[kjyfxyvs] burning a book from both ends

error correcting code error model: errors are more likely to occur at the start and end of a string than in the middle.

important metadata, for example, the length of a string or a cryptographic signature, is often stored at the start or end of a string.  it is especially important to detect or correct errors in these locations.  errors at these locations might be bursty: suffix or prefix truncation.

let the original string have length 2*n.  interleave [1 .. 2*n] with [n+1 .. 2*n, 1 .. n].  this results in a string of length 4*n, so redundancy ratio 2.  this achieves error detection but not error correction.

as an aesthetic choice, reverse the latter string before interleaving, i e., [n .. 1, 2*n .. n+1].

this might be an interesting way to print a book.  print the interleaved pages upside down.