nemja

> the question "Why should we parse kata-analyze differently from lz-analyze? Isn't this still by far the smallest problem? Score estimates are only reported in the former, so some differences...