From the Dirichlet Integral to Lobachevsky's Formula: A Formalization in Lean 4

In the vast landscape of mathematics, there is a quiet corner dedicated to the study of how things add up over time, particularly when those things wiggle back and forth. This is the realm of real analysis, where mathematicians examine the behavior of functions that change continuously. One of the most famous puzzles in this field involves a specific curve that rises and falls like a wave, getting smaller and smaller as it stretches out toward infinity. The question is simple to state but tricky to solve: if you add up the area under this wiggling curve from the very beginning to the farthest point you can imagine, what total do you get? For over a century, mathematicians have known the answer, but proving it rigorously without making hidden assumptions has always been a delicate task. This is because the curve does not settle down quickly enough for standard rules of addition to apply directly; it relies on a precise cancellation of positive and negative areas to reach a finite sum. Understanding this behavior is crucial not just for pure math, but for the technology that underpins modern communication, as these same wiggling patterns are used to reconstruct signals and images from raw data.

Recently, two researchers, Daniel Goldberg and Antoine Vinciguerra, decided to tackle this classic problem using a computer program designed to check mathematical proofs with absolute certainty. They did not just write down the solution; they built a complete, step-by-step logical argument inside a software system called Lean 4, which acts like a tireless auditor that refuses to accept a single step unless it is justified by the rules of logic. Their goal was to formalize the Dirichlet integral, the name given to the calculation of that specific wiggling area, and to show how it connects to a broader set of rules for integrating periodic functions. The challenge they faced was that the standard way computers handle area calculations, known as the Lebesgue integral, cannot directly process this particular curve because the total size of its wiggles is infinite, even though the net area is finite. To get around this, the researchers had to find a clever detour that avoided the infinite problem while still leading to the correct answer.

Instead of trying to force the computer to accept the original wiggling curve directly, the team first looked at a modified version of it where the curve was squared. This squared version behaves much more nicely; its total area is finite and well-behaved, allowing the computer to calculate it using standard methods. The researchers then proved a specific relationship between the area under the original wiggling curve and the area under this squared version. By calculating the area of the squared curve first, they could mathematically transfer that result back to the original problem. This approach allowed them to bypass the difficulties of conditional convergence, where the order of addition matters, and arrive at the famous result that the total area is exactly half of the number pi. This was not a guess or a simulation; it was a rigorous proof that the limit of the area, as the boundary moves further and further out, converges to this specific value.

Having solved the main puzzle, the team used their new tools to explore what else could be derived from it. They showed how this integral acts as a filter that can turn a smooth, continuous wave into a sharp, step-like jump, a behavior that is fundamental to how digital signals are processed. They also discovered and proved a collection of other identities involving products of these wiggling functions, showing how different frequencies interact when multiplied together. These results are not just abstract curiosities; they provide the mathematical foundation for understanding how signals can be reconstructed from samples, a concept central to the Shannon sampling theorem used in digital audio and image processing. The researchers demonstrated that by understanding the behavior of these specific integrals, one can derive precise formulas for how different wave patterns combine and cancel each other out.

The final and perhaps most surprising achievement of their work was the formalization of a formula discovered by Nikolai Lobachevsky, a mathematician best known for his work on non-Euclidean geometry. Lobachevsky had found a rule that allows the area under a wiggling curve multiplied by a repeating pattern to be calculated by looking at just a small slice of that pattern. The researchers proved that this rule holds true for any continuous, repeating function that has a specific kind of symmetry, using the computer to verify that the infinite sum of the wiggles can be reduced to a simple calculation over a short interval. They achieved this by showing that any such repeating function can be closely approximated by a sum of simple cosine waves, and since the rule works for each individual wave, it must work for the whole function. This provides a machine-checked proof of a general identity that had previously only been understood through human intuition and traditional paper-and-pencil methods.

The work of Goldberg and Vinciguerra demonstrates that even centuries-old mathematical truths can benefit from the precision of modern computer verification. By breaking down the problem into manageable pieces and navigating around the obstacles that confuse standard integration methods, they have created a solid foundation for future research in signal processing and harmonic analysis. Their formalization confirms that the Dirichlet integral is indeed the limit of areas over bounded intervals, and it establishes a reliable framework for Lobachevsky's formula. This achievement suggests that similar rigorous approaches could be applied to more complex versions of these integrals, potentially leading to new insights in how we understand the mathematical structures that govern the physical world. The paper stands as a testament to the power of combining deep mathematical insight with the unforgiving logic of computer verification, turning a classical puzzle into a verified fact.