Order parametr for design of proteinlike heteropolymers

preprint withdrawn. A revised version of this paper, with different authors appears in E. Nelson, P. Wolynes, and J. Onuchic, in Optimization in computational chemistry and molecular biology, C. Floudas and P. Pardalos editors, (1999)). The main failing of my approach in these papers is the fact that the Hamiltonian and order parameters are based on pair distances only, and therefore do not break the local gauge (reflection) symmetry. Consequently, the order parameters cannot detect the difference between (for example) a partially compact topology and its mirror image. I take full responsibility for this problem, the withdrawal of this preprint, and the comments made above. Erik D. Nelson