(e) We adapt our previous scheme for copying subsegments to incorporate
a second permutation τ of the same values {0, 1, 2, . . . , l−1}. For each
segment, we generate a random pair of permutations σ and τ ; as before,
we use σ to choose the order of indices of subsegments to be copied from
the segment of SRC (as before); we then further apply the permutation τ to
this subsegment index to obtain the subsegment copied to in DEST. That
is, the new scheme is to copy first subsegment σ(0) in SRC to subsegment
τ (σ(0)) in DEST, then subsegment σ(1) to τ (σ(0)), and so on.