The paper explores a surprising fact in computer science and mathematics: some infinite collections of objects can be completely searched by a computer in a finite amount of time. This seems impossible at first, since searching an infinite set naively would take forever, but the trick lies in how the infinite set is structured. The authors use ideas from topology, the branch of mathematics concerned with notions of nearness and continuity, to build and study many such "searchable" (or compact) infinite types. They then use ordinals, which are a way of counting and ordering things that extends beyond the ordinary whole numbers into the infinite, to measure how complicated these searchable types are.
A central goal is to find ordinal systems where a single description simultaneously captures two versions of an ordinal: a "discrete" one, useful for familiar counting and comparison, and a "compact" one, useful for exhaustive searching. The paper studies two such systems. The first uses classical Brouwer codes, a well-known way of representing countable ordinals through tree-like structures. The second is a more general, self-referential system built using a technique called inductive-recursion. In both cases, the authors analyze which points in the ordinal behave like isolated islands and which are limit points, meaning they are approached by infinitely many other points, an idea borrowed directly from topology.
A key finding is that two very natural and desirable properties of ordinals, the ability to always compare any two elements (trichotomy) and the ability to always find a least element in any decidable subset, cannot both be achieved at the same time in a "constructive" mathematical setting, where proofs must provide explicit computational witnesses. The paper also establishes that ordinals built from Brouwer codes satisfy a principle related to distinguishability of points, connecting to the topological notion of total separatedness. All results are formally verified using the Agda proof assistant, extending earlier work from a simpler formal system to a richer and more modern foundation for mathematics.