We develop a fully rigorous geometric-topological framework for analysing finite truncations of decision problems and complexity classes. Decision procedures are represented extensionally as Boolean functions restricted to bounded input length and embedded canonically into high-dimensional Hamming cubes. The framework provides a mathematical basis for controlled empirical exploration while avoiding unjustified asymptotic extrapolation.
Julius Wittig (2026) studied this question.