Next generation telescopes will be all digital instruments. This fact will allow the full integration of archives into telescopes operation. This is a main step if one wants to fulfill the requirement of usefulness of both scientific and technical data contained in the archives. A great change with respect to the past is required involving the modes of scheduling, observing, calibrating etc. New models of approaching the whole scientific data flow starting from the archiving problem are studied in several Institutions. These new models of observation, whose consequences in terms of human factor seem to be as important as the technological implications, promise to ensure a high and constant quality level of the data to be archived. The current status of this topic is reviewed. Some thoughts on the data model presently used in astronomy are also briefly given.