Anne L'Huillier, at the Commissariat à l'Énergie Atomique in Saclay (France), observed in 1987 that passing infrared laser light through a noble gas generates harmonics of much higher order than expected from conventional nonlinear optics — a phenomenon she characterizes experimentally over the following years, but which lacks a theoretical explanation capable of controlling or predicting its behavior. In 1993, Paul Corkum, at Canada's National Research Council, fills that gap with the semiclassical 'three-step' (or recollision) model: the electron ionizes via tunneling, is accelerated by the laser field, and, as the field reverses, recollides with the ion it came from, emitting the high-order harmonic. The model precisely explains why the harmonic spectrum cuts off at a determined energy and becomes the theoretical basis of attosecond physics, predicting exactly which combination of wavelength and laser intensity allows the generation of ultrashort pulses. Building on this experimental and theoretical foundation, Pierre Agostini, at the CEA in Saclay, succeeds in 2001 in producing and measuring for the first time a train of attosecond light pulses (10⁻¹⁸ seconds). That same year, Ferenc Krausz, at the Vienna University of Technology, manages to isolate and measure a single attosecond pulse — rather than a train — opening the way to experiments with greater temporal precision. None of the three pieces would have sufficed on its own: L'Huillier's phenomenon was unusable without Corkum's model explaining how to control it, Corkum's model remained theory on paper without the experimental demonstration by Agostini and Krausz, and that demonstration would not have existed without the two prior foundations. Together they give rise to attosecond physics, allowing, for the first time in history, the motion of electrons within atoms and molecules to be photographed — processes previously too fast to observe directly — with potential applications in ultrafast electronics, medical diagnostics, and atomic-scale materials characterization.