The factors that determine the stability of globular proteins will be briefly discussed. Both secondary and tertiary structure formation can be described as a balance between nonpolar interactions that favor the formation of compact nonspecific states and the requirement for hydrogen bond formation which can be viewed as a structural constraint imposed by the high free energy cost associated with burying unsatisfied hydrogen bonding groups. Many observed structural features of proteins can now be predicted quantitatively from first principles. This development points to a new generation of fold recognition scoring functions and new ways of organizing proteins into structural families. Given a three dimensional structure it should be increasingly possible to deduce function based on similarities in the surface properties of proteins with related functions. An example of this approach is provided by electrostatic homology relationships which have become evident from applications of the GRASP program.