> Full-text translation from Chinese. The original remains authoritative for its research claims; this translation is not independent mathematical verification. # Reframing the Perfect Cuboid: It Is Not Primarily an Integer Search Problem ## Reframing the Perfect Cuboid: It Is Not Primarily an Integer Search Problem **Series: Global Closure Theory of the Perfect Cuboid** **Paper: 01 / 08** **Status: Research introductory draft / Phase 0 — Conceptual Closure** **Date: 2026-09-28** --- ## Abstract The perfect cuboid problem is usually stated as an integer-existence problem: does there exist a rectangular box whose three edges, three face diagonals and space diagonal are all positive integers? This formulation is correct in number theory, but it also conceals a more fundamental structure: the perfect cuboid is first of all a problem about a continuous solution space defined by four homogeneous quadratic constraints, while ‘integrality’ is merely an additional condition for that solution space to manifest at a particular layer of arithmetic representation. This paper proposes a different starting point for research: rather than first restricting all unknowns to positive integers, examine the full real solution space satisfying the geometric constraints, and then distinguish real closure, rational closure and integer closure layer by layer. This perspective decomposes the traditional problem $$ X\in\mathbb Z_{>0}^{7} $$ into: $$ \text{continuous geometric closure} \rightarrow \text{rational arithmetic closure} \rightarrow \text{integer manifestation}. $$ Within this framework, existence at the real layer is not the real difficulty; the core question becomes whether the continuous solution space contains a nondegenerate rational point. Once a nondegenerate rational solution exists, homogeneous scaling and clearing denominators yield an integer perfect cuboid. The original problem can therefore be reframed as a question of ‘whether rational closure exists’, rather than primarily as a massive integer-search problem. This paper does not claim that a perfect cuboid must exist, nor does it infer the existence of rational points from the infinitude of the real solution space; on the contrary, we strictly distinguish the two. Its purpose is to establish a unified formal framework for subsequent research on ‘common initial state–constraint-induced differentiation’, ‘radix observation’, ‘exact rational reconstruction’, ‘first manifestation’ and ‘arithmetic freedom’. --- ## 1. The Traditional Problem Let the three edges of a rectangular box be $$ a,b,c>0. $$ Its three face diagonals are respectively $$ d,e,f>0, $$ and its space diagonal is $$ g>0. $$ A perfect cuboid requires: $$ a^2+b^2=d^2, $$ $$ a^2+c^2=e^2, $$ $$ b^2+c^2=f^2, $$ and $$ a^2+b^2+c^2=g^2. $$ The traditional problem additionally requires: $$ a,b,c,d,e,f,g\in\mathbb Z_{>0}. $$ Thus the problem can be written as: $$ \exists X\in\mathbb Z_{>0}^{7} \quad \text{such that} \quad F(X)=0, $$ where $$ X= (a,b,c,d,e,f,g), $$ and $$ F(X)= \begin{pmatrix} a^2+b^2-d^2\\ a^2+c^2-e^2\\ b^2+c^2-f^2\\ a^2+b^2+c^2-g^2 \end{pmatrix}. $$ This is a fully legitimate number-theoretic formulation. But this formulation also binds two things together: 1. Whether the geometric constraints can close; 2. Whether the resulting values simultaneously lie on positive-integer lattice points. This paper argues that these two layers should first be separated in research. --- ## 2. Geometric Closure and Arithmetic Closure Are Not the Same Problem If we temporarily remove the integer restriction and require only $$ a,b,c\in\mathbb R_{>0}, $$ we can directly define $$ d=\sqrt{a^2+b^2}, $$ $$ e=\sqrt{a^2+c^2}, $$ $$ f=\sqrt{b^2+c^2}, $$ and $$ g=\sqrt{a^2+b^2+c^2}. $$ Thus every positive-real triple $$ (a,b,c) $$ naturally corresponds to a real solution satisfying $$ F(X)=0 $$ . In other words, if the question is only ‘whether a rectangular box geometrically satisfies the four square relations’, the answer is not difficult. The real difficulty is whether $$ F(X)=0 $$ and $$ X\in\mathbb Z_{>0}^{7} $$ can hold simultaneously. We should therefore separate the following three questions: ### 2.1 Real Closure Does there exist $$ X\in\mathbb R_{>0}^{7} $$ such that $$ F(X)=0? $$ The answer is clearly yes. ### 2.2 Rational Closure Does there exist a nondegenerate $$ X\in\mathbb Q_{>0}^{7} $$ such that $$ F(X)=0? $$ This is the core question directly equivalent to integer existence for the perfect cuboid. ### 2.3 Integer Closure Does there exist $$ X\in\mathbb Z_{>0}^{7} $$ such that $$ F(X)=0? $$ This is the classical Perfect Cuboid problem. --- ## 3. Why Rational Closure Is the Core All four equations are homogeneous quadratics. If $$ X=(a,b,c,d,e,f,g) $$ is a solution, then for any $$ \lambda>0, $$ we have $$ F(\lambda X)=\lambda^2F(X)=0. $$ Thus, once we find a positive rational solution $$ X\in\mathbb Q_{>0}^{7}, $$ we can write $$ a=\frac{p_a}{q_a}, \quad b=\frac{p_b}{q_b}, \quad \ldots, \quad g=\frac{p_g}{q_g}, $$ where all $$ p_i,q_i\in\mathbb Z_{>0}. $$ Let $$ L= \operatorname{lcm} (q_a,q_b,q_c,q_d,q_e,q_f,q_g), $$ Then $$ LX\in\mathbb Z_{>0}^{7}. $$ Moreover, by homogeneity, $$ F(LX)=L^2F(X)=0. $$ This gives: $$ \boxed{ \exists X\in\mathbb Q_{>0}^{7},\ F(X)=0 \iff \exists Y\in\mathbb Z_{>0}^{7},\ F(Y)=0 } $$ This means: > The perfect cuboid problem need not begin with ‘finding enormous integers’ as the entry point for research. It can first be rewritten as: $$ \boxed{ \text{Does the continuous solution space contain a nondegenerate rational point?} } $$ Once a rational point is found, the integer solution is merely a matter of choosing a scale. --- ## 4. The First Important Distinction: Infinitely Many Real Solutions Do Not Imply a Rational Solution It must be emphasized that having a large $$ \mathcal M(\mathbb R) $$ does not directly imply that $$ \mathcal M(\mathbb Q) $$ is nonempty. Here $$ \mathcal M(K) = \{X\in K_{>0}^{7}:F(X)=0\} $$ denotes the closed solutions defined over the number field or set $K$. Although $$ \mathcal M(\mathbb R) $$ contains infinitely many solutions and even continuous families, this does not mean it must pass through a nondegenerate rational point. The following inference is therefore insufficient: $$ \text{The solution space is infinite} \Rightarrow \text{A rational solution must exist}. $$ This paper does not use this inference. What we actually seek to study is: $$ \boxed{ \mathcal M(\mathbb Q)\overset{?}{\neq}\varnothing } $$ That is: > Given that real closure is known to exist, what additional structures can force, permit or prevent rational closure from appearing? This question is the core of subsequent research. --- ## 5. From ‘Integer Search’ to ‘Closure Manifestation’ Traditional search strategies usually specify a height bound directly: $$ 1\le a,b,c,\ldots,g\le H, $$ They then search for points satisfying the conditions in a finite integer region. The concept is: $$ \text{Integer lattice} \rightarrow \text{constraint test} \rightarrow \text{accept/reject}. $$ This paper proposes another order of research: $$ \text{constraint space} \rightarrow \text{continuous closure} \rightarrow \text{arithmetic specialization} \rightarrow \text{integer manifestation}. $$ The two approaches are not mutually exclusive. Integer search can still serve as: - Counterexample search; - Lower-bound exploration; - Candidate-pattern discovery; - Numerical verification; - Algorithm comparison; - Height-distribution research. But if the aim is to understand ‘why solutions exist or do not exist’, searching only at ever-greater integer heights may not directly reveal the actual obstruction. --- ## 6. The Basic Form of the Solution Space Let $$ \mathcal M_{\mathbb R} = \left\{ X\in\mathbb R_{>0}^{7} : F(X)=0 \right\}. $$ Since we can freely choose $$ (a,b,c)\in\mathbb R_{>0}^{3}, $$ and then uniquely determine $$ (d,e,f,g), $$ within the positive-real region this solution space can be viewed as generated by three free parameters. Thus, in a locally regular region: $$ \dim \mathcal M_{\mathbb R}=3. $$ But there is an obvious scale degree of freedom: $$ X\sim \lambda X, \qquad \lambda>0. $$ If we care only about shape rather than overall size, we can therefore define the scale quotient space $$ \mathcal S_{\mathbb R} = \mathcal M_{\mathbb R}/\mathbb R_{>0}. $$ Its generic geometric number of degrees of freedom is $$ \dim \mathcal S_{\mathbb R}=2. $$ The actual geometric research object can therefore be viewed as a shape space with two degrees of freedom. The original problem: $$ X\in\mathbb Z_{>0}^{7} $$ then becomes: > Does this shape space contain a point that can be represented by rational ratios? --- ## 7. Normalization: Remove the Scale First Since scale does not affect whether the four constraints hold, we can choose a normalization condition. For example, let $$ g=1. $$ Define $$ x=\frac{a}{g}, \qquad y=\frac{b}{g}, \qquad z=\frac{c}{g}, $$ and $$ u=\frac{d}{g}, \qquad v=\frac{e}{g}, \qquad w=\frac{f}{g}. $$ The conditions become: $$ x^2+y^2=u^2, $$ $$ x^2+z^2=v^2, $$ $$ y^2+z^2=w^2, $$ and $$ x^2+y^2+z^2=1. $$ All solutions are then placed in the same normalized world. We no longer ask: > Are the edge lengths a hundred, a billion or $10^{100}$? Instead, we ask: > Does there exist a set of exact rational ratios for which all four closure conditions hold simultaneously? This is the shift from ‘numerical size’ to ‘ratio structure’. --- ## 8. The Perfect Cuboid as a Manifestation Problem This paper uses the term ‘manifestation’ to describe the following transition between layers: $$ \mathbb R \rightarrow \mathbb Q \rightarrow \mathbb Z. $$ ### 8.1 Real Manifestation The geometric constraints form a continuous valid state: $$ F(X)=0, \qquad X\in\mathbb R_{>0}^{7}. $$ ### 8.2 Rational Manifestation The same closed state additionally satisfies: $$ X\in\mathbb Q_{>0}^{7}. $$ ### 8.3 Integer Manifestation Clear the denominators using a common scale: $$ LX\in\mathbb Z_{>0}^{7}. $$ Thus an integer Perfect Cuboid can be reinterpreted as: $$ \boxed{ \text{An exact manifestation of a continuously closed world at the rational-representation layer} } $$ rather than being understood from the outset as an accidental collision of seven enormous positive integers. --- ## 9. ‘First Manifestation’ and ‘Existence’ Are Not the Same Question Suppose there exists a primitive integer solution $$ P= (a,b,c,d,e,f,g), $$ with $$ \gcd(a,b,c,d,e,f,g)=1. $$ Then for every $$ k\in\mathbb Z_{>0}, $$ we have that $$ kP $$ is also an integer solution. Thus: $$ \boxed{ \text{One integer solution exists} \Rightarrow \text{Infinitely many integer solutions exist} } $$ But all of these solutions may belong to the same scale orbit. We therefore need to distinguish: $$ \text{solution count} $$ and $$ \text{primitive shape count}. $$ Finding the first Perfect Cuboid does not automatically answer: - Whether there is only one primitive shape; - Whether there are infinitely many primitive shapes; - Whether there is a one-parameter family of rational solutions; - Whether there is a two-parameter family of rational solutions; - How the number of primitive solutions grows with height. These belong to the subsequent questions of a ‘second expansion’. --- ## 10. The First Research Hypothesis of This Series For now, this paper does not declare that a Perfect Cuboid must exist, but establishes only a hypothesis for study: ### Hypothesis PC-RC There exists at least one nondegenerate rational closure point: $$ \mathbf{H}_{\mathrm{PC-RC}}: \qquad \mathcal M(\mathbb Q)\neq\varnothing. $$ If this hypothesis holds, homogeneous scaling immediately gives: $$ \mathcal M(\mathbb Z_{>0})\neq\varnothing. $$ Thus what subsequent work actually needs to address is not: $$ \text{How can we directly hit an enormous integer point?} $$ but rather: $$ \text{How can we identify, generate or approach a rational closure point in the continuous solution space that admits exact reconstruction?} $$ --- ## 11. From a Search Model to a Generative Model We can abstract the two methods as: ### Traditional Search Model $$ X_1,X_2,\ldots,X_N\in\mathbb Z_{>0}^{7} $$ Test each one: $$ F(X_i)\overset{?}{=}0. $$ ### Constraint-Generation Model Choose an initial state $$ X_0, $$ and then construct an update operator $$ \Phi, $$ such that $$ X_{n+1}=\Phi(X_n), $$ with the hope that: $$ \|F(X_{n+1})\| < \|F(X_n)\|. $$ Eventually obtain: $$ X_n\rightarrow X^\ast, $$ where $$ F(X^\ast)=0. $$ If $$ X^\ast\in\mathbb Q_{>0}^{7}, $$ the problem is completed. If we can obtain only $$ X^\ast\in\mathbb R_{>0}^{7}, $$ exact rational reconstruction and verification are still needed. Thus subsequent research will center on: $$ \boxed{ \text{constraint-driven rational closure discovery} } $$ rather than merely brute-force integer search. --- ## 12. Floating-Point Values Are Not Rational Solutions Subsequent work in this paper will permit floating-point values as an exploration tool, but must strictly distinguish: $$ \operatorname{fl}(X) $$ and $$ X\in\mathbb Q^{7}. $$ Even if the computer obtains: $$ \|F(X)\|<10^{-100}, $$ this does not imply: $$ F(X)=0. $$ Moreover, because the equations are homogeneous quadratics, if we scale up an approximate solution: $$ X\mapsto \lambda X, $$ the residual scale becomes approximately: $$ F(\lambda X)=\lambda^2F(X). $$ Thus ‘very close’ does not automatically become ‘exact’ when scaled up. Subsequent computation must therefore follow: $$ \text{numerical discovery} \rightarrow \text{rational reconstruction} \rightarrow \text{exact symbolic verification}. $$ Any result called a candidate Perfect Cuboid must be re-verified using exact arithmetic at the final stage. --- ## 13. The Reformulation Obtained in This Paper The traditional problem: $$ \boxed{ \exists X\in\mathbb Z_{>0}^{7} : F(X)=0 } $$ can equivalently be rewritten as: $$ \boxed{ \exists [X]\in\mathcal S_{\mathbb Q} } $$ where $$ \mathcal S_{\mathbb Q} = \frac{ \{X\in\mathbb Q_{>0}^{7}:F(X)=0\} }{ X\sim \lambda X,\ \lambda\in\mathbb Q_{>0} }. $$ This paper therefore proposes: > The Perfect Cuboid can first be viewed as an existence problem for a rational shape class; an integer representative is merely the arithmetic manifestation of that shape class after scaling up by a common factor. This reformulation does not solve the conjecture itself, but changes the order of subsequent research. --- ## 14. Subsequent Research Questions Within this framework, subsequent questions naturally divide into several layers. ### Q1. Can Constraint Dynamics Differentiate a Common Initial State into a Closed State? Starting from $$ X_0=(s,s,s,s,s,s,s) $$ an undifferentiated state of this kind, can we define a natural constraint flow that brings the system into $$ \mathcal M_{\mathbb R}? $$ ### Q2. How Can We Distinguish Ordinary Real Closure from Rational Closure? Is there a computable quantity that measures how far a closed point is from ‘admitting exact rational reconstruction’? ### Q3. Can the Radix Serve as an Additional Observation Operator? For different radices $$ B=2,3,4,5,\ldots, $$ does the same underlying numerical structure display different finite representations, carry structures and local closure patterns? ### Q4. Are There Cross-Radix Invariants? If certain patterns remain stable across multiple radices, do they impose stronger restrictions on rational closure? ### Q5. After the First Rational Solution Appears, How Do Primitive Solution Families Expand? Does there exist a nontrivial solution-preserving transformation: $$ T(P)\not\sim P, $$ for which, nevertheless, $$ F(P)=0 \Rightarrow F(T(P))=0? $$ ### Q6. How Many Arithmetic Degrees of Freedom Are There? Even if the geometric shape space has two degrees of freedom, the actual family of rational solutions may be: $$ 0\text{-dimensional}, $$ $$ 1\text{-dimensional}, $$ or $$ 2\text{-dimensional}. $$ This will be a second-stage question after the first solution. --- ## 15. Methodological Principles Subsequent research in this series follows these principles. ### Principle A: Do Not Treat Approximation as Proof Any floating-point result is used only as a candidate-generation tool. ### Principle B: Do Not Mistake Continuous Existence for Rational Existence $$ \mathcal M(\mathbb R)\neq\varnothing $$ does not imply $$ \mathcal M(\mathbb Q)\neq\varnothing. $$ ### Principle C: Scale Is Not the Core Degree of Freedom All purely scaled solutions are regarded as the same shape class. ### Principle D: Study Structure before Height Where possible, do not prioritize ever-larger upper bounds for integer search, but first study the closure mechanism itself. ### Principle E: Formal Results Must Be Exactly Verified If a candidate rational solution is found, final verification must use exact arithmetic, not an IEEE floating-point approximation. ### Principle F: Separate the Existence Question from the Freedom Question First stage: $$ \mathcal M(\mathbb Q)\overset{?}{\neq}\varnothing. $$ Second stage: $$ \#\mathcal S_{\mathbb Q}, \quad \dim_{\mathrm{arith}}\mathcal S_{\mathbb Q}, \quad N(H) $$ How do these behave? --- ## 16. Conclusion The classical formulation of the Perfect Cuboid requires seven positive integers to satisfy four square relations simultaneously, and has therefore long been naturally regarded as a highly difficult Diophantine search problem. The first step proposed here is not to reject that formulation, but to separate its layers. We first define the continuously closed world: $$ \mathcal M_{\mathbb R} = \{X\in\mathbb R_{>0}^{7}:F(X)=0\}. $$ Then quotient out the scale degree of freedom: $$ \mathcal S_{\mathbb R} = \mathcal M_{\mathbb R}/\mathbb R_{>0}. $$ We then ask the actual question: $$ \boxed{ \mathcal S_{\mathbb Q}\overset{?}{\neq}\varnothing } $$ If the answer is yes, an integer Perfect Cuboid can be generated directly by a common scale. Thus the original question can change from: $$ \text{‘At what enormous integer height can we find the first point?’} $$ to: $$ \text{‘When does the continuously closed world produce its first exact rational manifestation?’} $$ This reformulation will be the starting point of the entire subsequent series. The next paper will address: # The Continuous World and the Scale Quotient Space: The Topological Starting Point of the Perfect Cuboid The focus will move from the equation system alone to: $$ \mathcal M_{\mathbb R}, \qquad \mathcal S_{\mathbb R}, $$ and the shape degrees of freedom, boundaries, degenerate regions and observable structures that actually remain for the Perfect Cuboid after removing the overall scale. --- ## Provisional Contents of the Series 1. **Reframing the Perfect Cuboid: It Is Not Primarily an Integer Search Problem** 2. **The Continuous World and the Scale Quotient Space: The Topological Starting Point of the Perfect Cuboid** 3. **Common Initial State and Constraint-Induced Differentiation** 4. **The Radix as an Observation Operator: Radix–Carry Global Closure** 5. **From Floating-Point Approximation to Exact Rational Closure** 6. **The First Manifestation Problem: From a Rational Solution to the Minimal Primitive Cuboid** 7. **After the First Solution: Arithmetic Freedom of the Perfect Cuboid** 8. **Upper and Lower Bounds of the Solution Space and a Second Expansion** --- **End of Paper 01**