Electron-proton colliders with center of mass energies between 14 GeV and 30 GeV and luminosities at the 10(sup 33) level have been proposed recently as a means for studying hadronic structure (1). Electron beam polarization appears to be crucial for the majority of experiments. Two accelerator design scenarios have been examined in detail: colliding rings (2) and recirculating linac-on-ring (3). Although the linac-on-ring scenario is not as well understood as the ring-ring scenario, comparable luminosities appear feasible, while the linac-on-ring option presents a significant advantage with spin manipulations. Rf power and beam dump requirements make the linac-on-ring option viable only if the electron linac recovers the beam energy, a technology demonstrated at Jefferson Lab's IR FEL, with cw current up to 5 mA and beam energy up to 50 MeV (4). We begin with a brief overview of the Jefferson Lab energy recovery FEL and summarize the benefits of energy recovery. The feasibility of an energy recovery linac-ring collider is investigated and two conceptual point designs are shown. Luminosity projections for the linac-ring scenario based on fundamental limitations are presented. Accelerator physics issues are discussed and we conclude with a list of required R and D for the realization of such a design.