Transportation safety is a highly contentious issue in the design of cities and communities. To enhance community livability, urban designers, architects and city planners often encourage the placement of street trees, aesthetic street lights, and other roadside features in a buffer zone between the pedestrian realm and the vehicle travelway. While such designs clearly enhance the aesthetic quality of a roadway, conventional geometric design practice regards roadside features located in the clear zone as fixed-object hazards, and strongly discourages their use. This study examines roadside safety in urban environments to better understand the nature of urban fixed-object crashes, as well as the safety impacts of livable streetscape treatments. While the prevailing assumption is that livable street treatments have a negative impact on a roadways safety performance, the existing empirical evidence indicates that such designs are much safer than more conventional roadside designs. Current safety objections to the use of livable street treatments are not based on empirical evidence, but are instead the result of a design philosophy that systematically overlooks the real-world operating behavior of road users. This study details the origin and evolution of this philosophy, termed passive safety, and subjects it to an empirical test to evaluate its applicability to urban arterial roadways. It finds that passive safety assumptions do not meaningfully explain empirical observations of crash frequency and severity. To enhance contemporary geometric design practice, this study then proceeds to more thoroughly examine the nature and characteristics of urban roadside crashes, and proposes a new design approach, termed positive design that better addresses the twin goals of safety and livability.