---

Discrete Order Geometry (DOG) and Classical Fractal Geometry: Paradigm Comparison and Pain Point Resolution

Author: Zhang Suhang

Luoyang

Abstract: Classical fractal geometry, with its core assumptions of infinite self‑similarity and connected nesting, suffers from several inherent pain points in both theory and application: the contradiction between infinite processes and physical forms, the undecidability of certain decision problems, inaccuracies in box‑counting dimension estimation, and physical paradoxes arising from infinite construction. Based on the axiomatic system of Discrete Order Geometry (DOG), this paper analyzes the roots of these pain points point by point, and demonstrates how DOG systematically resolves them through finite discreteness, order lattice points, continued fraction hierarchical convergence, and state‑machine modeling. On this basis, this paper further argues that connected adjacency is a local special case, while discrete, spatially separated order is the universal norm of the cosmos. DOG is not a patch or improvement of fractal geometry, but a paradigm shift from “infinite connected self‑similarity” to “finite discrete order”, and at the same time accomplishes a historical reset of geometry.

Keywords: Discrete Order Geometry; fractal geometry; paradigm comparison; finite state machine; continued fractions; connectedness as special case

---

I. Introduction

Fractal geometry has achieved great success in describing natural forms such as coastlines, snowflakes, and clouds. However, its core assumptions—infinitely fine self‑similar structures and connected nesting—have also brought several unavoidable pain points. These pain points arise not from insufficient computational techniques, but from inherent limitations of the underlying paradigm.

Discrete Order Geometry (DOG), with its core principles of finite discrete order lattice points, order priority, and closed recursion, offers a distinctly different geometric worldview. This paper aims to compare, point by point, the classical pain points of fractal geometry with DOG’s solutions, demonstrating that DOG can systematically resolve them. On this basis, this paper further accomplishes a historical reset of geometry.

---

II. DOG Resolution of Theoretical and Philosophical Deficiencies

Pain Point 1: Infinite Process vs. Physical Form

· Fractal: Objects such as the Koch snowflake and the Mandelbrot set are strictly defined as limits of infinite recursion. Infinitely fine structures do not exist in nature; fractals can only be “approximate descriptions.”

· DOG: Axiom 1 of DOG explicitly states that all system states are constrained to a finite, countable set of order lattice points. Continued fraction convergence is truncated at finite orders, directly yielding usable numerical values. No infinite processes, no limit concepts required.

Pain Point 2: Unstable Logical Core of Key Concepts

· Fractal: Concepts such as self‑similarity and fractal dimension are philosophically controversial and difficult to map clearly onto physical entities.

· DOG: DOG is built on three testable axioms (spatial discretization, order priority, closed recursion). Order lattice points, hierarchical deflection angles, and coupling constants have clear geometric and physical referents, with no ontological ambiguity.

Pain Point 3: Descriptive Limitations and Fading Hype

· Fractal: Primarily used for morphological description, with limited predictive and constructive power.

· DOG: DOG is a constructive paradigm. It provides evolution equations, state machines, and recursive rules, enabling orbit generation, chaos trend prediction, and the design of artificial order systems—not limited to post‑hoc description.

---

III. DOG Resolution of Application and Computational Bottlenecks

Pain Point 4: Unreliability of Box‑Counting Methods

· Fractal: Computation of fractal dimension is highly sensitive to boundary choices and measurement regions; different researchers may obtain different results.

· DOG: DOG does not use dimensional fitting. Hierarchical order levels are determined directly by the order of continued fraction convergence, each level corresponding to a definite order lattice density, free from boundary sensitivity.

Pain Point 5: Undecidability of Certain Decision Problems

· Fractal: For example, whether the attractor of a given iterated function system intersects a line segment has been proven undecidable.

· DOG: DOG transforms the system into a finite state machine (finite set of order lattice points + deterministic transition rules). On a finite state machine, decision problems such as intersection, reachability, and periodicity are all decidable (via exhaustive search or graph algorithms).

Pain Point 6: Limited Predictive Power for Generalization

· Fractal: Fractal dimension was once used to predict the generalization capacity of neural networks but later proved unreliable.

· DOG: DOG does not rely on a single scalar fit. It classifies evolutionary paths by topological structures: stable order loops, periodic order chains, and escape order trees. This classification has predictive power (e.g., determining whether a system will lock in or escape in the long term).

Pain Point 7: Physical Paradoxes from Infinite Construction

· Fractal: Measuring an infinitely fine coastline would require infinite energy, implying a deep conflict with the Heisenberg uncertainty principle.

· DOG: DOG fundamentally avoids infinity. Axiom 1 explicitly states that all motion states are constrained to a finite set of lattice points. No infinite fineness, hence no energy divergence or measurement paradox.

---

IV. Paradigm Comparison Summary

Dimension Classical Fractal Geometry DOG Discrete Order Geometry

Structural basis Infinite recursion, connected nesting Finite discrete, separable order

Scale treatment Continuous scale‑free Continued fraction hierarchical convergence (finite truncation)

Decision problems Undecidable cases exist Finite state machine, all decidable

Computational mode Exponential explosion, serial dependency Lattice parallel, state transitions

Physical reference Descriptive, approximate fitting Constructive, predictable, enumerable

Paradox risk Infinity‑induced energy/measurement paradox Finite structure, no paradox

---

V. Historical Contribution and Paradigm Boundary of Fractal Geometry

Before proceeding to DOG’s paradigm reset, it is necessary to clarify the historical status of fractal geometry.

The greatness of fractal geometry lies in its systematic introduction of “self‑similarity” and “scale‑freeness” into geometry, revealing the order behind numerous irregular forms in nature. Coastlines, snowflakes, clouds, mountains, leaf veins—objects previously indescribable by Euclidean geometry—acquired mathematical expression under the fractal framework. This is one of the major mathematical breakthroughs of the 20th century.

However, fractal’s core assumptions—infinite recursion, connected nesting—determine its boundary of applicability: it describes continuous, connected, gapless self‑similar structures. For discrete, spatially separated, gapped ordered systems (such as galaxy clusters, satellite systems, biological population distributions), fractal geometry has no direct modeling capability.

This is not a “mistake” of fractals, but a “choice.” Just as Euclidean geometry chose flat space and Riemannian geometry chose curved space, fractal geometry chose connected self‑similarity. Each paradigm has its own domain of applicability. The problem is not with fractals themselves, but with whether later generations mistake its assumptions for universal prerequisites of geometry.

---

VI. Paradigm Reset: Connectedness is a Special Case, Discrete Separated Order is the Norm

Fractal geometry inherited a deep implicit assumption from Euclidean and Riemannian geometry: a geometric system must be based on connected, adjacent, gapless structures. Whether the continuous curve of a coastline or the connected branches of a snowflake, fractals always study structures that are “linked together.”

Yet, from the perspective of the entire cosmos, connected adjacency is a local special case; discrete, spatially separated order is the universal norm:

· Galaxies are separated by vast voids, forming clusters through gravitational order, not physical contact.
· Stars and planets face each other across empty space, forming systems through orbital order.
· Inside an atom, the electron and nucleus are mostly empty space, forming the atom through quantum order.
· Biological population distributions, city clusters, communication networks… in the vast majority of ordered systems, the units are not physically connected.

Why did traditional geometry (including fractals) mistake a special case for the norm?

Because geometry originated from land surveying—a local, continuous, connected Earth‑surface scenario. Euclidean geometry worked well for measuring fields and building houses, so “connectedness” became a tacit prerequisite for geometry. Riemannian geometry inherited this assumption, writing connectedness into the definition of manifolds. Fractal geometry likewise followed this tradition, focusing on connected self‑similar structures.

The result: for two thousand years, geometry has applied “local special case” axioms to the entire universe.

DOG’s three axioms are precisely designed to correct this historical deviation:

1. Axiom of spatial discretization: Does not require continuity, only a finite countable set of order lattice points.
2. Axiom of order priority: Does not require physical connection, only order isomorphism.
3. Axiom of closed recursion: Does not require adjacent evolution, only a recursive closed loop.

DOG does not negate traditional geometry (including fractals). Rather, it resets it: traditional geometry is a “connected special case” subset of the DOG system. DOG expands the boundary of geometric research from the local Earth surface to the entire cosmos, giving discrete, spatially separated ordered structures their first legitimate geometric home.

---

VII. Conclusion

The pain points of classical fractal geometry—whether the contradiction between infinite processes and physical forms, undecidability, dimensional unreliability, or physical paradoxes—all stem from its core assumption of “infinite connected self‑similarity.”

Through a paradigm shift, DOG replaces “infinite recursion” with “finite discrete order,” “iterated function systems” with “state machines,” and “continuous scale‑freeness” with “continued fraction truncation,” resolving these pain points in one stroke.

More importantly, DOG accomplishes a historical reset of geometry:

· Connected adjacency is a special case of local phenomena on the Earth’s surface.
· Discrete, spatially separated order is the universal norm of the cosmos.

DOG is not a patch or improvement on fractal geometry, but an independent, self‑consistent, computable, decidable, and fully new geometric path. It returns traditional geometry (Euclidean, Riemannian, fractal) to their proper place as special cases, while providing the first dedicated geometric paradigm for discretely ordered cosmic structures.

---

References

[1] Mandelbrot B. The Fractal Geometry of Nature. 1982.
[2] Zhang S H. Discrete Order Geometry (DOG): A New Geometric Paradigm Based on Fractal Nesting and Continued Fraction Scaling. 2026.
[3] Zhang S H. A Complete Solution System for the Three‑Body Problem within the Framework of Discrete Order Geometry (DOG). 2026.