Program analysis has long relied on logical frameworks that capture complementary yet partial aspects of program behavior. For example, Hoare Logic formalizes partial correctness; Incorrectness Logic characterizes reachable error states; and Sufficient Incorrectness Logic pinpoints causes of bugs without false alarms. Used in isolation, these logics illuminate different aspects but yield an incomplete view. In this ongoing work, we outline the principles of a unified program logic exploiting the orthogonal, similar, and complementary features of each logic to reduce current fragmentation and improve the precision of the analysis. Central to our approach is a new notion of triples that simultaneously express over- and under-approximations in both the forward and backward directions at the same time, avoiding a commitment to a single reasoning direction. We show that our formalism subsumes existing logics and paves the way for new interesting logics for program analysis.