This paper addresses the optimality of the fixed delay tree search with decision feedback (FDTSIDF) detector for data without a minimum run length constraint. The evaluation is performed by comparing the minimum distance for FDTSIDF with the minimum distance associated with a maximum likelihood sequence detector (MLSD). Using this distance information, potential improvements through coding areaddressed.An alternative realization to the direct implementation of FDTS/DF is considered by examining the detection process in the context of a signal space.