An area of physics exhibiting rapid progress and strongly focused interest is that of extremely fast phenomena; both the development of suitable methods and the understanding of such processes are at issue. While the term ultrafast was originally used for processes occurring in the (first upper, then lower) femtosecond range, today timescales of 1 fs or fractions thereof are accessible. This does not represent a marginal improvement, but encompasses a qualitative change: from the range where in principle the Born-Oppenheimer approximation, albeit with diabatic extensions, can be applied, to that of true nonadiabatic response, where the motion of individual electrons becomes the focus of attention. Therefore, the opportunity now exists to investigate the motions of electrons within many-electron systems, including the development of coupled many-body responses like the formation of band structure and the buildup of screening responses. First attempts to push into this range by looking at, for example, the ultrafast charge transfer at surfaces have been made successfully in the last 15 years with the core hole clock method. However, the latest developments utilizing attosecond laser pulses promise to provide a general tool to investigate in detail, and with few restrictions, this realm of electronic motions. While significant methodological development remains necessary for complete utilization and interpretation of measurements made on the attosecond timescale, the results obtained so far are extremely promising. It is timely, therefore, to give an overview of some past approaches, the most recent developments, and to identify the types of problems that should be attacked in the future using this exciting new tool. We will focus in particular on processes occurring at surfaces, because they contain in nuce the ingredients operative in atoms and molecules together with the important effects of solids.