Roth's theorem is a landmark result in number theory saying that algebraic numbers cannot be approximated by rationals "too well." Over the decades, mathematicians have generalized this idea in many directions, including to settings called adelic curves, which are abstract frameworks that treat number fields and function fields in a unified way. The central question becomes: how many rational points can simultaneously satisfy a whole system of such approximation conditions? This paper gives an explicit upper bound on that count, meaning a concrete formula rather than a proof that the count is merely finite.
Earlier work on this problem came with awkward restrictions, most notably a requirement that solutions had to have "large height," a technical measure of arithmetic complexity. This paper eliminates that restriction entirely, so the bound applies to all solutions without exception. The authors also relax the conditions on the parameters appearing in the inequalities, making the result applicable in a wider range of situations.
The other main contribution is a new, constructive proof of a result originally due to Vojta, Dolce, and Zucconi, which produces a certain auxiliary system of inequalities needed in the argument. The original proof used a compactness argument borrowed from topology, which is elegant but non-constructive and gives no concrete numbers. The authors replace this with a combinatorial approach based on the pigeonhole principle and a finite grid partition. The payoff is that every constant and parameter in the final bound is now explicit and trackable, which is valuable both for theoretical purposes and for potential computational applications.