The FEL experiment in the visible and near UV on the Super ACO storage ring has given, since 1989, important informations on the SRFEL dynamics and, furthermore, a very good beam stability has been achieved. In addition, the operation at 350 nm with this good stability and a long beam lifetime allowed us to perform the first user experiment in biology and to start with a campaign for using the laser as photons source for experiments in other domains, coupling FEL light and the Synchrotron Radiation. For this, FEL starts to be very competitive with respect to the other conventional laser sources, provided that it could oscillate further in the UV, say at 300 nm and below. So, the real challenge is now given by the lasing at shorter wavelengths and, for this, by the optical technology existing nowadays. Since 1992 the efforts have been concentrating to look for every kind of solution allowing us to overcome the problem of having a very low gain. From an optical point of view, in the range of wavelengths explored, there is a lack of transparents dielectric materials for substrates and coatings. Substrates are required at the same time to be relatively notmore » absorbing (a few tens 10{sup -6}), to have a very good surface quality (RMS roughness below 10 {Angstrom}) because of scattering losses dramatically increasing in this spectral range and, due to the thermal load of the undulator emission, to have adequate thermal characteristics. In order to fulfill all these requirements, a good characterisation and modelisation of the substrates is needed, especially to correlate thermal loading and mechanical deformations from one hand, and roughness and scattering losses from the other hand. Coatings must be not absorbing too and, above all, the most amorphous as possible (this could be obtained with IBS deposition technique), in order to insure a good reproduction of the substrate roughness at the interfaces and on the top layer and an higher resistance to the XUV photons load.« less