detector's physical response time.

Corrected reasoning

Ask exactly what a check accepts. The following changed grid retains the original histogram, sum, unique central 1, horizontal and vertical reflection, half-turn symmetry, and both diagonal signatures:

4  3  3  3  4
2  2  1  2  2
4  3  3  3  4

Four positions have changed. Therefore those particular predicates are non-injective: more than one content state passes them. This is a concrete limit of those predicates, not a refutation of the construction or of separate KEX integrity, identity, provenance, or authorization mechanisms.

A validator comparing all fifteen cells with the source rejects the changed grid. So does the candidate full-field f rule. These add constraints beyond aggregate balance. Even a complete public admissibility rule establishes only that content is an allowed state; origin authentication additionally needs a binding between an authorized actor and the relevant content, context, and event. NIST's MAC definitions make this distinction concrete through keyed authentication mechanisms. An ordinary unkeyed hash comparison also requires a trustworthy reference or authenticated binding to establish origin. [T1; T2]

The useful research question is therefore which classes of accidental and coordinated changes the native structural checks detect, what information they preserve, and how they combine with KEX's actual authorization and retained-identity mechanisms. No conclusion about those other mechanisms follows from this small counterexample.

Worked example

Run the original and changed fixtures through identical checks. Both balance signatures pass; exact-source comparison fails for the changed fixture. Showing both outcomes avoids either overclaiming the balance test or discarding its legitimate diagnostic use.

*Contextual facility: compare the fixed four-cell change, its unchanged summary statistics, and the full-field rejection.*

5. Joining boundaries while preserving the intended origin and identities

Intended rule

The source motivates spatially connected arrays and an outward construction with an unrepeated center. These requirements make the join operation important: a native continuation and a repetition of independent local arrays need not be the same operation.

Mistaken interpretation

The earlier responses infer universal zero-distortion tiling and infinite computing power from matching edge labels. They also describe perimeter values as if all boundary cells were 4. The actual left and right edges are [4,3,4], and the top and bottom edges are [4,3,2,3,4]. [S1: lines 34, 123, 171, 189, 280, 390–392]

Corrected reasoning

Opposite edges do match, which supports a well-defined label-compatibility check. What happens next depends on the join:

1. Repeat full rectangles without overlapping cells. Across the horizontal seam, [4,3,4] meets [4,3,4]. Those three new neighbor differences are 0, whereas all internal orthogonal differences were 1. Along the center row near the seam the sequence includes 2,3,3,2.
2. Identify matching boundary columns or rows. A horizontal join can instead share the column [4,3,4]. The center-row sequence then contains 2,3,2. Values agree at the shared positions, but two source identities occupy the same joined coordinate until an explicit aliasing or ownership rule resolves them.

Both are analyst-defined test joins, not recovered native operators. Neither proves that a stencil, gradient, physical flux, routing transition, or identity mapping is preserved. For example, the second difference at the center-row right boundary is −1 in the first join and −2 in the shared-boundary join; the single-origin f continuation gives 0 there. The comparison establishes operator dependence, not that one join is universally incorrect.

A two-by-two repetition contains four local centers. Namespacing their identities can distinguish them, but cannot make four occurrences of label 1 into a single global occurrence. The author's instruction against repeating the one must be scoped before local-copy tiling can be treated as native whole growth. This is precisely why a hierarchy contract is needed.

Worked example

Four nonoverlapping copies contain 60 distinct source-cell identities and 60 coordinate positions. Shared-boundary copies retain 60 source identities but occupy 45 positions, including 13 multiply represented positions. All shared values agree. Value agreement and identity resolution are two separate checks.

Unlimited repetition can be a mathematical construction under a declared rule; physically instantiated arrays still require resources, communication, synchronization, and error handling. Their scaling behavior must be measured for the specified workload.

*Contextual facility: choose the two test joins, inspect seam differences and identity multiplicities, and keep the native-join question visible.*

6. Through 9, then another array at whole 2

Intended rule

The author's precise statement is: “this goes ALL the way out to 9 AND when it passes 9 you simply add another array at whole 2 one step up... its designed for chip truth tables all the way out”. [S1: line 230]

This supplies a boundary and a next-whole intent. It does not justify replacing the construction with a flat conventional numeral encoding.

Mistaken interpretation

The earlier AI introduces a “base-10” or “radix-9” label, calls the first array saturated, makes the entire prior array a local 1, and draws a next array containing several zeros and repeated ones. That drawing conflicts with the stated preservation requirements and cannot serve as a source-faithful transition oracle. The prior-whole-as-local-1 reading appears in the AI exposition, not in the quoted author sentence. It remains a candidate until an authored mapping confirms its exact meaning. [S1: lines 232–266, 320–325, 364–366]

A cap at 9 does not alone identify the radix, digit alphabet, value decoding, or carry arithmetic. Labels could designate positions, extents, states, or whole relations. An ordinary decimal interpretation would add a particular ten-symbol digit and place-value contract; it must not be imported merely because the largest displayed single-symbol label is 9.

Corrected reasoning

The missing executable specification can be named precisely:

1. State domain and label semantics: what a cell and its label represent; which states are admissible.
2. Extension geometry through 9: how the present crop expands, with its full boundary and center uniqueness preserved.
3. Input operation and operand encoding: which truth-table operation is applied to which inputs.
4. Overflow measure and trigger: what quantity passes 9, and at what transition or observation point.
5. Whole-level and step-up transform: whether “up” changes hierarchy, frame, dimension, placement, or another native relation.
6. Post-overflow state and residual: what is retained, consumed, reset, created, or carried.
7. Native identity embedding and inverse: how prior whole and cell identities remain addressable, without an unauthorized duplication or collapse.
8. Truth-table routing and output decoding: how each admissible input reaches an identifiable output and how the result is read.
9. Boundary join and ownership: how adjacent or shared positions, connections, and records are assigned.
10. Reset, failure, concurrency, and cascade policy: what happens for simultaneous triggers, unavailable capacity, interruption, repeated promotion, and recovery.

A non-destructive envelope can illustrate how one record retains another by identity, but it is not the native BOX operation. Existing source method control explicitly prevents that substitution. The bounded search located typed PlaceValue constructors, existing geometric adapters, and other register/overflow studies; none supplied the exact new spatial-radix transition. This is a scoped recovery result, not a claim that the author's developed definitions do not exist elsewhere. [S3–S6]

Worked example

The test harness can classify external test labels 9 and 10 as not above / above the stated cap. It returns no invented resulting array or numeric answer. At 10 the truthful result is “native promotion required but unspecified.” The interactive facility exposes the missing-operation list and the source sentence, providing a concrete place to attach the native rule when recovered.

*Contextual facility: inspect the boundary contract and candidate retained-identity envelope separately. Native promotion remains pending an authored operator and fixture.*

7. Chip truth tables: what a meaningful demonstration would show

Intended rule

The author explicitly identifies chip truth tables as the target. That invites a serious evaluation of whether source-defined state and spatial relations can be embodied as an efficient, inspectable computing structure. It does not require treating the proposal as only an architectural metaphor.

Mistaken interpretation

The supplied exposition describes ordinary arrays as necessarily processed in linear order, all carry as slow ripple carry, and routing as a way to make the correct answer appear without logical operations or delay. It then promotes a geometric intention into near-light computation and infinite scalability. [S1: lines 111–113, 269–282, 327–331, 372–402]

Corrected reasoning

Layout and traversal. Row-major and column-major are storage/addressing schemes, not requirements to visit every cell in sequence. An ordinary row-major address can be computed directly as base + element_size × (row × width + column). NumPy's primary documentation distinguishes array indexing from memory strides and accommodates multiple strided layouts. Center-out access can be expressed on conventional storage; a spatial design must show its benefit against a specified baseline and workload. [T3]

Routing and computation. A truth table must identify inputs, outputs, and a function. Mapping an input to a destination still requires physical encoding, distinguishable state-dependent behavior, and a readout. Routing may be the embodiment of the computation, but a labelled static grid alone does not establish that it performs the intended function. Existing FPGA documentation describes logic resources and configurable routing together, including dedicated arithmetic/carry structures. Specialized carry-select and interconnect designs already make the engineering comparison more precise than “ordinary processors only ripple.” [T4; T5]

Alternative physical models. Conventional Boolean-gate implementation is not the only conceivable model. Fredkin and Toffoli's conservative-logic work is a primary example of explicitly defined physical/dynamical models connected to computation. Its relevance is methodological: an alternative needs primitives, composition, and a correctness account. It does not validate this grid by analogy. [T6]

Timing. Replacing one implementation mechanism does not remove the time for propagation, state resolution, sensing, and communication. A routing path has a physical length and medium-dependent dynamics; a labelled diagram supplies neither. The SI constant c describes light in vacuum, not a measured chip's end-to-end computation latency. A timing claim needs a target substrate, path lengths, device/interconnect model, input-output measurement protocol, and uncertainty. FPGA timing documentation explicitly treats routing and connected-block delays as parts of the analyzed design. [T7; T8]

Worked evaluation plan

Freeze one authored operation, its full input domain, state transition, hierarchy rule, and output decoder. For a small domain, enumerate every input and compare the outputs with the authored truth table. Include boundary and overflow cases, reset, ambiguous inputs, and coordinated changes. Then separately build or simulate the physical implementation, test noise and timing corners, measure end-to-end latency and energy, and compare with an appropriate implementation of the same function. A software demonstration, HDL simulation, placed-and-routed estimate, and physical device measurement must each be labelled as that kind of evidence.

These steps can establish genuine technical merit without requiring either blanket endorsement or blanket rejection from a small picture.

8. Research conclusion and reproducibility

The construction supplies a clear center-relative organization and a specific invitation to investigate how symbol, presence, whole, and execution relate. Several early misconceptions are resolved simply by obeying the source: retain the single center, do not invert it into an unrequested height map, do not invent paths outside the supplied crop, preserve (-1+), and do not insert zeros or repeated origins into the next-whole diagram.

The finite study establishes the shown symmetries and explicit diagonal counts. It also identifies which stronger statements require additional structure: admissible-state validation beyond a public balance signature, actor authentication beyond content consistency, seam operators beyond equal labels, and a native whole transition beyond the threshold phrase. The most useful next contribution is the authored through-9 construction and a worked whole-2 truth-table transition with retained identities. That would let the interactive study test the intended mechanism directly.

The accompanying Python standard-library harness reproduces 33 named checks and writes a machine-readable report. Some checks confirm source properties; others successfully reproduce a counterexample or an unresolved-contract status. “All checks passed” means that the finite analysis is reproducible, not that all supplied engineering claims were validated. No browser, application, chip, or physical device was executed as part of these checks.

Source and reference register

Supplied and recovered project sources

S1. *Pasted text(20261018-081256).txt*, supplied conversation, 403 lines, 41,939 bytes. Retained Library item retained-source:S1; SHA-256 24e99bf74ad6ff49c3598acb83e0c4da7c312386a11ad763091a2dd4a185200d. Short author corrections and longer prior AI responses are distinguished throughout. The file's final request for a discussion without assertions is at line 349; its following response continues to make unsupported proof claims.

S2. Keddeh geometry README, *Mesh cinematic*, paragraphs on the central reference and publisher integration. Recovered repository: docs/geometry/README.md. The central identity and presentation adapter are distinguished from physical evolution laws.

S3. *Keddeh Maths — Learning Method Control*, 2 October 2026, “Box-in-box is not to be reconstructed from analogy,” recovered lines 10–12. This is source-defined method control, not a refusal to analyze the BOX object.

S4. *KEX-BRAINK Mathematical Foundations and Proof Obligations*, typed state algebra and PlaceValue(Symbol, Radix, Position), recovered lines 47 and 347. This establishes a typed place-value constructor, not the new spatial-radix carry.

S5. *3D rendering, CENTRAL 1 and tensor-shell implementation trace*, 18 October 2026, sections 1–4. Positive evidence of existing geometry/rendering, with edition-specific origins and a scoped unresolved native BOX-to-shell linkage.

S6. Existing origin study, register/overflow discussion, recovered source/osw/public/origin-v2-9.txt, lines 982–1034. These are other source-specific hardware checks; they are not substituted for the new whole-2 rule.

S7. Opaque singular-origin implementation, recovered source/osw/public/singular-origin.js, lines 2–5. The literal .1 is expressly a string identity; numeric projection coordinates are separately typed.

S8. Lossless whole-origin classification, recovered source/kex-origin.mjs, lines 2–3. (-1+) is retained as a distinct source lexeme with arithmetic normalization disabled.

The full bounded search manifest is supplied with the study. It scanned 153 eligible text/source files in the identified recovered-source areas; its exclusions and exact search terms are recorded. No universal absence claim follows from that search.

Primary technical references

T1. NIST, Message Authentication Code glossary. Used for the distinction between data consistency and keyed origin/integrity authentication.

T2. NIST, FIPS 198-1: The Keyed-Hash Message Authentication Code. Historical primary mechanism reference, not an assertion about its current regulatory status.

T3. NumPy, The N-dimensional array: indexing and internal memory layout. Used for indexing/stride distinctions, not a benchmark of the proposed structure.

T4. Intel, FPGA Architecture Overview. Logic resources and configurable routing are components of the implemented circuit.

T5. Altera, Stratix 10 Carry Chain. Dedicated adders, carry-select structure, and interconnect placement illustrate an actual arithmetic implementation.

T6. Edward Fredkin and Tommaso Toffoli, Conservative Logic, MIT-LCS-TM-197, May 1981; journal version, *International Journal of Theoretical Physics* 21 (1982), 219–253, DOI 10.1007/BF01857727.

T7. Intel, Timing Path and Clock Analysis. Search-index primary documentation excerpt checked; direct page retrieval failed during this review. Used only for the stated treatment of block and routing delays, not for numerical performance claims.

T8. BIPM, SI base unit: metre. Defines c as the speed of light in vacuum; supplies no chip timing result.