Digital Libraries constitute a relatively young scientific field, whose life spans roughly the last ten years. Despite its young age, the field has made a long journey from its initial conception to the present state of the art and has reached a level of maturity that warrants taking stock of its course so far and paving a path for the immediate and long-term future. This talk analyzes the recommendations of several activities that were initiated by the DELOS Network of Excellence, especially of the last DELOS Brainstorming Workshop in July 2004. It identifies the main directions that future work on Digital Libraries should follow, emphasizing the design of systems that are person-centric, collaborative, distributive, generic, and appropriate for a wide variety of applications.