Atom Optics is a new branch of Atomic Physics, where one tries to act upon atoms, likewise one can act upon light (photons) in ordinary optics. The goal is therefore to reflect, focus, diffract, make interfere atoms1 .Afew years ago, this was considered as a fundamental goal, as had been the case decades ago for electron optics, or neutron optics. Indeed, it is always important to investigate the quantum behaviour of larger and larger objects, and studying the wave like behaviour of atoms has attracted a lot of attention. Moreover, atoms are objects that are coupled with the environment in a controlable way, in particular via quasi-resonant interaction with light. There is thus a hope to investigate how this coupling with the environment is relevant to the old problem of the division between the quantum world and the classical world2 ,and how it is related to the question of quantum decoherence. To keep in the domain of Atom Optics, the question of the preservation of coherence of the de Brogue waves describing the atomic motion is crucial. Another basic issue is the question of the collisions between atoms, or between an atom and a surface, when the atom is described as a matter-wave. From this point of view, the very interesting recent results in the domain of cold atoms collisions are only a part of what will probably come out, and new approaches may be fruitful. For instance, some collision effects may be considered as non-linear effects of atom optics... On the other hand, almost from the start, Atom Optics has been a domain where the question of applications has been raised. The experience of the many useful applications existing in the domain of electron and neutron optics has clearly shown the way. A first reason for expecting interesting applications to atom optics is the small value of the de Brogue wavelength, in the nanometer range or below, even at very low energies. This opens for instance the possibility of high resolution lithography, or of nanoprobes, at energies low enough to have a soft, non damaging interaction with a surface. Another important class of applications is in atomic interferometry. It can be shown that atomic interferometers (using atoms of mass Mat) are much more sensitive to inertial or gravitational effects3 than photonic interferometers (using photon of energy hw), by a ratio of the order of MatC2 / ho) . Note that this statement is true only under the condtion that the atomic and photonic interferometers under comparison have equivalent geometries, so that present day interferometers do not yet take full advantage of the 1011 factor of potential improvement. However, impressive results have already be demonstrated6. As a matter of fact, many of the papers presented in this symposium are linked to applications, either with atomic interferometers, or related to the possibility of deep focusing of an atomic beams. Most of these applications a Unite de recherche associée au CNRS can be easily understood by analogy with photon optics. However, there are also several features of atom optics that have no analogy in photon optics. The most important one is probably the possibility to increase the luminosity of an atomic beam (i.e. the flux of atoms per per unit time, surface, and solid angle) or —equivalently — to increase the atomic density in the phase space. This is possible by use of laser cooling, which is a dissipative process, and therefore is not constrained by the Liouville theorem. Another distinctive feature of atom optics —this one not particularly an advantage —is the lack of a sudden interface, that would be analogous to a glass-vacuum interface for light. By sudden interface, we mean that some characteristic parameter relevant to the propagation (the index of refraction in the case of light) changes on a scale small compared to the wavelength of the waves that propagate. In atom optics, the interfaces are usually soft, since the potentials acting on the atomic motion usually change on a scale large compared to the de Broglie wavelengths. In this overview, I intend to recall some basic principles of atom optics, and to point out some important analogies and differences with photon optics4 .Iexpect this discussion to be illustrated by the various presentations at this symposium, for which I hope to offer a framework. In addition, I will take the liberty to illustrate my presentation by some examples chosen among results obtained recently in my laboratory, on the subject of atomic mirrors. The comparison between atom optics and photon optics would of course be incomplete if I would not adress the question of the impact of Bose Einstein condensates onto Atom Optics. As many of us, I am convinced that we are assisting to a revolution as important for Atom Optics as the invention of the laser in the domain of Photon Optics, and we are lucky to have a full session on this hot subject.