improve parser speed even further

[rkeytacked]

The switch from 1.0.* (isbn-core) to 1.1.* (fast-isbn) already improved the parsing speed from >4M ops/s to >14M ops/s on the reference machine. Theoretically it should be possible to reduce the number of comparisons and table lookups in order to evaluate the correct hyphenation pattern and checksum validity.

A good trade-off between table sizes and lookups might be triples of digits (000 – 999) with lookup tables of size 1000.