Currently the title of a work is truncated to 100 characters before its permanent work ID is calculated. The underlying cause is a fixed-length field in the software for which the algorithm was originally written.

That limitation doesn't apply to us, so we can remove the restriction. This will nearly eliminate the chance that multi-volume books with very long titles will be treated as interchangeable versions of a single work. The downside is that our algorithm will no longer be compatible with the original algorithm.