Coupled-Resonator Optical Waveguides (CROWs) are chains of resonators in which light propagates by virtue of the coupling between the resonators. The dispersive properties of these waveguides are controllable by the inter-resonator coupling and the geometry of the resonators. If the inter-resonator coupling is weak, light can be engineered to propagate slowly in these structures. The small group velocities possible in CROWs may enable applications in and technologies for optical delay lines, interferometers, buffers, nonlinear optics, and lasers. This thesis reports on achieving and controlling the optical delay in passive and active CROWs. Both theoretical and experimental results are presented. Transfer matrices, tight-binding models, and coupled-mode approaches are developed to analyze and design a variety of coupled resonator systems in the space, frequency, and time domains. Although each analytical method is fundamentally different, in the limit of weak inter-resonator coupling these approaches are consistent with each other. From these formalisms, simple expressions for the delay, loss, bandwidth, and a figure of merit are derived to compare the performance of CROW delay lines. Using a time-domain tight-binding model, we examine the resonant gain enhancement and spontaneous emission noise in amplifying CROWs to find that the net amplification of a propagating wave does not always vary with the group velocity but instead depends on the termination and excitation of the CROW. CROWs in the form of high-order (> 10) weakly coupled passive polymer microring resonators were fabricated and measured. The measured transmission, group delay, and dispersive properties of the CROWs agreed with the theoretical results. Delays in excess of 100 ps and slowing factors of about 25 over bandwidths of about 20 GHz were observed. The main limitation of the passive CROWs was the optical losses. To overcome the losses and to enable electrical integration, we demonstrated active CROWs in the form of current injection InP-InGaAsP Fabry-Perot laser arrays. Even though the losses could be completely compensated, the transmission spectra and signal-to-noise ratio depended strongly on the injection current and resonator position. The signal-to-noise ratio degraded rapidly away from the input. Our results highlight possible avenues to operate laser arrays as loss-compensated or amplifying CROWs.