This paper offers a perspective on the directions in which hypermedia infra
structure research will move in the next several years. The perspective is
based on the authors' experiences and insights from a decade of active part
icipation in this research area. After a review of hypermedia infrastructur
e research, the paper focuses on two particular threads of such research na
med "multiple open services" and "structural computing". We believe that th
ese threads show much promise for the future.