# The Center-Generated Bidirectional Offset Spiral Proposition ## —Formalized Geometry of Complete Turns, Positive Thickness, Non-Overlapping, and Spiral Annular Bands **Author:** Neo.K **Version:** v0.1 **Date:** July 27, 2026 **Nature of Research:** Geometric Proposition Paper / Formal Foundation Draft **Status:** Basic tubular geometry and classification propositions are directly provable; contact saturation and global optimization problems remain to be studied. --- ## Abstract This paper formalizes a class of planar geometry generated outward from a center, possessing positive thickness, non-overlapping interiors, but allowing boundary contact. Its basic skeleton is a regular curve originating from a center, with its direction continuously rotating with respect to the parameter; it is then simultaneously offset outward and inward along the curve's normal, forming a band-like region with a positive two-dimensional measure. This paper first clarifies: a total turning angle of $360^\circ$ for a planar curve is insufficient to deduce that the curve is a circle. Only when the curvature is a constant everywhere is a complete $2\pi$ turning unit a circle. When the curve continues to increase its radial scale after completing a full turn, the centerline is no longer a single circle but forms an outwardly expanding spiral. Let the centerline be $$ \gamma:[0,L]\to\mathbb R^2, $$ Given a thickness radius $\rho>0$, its bidirectional normal offset band is defined as $$ S_\rho(\gamma) = \left\{ \gamma(s)+tN(s): 0\le s\le L,\; -\rho\le t\le\rho \right\}. $$ If the normal parameterization is injective in the interior, the interiors of different layers do not overlap; if only boundary tangency or joining is allowed, it can be viewed as a limiting case of contact saturation. For a valid open curve, the area of the normal band without end caps is $$ \mu_2(S_\rho(\gamma))=2\rho L, $$ After adding semicircular caps at both ends, the area of the complete tubular neighborhood is $$ \mu_2(T_\rho(\gamma)) = 2\rho L+\pi\rho^2. $$ This paper proposes the "Center-Generated Bidirectional Offset Spiral Proposition": if the centerline starts from the center, its radial scale does not regress, it maintains a positive radial increment after completing a full turn, and it satisfies the non-overlapping interior condition, then its global geometry is a spiral centerline; after bidirectional offsetting, it forms a spiral annular band. If the radial increment is zero, it degenerates into a circle and its annulus; if the inward offset fills the central hole, it degenerates into a disk or a spiral-generated disk. This paper also distinguishes between separated, contact, and overlapping spiral bands, explaining that the radial pitch of an Archimedean spiral can only serve as a proxy for layer spacing, whereas the precise non-overlapping condition should be controlled by the curve's reach. **Keywords:** Center generation, bidirectional offset, spiral annular band, complete turn, positive thickness, tubular neighborhood, reach, curvature, non-overlapping, boundary contact --- # 1. Starting Point of the Problem Consider a "crooked line" in a two-dimensional plane. We require: 1. It starts from a certain center position; 2. It continuously extends outward; 3. It can turn; 4. The interiors of different parts cannot overlap; 5. Boundaries can be tangent, joined, or share contact points; 6. The final geometry has a positive two-dimensional measure; 7. The centerline is simultaneously offset inward and outward. This set of conditions does not describe an ordinary zero-thickness curve, but rather describes: $$ \boxed{ \text{A center curve and its positive-thickness normal band} }. $$ If the centerline only completes one full turn and the radius is fixed, the most natural shape is a circle. If it continues to generate outward after completing one full turn, the most natural global shape is a spiral. --- # 2. Basic Space and Measure Let $$ \mathbb R^2 $$ be the two-dimensional Euclidean plane, and the two-dimensional Lebesgue measure be denoted as $$ \mu_2. $$ Let $$ \gamma:[0,L]\to\mathbb R^2 $$ be a regular curve of finite length that is piecewise $C^2$. If $\gamma$ is merely an ordinary rectifiable curve, then $$ \mu_2(\gamma([0,L]))=0. $$ Therefore, a "non-zero minimum measure" cannot be provided by the pure centerline itself, but requires the introduction of positive thickness. --- # 3. Positive Thickness and Bidirectional Normal Offset Assume $\gamma$ is parameterized by arc length: $$ \|\gamma'(s)\|=1. $$ Define the unit tangent vector: $$ T(s)=\gamma'(s). $$ Select the unit normal vector: $$ N(s)=R_{\pi/2}T(s). $$ Given the thickness radius $$ \rho>0, $$ define the normal parameterization: $$ F_\rho(s,t) = \gamma(s)+tN(s), \qquad (s,t)\in[0,L]\times[-\rho,\rho]. $$ Its image set $$ \boxed{ S_\rho(\gamma) = F_\rho \left( [0,L]\times[-\rho,\rho] \right) } $$ is called the bidirectional normal offset band of the center curve. Its two local boundaries are $$ \gamma_+(s) = \gamma(s)+\rho N(s), $$ $$ \gamma_-(s) = \gamma(s)-\rho N(s). $$ --- # 4. Complete Tubular Neighborhood The normal offset band $S_\rho(\gamma)$ does not include the circular caps extending in the direction of the curve's endpoints. If we take the Minkowski sum of the curve and a closed disk $B_\rho$ of radius $\rho$, defining: $$ \boxed{ T_\rho(\gamma) = \gamma([0,L])\oplus B_\rho } $$ Then $T_\rho(\gamma)$ is the complete tubular neighborhood. For an open curve, it equals: $$ \text{Normal band} + \text{Two semicircular end caps}. $$ This paper will: - Call $S_\rho(\gamma)$ the offset band without end caps; - Call $T_\rho(\gamma)$ the complete thickened region. --- # 5. Non-Overlapping and Allowed Contact ## Definition 5.1: Strictly Non-Overlapping If $$ F_\rho: [0,L]\times(-\rho,\rho) \to\mathbb R^2 $$ is injective, the normal band is said to be strictly non-overlapping in its interior. That is, if $$ F_\rho(s,t)=F_\rho(s',t'), $$ and $$ |t|<\rho, \qquad |t'|<\rho, $$ then we must have $$ s=s', \qquad t=t'. $$ ## Definition 5.2: Boundary Contact Allowed It is allowed for different parameter points $$ (s,t)\ne(s',t') $$ such that $$ F_\rho(s,t)=F_\rho(s',t'), $$ but this can only occur when $$ |t|=\rho \quad\text{or}\quad |t'|=\rho. $$ Therefore $$ \operatorname{Int}S_\rho(\gamma) $$ must not self-overlap, but $$ \partial S_\rho(\gamma) $$ can be tangent or joined. --- # 6. Reach Condition For a closed set $X\subset\mathbb R^2$, its reach is defined as: the maximum radius of a unique projection neighborhood such that all points sufficiently close to $X$ have a unique nearest point. For the center curve $\gamma$, if $$ \operatorname{reach}(\gamma)>\rho, $$ then: 1. Points with a distance to the centerline less than $\rho$ have a unique nearest center point; 2. The normal parameterization remains injective within thickness $\rho$; 3. The interior of the normal band does not overlap; 4. The local radius of curvature is greater than $\rho$. Therefore $$ \boxed{ \operatorname{reach}(\gamma)>\rho } $$ is the standard strictly valid condition adopted in this paper. The contact-saturated boundary is written as $$ \boxed{ \operatorname{reach}(\gamma)=\rho }. $$ In this case, the normal band may experience tangency at the boundary for the first time, but the interior still has not produced any positive-area overlap. --- # 7. Curvature and Complete Turn Let the tangential angle be $$ \theta(s), $$ such that $$ T(s) = (\cos\theta(s),\sin\theta(s)). $$ The curvature is $$ \kappa(s)=\theta'(s). $$ The total turn of the curve over the interval $[a,b]$ is $$ \Delta\theta = \theta(b)-\theta(a) = \int_a^b\kappa(s)\,ds. $$ One complete directed turn is $$ \Delta\theta=2\pi. $$ --- # 8. Total Turn of $2\pi$ Does Not Equal a Circle We must distinguish between: $$ \int\kappa\,ds=2\pi $$ and $$ \kappa(s)\equiv\frac1R. $$ The former merely indicates that the total tangential rotation is $360^\circ$. Many non-circular simple closed convex curves also satisfy a total turn of $2\pi$, for example: - Ellipses; - Non-uniform convex ovals; - Smooth irregular convex curves. Therefore $$ \boxed{ \text{Total turn of }2\pi \not\Rightarrow \text{Curve is a circle} }. $$ --- # 9. Circle Generation Theorem ## Theorem 9.1: Constant Curvature Complete Turning Unit Let $\gamma$ be an arc-length parameterized curve satisfying $$ \kappa(s)\equiv\frac1R, \qquad R>0, $$ and $$ b-a=2\pi R. $$ on some interval $[a,b]$. Then this curve segment is a complete circle of radius $R$, up to a planar rigid body motion. ### Proof From $$ \theta'(s)=\frac1R, $$ we obtain $$ \theta(s) = \theta(a)+\frac{s-a}{R}. $$ Integrating again $$ \gamma'(s) = (\cos\theta(s),\sin\theta(s)), $$ yields $$ \gamma(s) = c+ R \left( \sin\theta(s), -\cos\theta(s) \right) $$ which holds for some fixed point $c$. Therefore $$ \|\gamma(s)-c\|=R. $$ Moreover, since $$ \theta(b)-\theta(a)=2\pi, $$ the segment completes exactly one full loop. Q.E.D. ## Corollary 9.2 In the geometric language of this paper: $$ \boxed{ \text{A circle is a complete }360^\circ\text{ turning unit with constant curvature} }. $$ --- # 10. Center Generation Condition "Generated outward from the center" can be expressed using polar coordinates. Let $$ \gamma(\vartheta) = r(\vartheta) (\cos\vartheta,\sin\vartheta), $$ where $$ 0\le\vartheta\le\Theta. $$ Require $$ r(0)=0, $$ and $$ r'(\vartheta)\ge0. $$ This means that as the polar angle increases, the curve will not regress to a smaller radius along the same ray. More generally, if arc-length parameterization is used, we can require $$ \frac{d}{ds} \|\gamma(s)\| \ge0 $$ to hold where differentiable. This paper refers to this type of condition as $$ \boxed{ \text{Radially non-degenerate center generation condition} }. $$ --- # 11. Radial Increment Per Loop Define the radial increment after one complete loop: $$ \Delta_{2\pi}r(\vartheta) = r(\vartheta+2\pi)-r(\vartheta). $$ We can distinguish: ## Zero Radial Increment $$ \Delta_{2\pi}r(\vartheta)=0. $$ ## Positive Radial Increment $$ \Delta_{2\pi}r(\vartheta)>0. $$ Zero radial increment means the curve returns to the same radius after each full turn. Positive radial increment means the curve moves to an outer layer after completing each loop. --- # 12. Zero Increment Degenerates to a Circle ## Proposition 12.1 If $$ r'(\vartheta)\ge0 $$ and for all applicable $\vartheta$ $$ r(\vartheta+2\pi)=r(\vartheta), $$ then $r$ is constant over each complete period. If we further require the curve to complete a full turn without repeated winding, the centerline degenerates into a circle: $$ \gamma(\vartheta) = R(\cos\vartheta,\sin\vartheta). $$ ### Proof By monotonic non-decrease: $$ r(\vartheta) \le r(\vartheta+\eta) \le r(\vartheta+2\pi) $$ holds for $$ 0\le\eta\le2\pi. $$ But the period endpoints are equal: $$ r(\vartheta+2\pi)=r(\vartheta). $$ Hence all intermediate values are equal. Q.E.D. --- # 13. Bidirectional Offset of a Circle If the centerline is a circle of radius $R$: $$ \gamma(\vartheta) = R(\cos\vartheta,\sin\vartheta), $$ then its inner and outer offset boundaries are: $$ \gamma_+(\vartheta) = (R+\rho) (\cos\vartheta,\sin\vartheta), $$ $$ \gamma_-(\vartheta) = (R-\rho) (\cos\vartheta,\sin\vartheta). $$ If $$ 0<\rho r(\vartheta), $$ then the centerline cannot be a single fixed-radius circle. If this positive increment persists, the centerline forms an outwardly expanding spiral curve. ### Proof A single circle must satisfy $$ r(\vartheta)\equiv R. $$ Therefore $$ r(\vartheta+2\pi)-r(\vartheta)=0. $$ which contradicts the positive increment condition. Q.E.D. --- # 15. Center-Generated Spiral This paper defines a center-generated spiral as a polar coordinate curve $$ \gamma(\vartheta) = r(\vartheta) (\cos\vartheta,\sin\vartheta), $$ satisfying: 1. $r(0)=0$; 2. $r'(\vartheta)\ge0$; 3. A complete turn exists; 4. At least some complete periods have a positive radial increment; 5. The curve is regular; 6. After thickening, it satisfies the non-overlapping interior condition. This definition does not require $$ r(\vartheta) $$ to be a linear function. Therefore it includes: - Archimedean spirals; - Non-uniform pitch spirals; - Variable curvature spirals; - Piecewise smooth spirals; - Curvature-concentrated spirals. --- # 16. Archimedean Spiral If $$ r'(\vartheta)=b>0, $$ then $$ r(\vartheta)=r_0+b\vartheta. $$ When starting from the center $$ r_0=0, $$ hence $$ \boxed{ r(\vartheta)=b\vartheta }. $$ After completing each loop, its radial difference along the same ray is $$ p = r(\vartheta+2\pi)-r(\vartheta) = 2\pi b. $$ where $p$ is called the radial pitch. --- # 17. Radial Pitch is Not a Precise Non-Overlapping Condition For an Archimedean spiral, the radial distance between the centerlines of two adjacent loops in the same polar angle direction is indeed $$ p=2\pi b. $$ Thus, a common approximate classification is: $$ p>2\rho $$ layer separation when, $$ p=2\rho $$ layer contact when, $$ p<2\rho $$ layer overlap when. But it must be noted: > The normal offset direction is generally not equal to the pure radial direction. Therefore $$ p\ge2\rho $$ cannot replace the precise global non-overlapping criterion in the general case. The precise condition should still be written as $$ \boxed{ \operatorname{reach}(\gamma)\ge\rho }. $$ Radial pitch is merely a convenient preliminary proxy. --- # 18. The Center-Generated Bidirectional Offset Spiral Proposition ## Main Proposition Let $$ \gamma(\vartheta) = r(\vartheta) (\cos\vartheta,\sin\vartheta) $$ be a piecewise $C^2$ regular curve, satisfying: 1. Center origin: $$ r(0)=0; $$ 2. Radially non-degenerate: $$ r'(\vartheta)\ge0; $$ 3. Complete turn exists: $$ \Theta\ge2\pi; $$ 4. The curve can continue to extend after completing one loop; 5. Given thickness: $$ \rho>0; $$ 6. Non-overlapping interior of the normal band: $$ \operatorname{reach}(\gamma)\ge\rho; $$ 7. Boundaries can contact, but must not produce positive-area interior overlap. Then we have: ### Case A: Zero Periodic Radial Increment If $$ r(\vartheta+2\pi)=r(\vartheta) $$ holds for all applicable $\vartheta$, then the centerline degenerates into a fixed-radius complete turning curve; under constant curvature conditions, it is a circle. Its bidirectional offset geometry is an $$ \boxed{ \text{Annulus} } $$ or, when the inner hole disappears, a $$ \boxed{ \text{Disk} }. $$ ### Case B: Positive Periodic Radial Increment If there exists $$ \Delta_{2\pi}r(\vartheta)>0, $$ then the centerline is not a single circle, but an outwardly expanding spiral curve. Its bidirectional normal offset region is a $$ \boxed{ \text{Center-generated bidirectional offset spiral annular band} }. $$ ### Case C: Contact Saturation If $$ \operatorname{reach}(\gamma)=\rho, $$ and tangency only occurs for the first time at the boundary, it forms a $$ \boxed{ \text{Contact spiral annular band} }. $$ If the contact sequence further encloses all internal gaps, its overall closure can be viewed as a $$ \boxed{ \text{Spiral-generated disk} }. $$ --- # 19. Precise Definition of the Spiral Annular Band ## Definition 19.1 For a center-generated spiral $\gamma$, define $$ \boxed{ \mathcal S_{\gamma,\rho} = S_\rho(\gamma) } $$ as the center-generated bidirectional offset spiral annular band. If end caps are added $$ \boxed{ \mathcal T_{\gamma,\rho} = T_\rho(\gamma) } $$ it is called the complete spiral tubular region. --- # 20. Geometric Classification ## 20.1 Strictly Separated Type If $$ \operatorname{reach}(\gamma)>\rho, $$ then both the interiors and boundaries of different normal fibers maintain a positive separation distance. It is called a $$ \boxed{ \text{Strictly separated spiral annular band} }. $$ ## 20.2 Contact Type If $$ \operatorname{reach}(\gamma)=\rho, $$ and only boundary contact occurs, it is called a $$ \boxed{ \text{Contact spiral annular band} }. $$ ## 20.3 Overlapping Type If $$ \operatorname{reach}(\gamma)<\rho, $$ then there exist points in the positive-thickness neighborhood with multiple nearest center points, and the normal band may exhibit interior overlap. In this case, it is called an $$ \boxed{ \text{Overlapping spiral band} }. $$ It does not satisfy the non-overlapping interior condition of the main proposition of this paper. --- # 21. Area Formula ## Theorem 21.1: Area of the Offset Band Without End Caps If $$ \operatorname{reach}(\gamma)>\rho, $$ then $$ \boxed{ \mu_2(S_\rho(\gamma)) = 2\rho L }. $$ ### Proof From the Frenet formulas $$ N'(s)=-\kappa(s)T(s), $$ we obtain $$ \partial_sF_\rho = (1-t\kappa(s))T(s), $$ $$ \partial_tF_\rho=N(s). $$ The Jacobian is $$ J(s,t) = 1-t\kappa(s). $$ Therefore $$ \begin{aligned} \mu_2(S_\rho(\gamma)) &= \int_0^L \int_{-\rho}^{\rho} (1-t\kappa(s)) \,dt\,ds\\ &= \int_0^L2\rho\,ds\\ &= 2\rho L. \end{aligned} $$ Q.E.D. ## Corollary 21.2: Tubular Area of the Complete Open Curve After adding semicircular caps at both ends $$ \boxed{ \mu_2(T_\rho(\gamma)) = 2\rho L+\pi\rho^2 }. $$ ## Corollary 21.3: Non-Zero Area As long as $$ L>0, \qquad \rho>0, $$ we have $$ \boxed{ \mu_2(S_\rho(\gamma))>0 }. $$ Therefore, the positive thickness condition ensures the geometry is no longer a two-dimensional zero-measure pure line. --- # 22. Area Conservation and Shape Differences For all valid non-overlapping centerlines of the same length and same thickness $$ \mu_2(S_\rho(\gamma)) = 2\rho L. $$ So $$ \boxed{ \text{Annuli, spiral annular bands, and other valid normal bands can have the same total area} }. $$ But they may still have different: - Curvature distributions; - Support functions; - Convex hulls; - Diameters; - Directional widths; - Topological contact structures; - Container embedding difficulties. Therefore $$ \boxed{ \text{Same total amount} \not\Rightarrow \text{Same geometry} }. $$ --- # 23. Local Circle and Global Spiral This paper distinguishes between the following two statements. ## Local Complete Turn Language If a certain turning unit has constant curvature, then $$ \boxed{ \text{This unit is a circle} }. $$ ## Global Extrapolation Language If the radial scale still increases after each complete turn, then $$ \boxed{ \text{The globally connected shape is a spiral} }. $$ Therefore $$ \boxed{ \text{A circle is a local constant-curvature complete turning unit; a spiral is the global connection language of multiple outward turning units} }. $$ --- # 24. Spiral-Generated Disk "Spiral annular band" and "spiral-generated disk" must be distinguished. ## Spiral Annular Band If gaps still exist between the thickened spiral layers, the region is not a complete disk. ## Spiral-Generated Disk If: 1. The layers are in continuous contact at the boundaries; 2. No positive-area overlap exists; 3. No open-set gaps are left between the center and the outermost layer; 4. The closure of the resulting region is a disk-like set; Then it is called a $$ \boxed{ \text{Spiral-generated disk} }. $$ This is a global filling condition and cannot be deduced solely from a single pitch equation. --- # 25. Topological Classification For a strictly separated finite open spiral band, excluding end caps, its parameter domain is $$ [0,L]\times[-\rho,\rho]. $$ Under interior injectivity, its topological type is equivalent to a closed rectangle $$ \boxed{ D^2 }. $$ For a closed circular centerline, its annular topological type is $$ \boxed{ S^1\times[0,1] }. $$ If the central hole disappears, the topological type becomes $$ \boxed{ D^2 }. $$ So $$ \text{Annulus} \quad\text{and}\quad \text{Open spiral band} $$ Even if both visually have "wrapping," their topological origins can still be different. --- # 26. Relationship with Pure Circle Language If the geometric language stipulates: > All complete $360^\circ$ turns must be completed with constant curvature, Then each independent complete turning unit is a circle. But if the centers or radii of these turning units continuously change, after joining them continuously, the whole is no longer a single circle. Therefore, the correct form is $$ \boxed{ \text{Each fixed-scale complete turning unit can be a circle; the multi-scale continuously extrapolated whole can be a spiral} }. $$ It cannot be directly written as $$ \text{All geometries with a total turning angle of }360^\circ\text{ are circles}. $$ --- # 27. Main Corollaries ## Corollary A: Necessity of Positive Thickness If a positive two-dimensional measure is required, one must study $$ S_\rho(\gamma) \quad\text{or}\quad T_\rho(\gamma), $$ rather than the pure centerline. ## Corollary B: Annulus Degeneration If the radial increment is zero and the curvature is constant, the centerline is a circle, and the bidirectional offset is an annulus. ## Corollary C: Spiral Annular Band Generation If there is still a positive radial increment after completing a full turn, the global centerline is not a circle, but a spiral curve; its bidirectional offset is a spiral annular band. ## Corollary D: Contact Saturation The precise boundary condition for a geometry that "can contact but cannot overlap" is $$ \operatorname{reach}(\gamma)=\rho. $$ ## Corollary E: Area is Independent of Curvature Distribution As long as the normal band is valid and non-self-intersecting $$ \mu_2(S_\rho(\gamma))=2\rho L. $$ Therefore, curvature affects the shape, but does not affect the total area of the band without end caps. --- # 28. New Research Problems ## Problem 28.1: Existence of Contact-Saturated Spirals Given $$ L,\rho,\Theta, $$ does there exist a center-generated spiral satisfying $$ \operatorname{reach}(\gamma)=\rho $$ and completing a specified total turn? ## Problem 28.2: Precise Conditions for Spiral-Generated Disks What conditions can guarantee that $$ \overline{S_\rho(\gamma)} $$ is a disk-like set, and there are no open-set gaps from the center to the outermost boundary? ## Problem 28.3: Minimum Enclosing Circle Under fixed $$ L,\rho,\Theta $$ which valid center-generated spiral minimizes the radius of the minimum enclosing circle of its bidirectional offset band? ## Problem 28.4: Maximum Support Pressure Under fixed length and thickness, which curvature distribution makes $$ h_{T_\rho(\gamma)}(u) $$ the most difficult to be mutually accommodated in certain directions? ## Problem 28.5: Constant Curvature and Curvature Concentration Is the most effective extrapolated geometry: - Constant curvature spiral; - Piecewise constant curvature; - Finite-width curvature concentration layer; - Contact-saturated variable curvature spiral? --- # 29. Formalization Interface If proceeding to Lean, Coq, or other formal systems in the future, it can be processed in the following order. ## Layer 1: Curve and Frenet Frame Formalize $$ T=\gamma', \qquad N=R_{\pi/2}T, \qquad T'=\kappa N. $$ ## Layer 2: Normal Mapping Formalize $$ F_\rho(s,t)=\gamma(s)+tN(s). $$ ## Layer 3: Jacobian Prove $$ \det (\partial_sF_\rho,\partial_tF_\rho) = 1-t\kappa(s). $$ ## Layer 4: Area Formula Prove $$ \mu_2(S_\rho(\gamma))=2\rho L. $$ under the conditions of injectivity and positive Jacobian. ## Layer 5: Constant Curvature Circle Prove that $$ \kappa\equiv1/R $$ and complete arc length $$ L=2\pi R $$ deduce a circle. ## Layer 6: Polar Coordinate Classification Formalize $$ r'(\vartheta)\ge0 $$ and periodic increment $$ r(\vartheta+2\pi)-r(\vartheta). $$ --- # 30. Limitations This paper does not prove: 1. All center-generated curves must be Archimedean spirals; 2. The radial pitch $2\pi b$ is the precise layer spacing for general normal bands; 3. $\operatorname{reach}(\gamma)=\rho$ necessarily fills the entire interior completely; 4. A contact spiral definitely forms a disk; 5. A specific spiral is the optimal solution for area, enclosing circle, or universal accommodation problems; 6. All curves with a total turning angle of $2\pi$ are circles; 7. This paper has solved the Kakeya problem, Moser's worm problem, or other universal covering problems. What this paper accomplishes is: - Precise definition of the geometric objects; - Logical classification of circles and spirals; - Reach formulation of non-overlapping and contact; - Positive area formula for bidirectional offset bands; - Fundamental propositions required for subsequent bridging theories. --- # 31. Conclusion This paper formalizes the center-generated bidirectional offset spiral. Its basic geometric chain is: $$ \boxed{ \text{Center origin} \longrightarrow \text{Continuous direction rotation} \longrightarrow \text{Complete }360^\circ \longrightarrow \text{Still extending outward} \longrightarrow \text{Spiral centerline} \longrightarrow \text{Bidirectional normal offset} \longrightarrow \text{Spiral annular band} }. $$ If there is no radial increment, and a constant curvature condition is added, then $$ \boxed{ \text{The centerline is a circle, and the offset region is an annulus or a disk} }. $$ If it has a positive radial increment, then $$ \boxed{ \text{The centerline is a spiral curve, and the offset region is a spiral annular band} }. $$ If $$ \operatorname{reach}(\gamma)=\rho, $$ then $$ \boxed{ \text{Forms a contact-saturated state that allows boundary contact but not interior overlap} }. $$ Its area under valid conditions is $$ \boxed{ \mu_2(S_\rho(\gamma))=2\rho L }. $$ Therefore, the core of this proposition can be condensed as: $$ \boxed{ \text{A circle is a fixed-scale constant-curvature complete turning unit; a spiral is the global geometry of a complete turning unit continuously extending outward; bidirectional offsetting transforms a zero-area centerline into a positive-area spiral annular band} }. $$ --- # Appendix A: Table of Symbols | Symbol | Meaning | |---|---| | $\gamma$ | Center-generated curve | | $L$ | Length of the center curve | | $T(s)$ | Unit tangent vector | | $N(s)$ | Unit normal vector | | $\theta(s)$ | Tangential angle | | $\kappa(s)$ | Curvature | | $\rho$ | Thickness radius | | $F_\rho(s,t)$ | Normal parameterization | | $S_\rho(\gamma)$ | Bidirectional offset band without end caps | | $T_\rho(\gamma)$ | Complete tubular neighborhood | | $r(\vartheta)$ | Polar coordinate radius function | | $\Delta_{2\pi}r$ | Radial increment after one loop | | $\operatorname{reach}(\gamma)$ | Tubular radius of unique nearest point | | $p$ | Radial pitch of the Archimedean spiral | --- # Appendix B: Shortest Proposition Version Let $\gamma$ be a planar regular curve starting from the center, with a non-decreasing radial scale, and completing at least one full turn. Take its bidirectional normal offset with a thickness radius $\rho>0$, requiring the interior of the normal band to be non-overlapping while boundaries can contact. Then: $$ \Delta_{2\pi}r=0 $$ and when curvature is constant, the centerline is a circle, and after thickening it is an annulus or a disk; $$ \Delta_{2\pi}r>0 $$ when, the centerline is an outwardly generated spiral curve, and after thickening it is a center-generated bidirectional offset spiral annular band; $$ \operatorname{reach}(\gamma)=\rho $$ describes the contact-saturated state where boundaries exactly contact; Under strictly valid tubular conditions: $$ \mu_2(S_\rho(\gamma))=2\rho L. $$