Changes in version 0.24:

Character recognition has been improved.
(merged 'LZ', 'TZ' and 'rt', '9' vs 'g', 'A' vs 'R', 'C' vs '(', 'G',
'J', 'L', 's', 'x', 'z').

Several filters can be now applied in sequence using more than one
"--filter" option.

The new filters "same_height", "upper_num" and "upper_num_only", have
been added.

The description of "OCRAD_result_blocks" in the manual has been fixed.

The license has been changed to GPL version 2 or later for better GPL
compatibility. GPLv2 and GPLv2+ programs may use the library now.
