Note that a strictly greedy algorithm could be confused by this situation and could align, for example,  to  in Figure 1 if the similarity among these two instances of the motif happens to be slightly stronger than the similarity among  and , and among  and , respectively. However, DIALIGN uses a greedy approach only for multiple alignment where an exact solution is not feasible, but for pairwise alignment, the program returns an optimal alignment with respect to the underlying objective function. Thus, under the above assumtion, a meaningful alignment will be produced even if  exhibits stronger similarity to  than to .