Changes from NSP version 0.69 to 0.71
-------------------------------------
Ted Pedersen, tpederse@d.umn.edu
University of Minnesota, Duluth
June 16, 2004
Makefile.PL
------------
Updated Makefile.PL with correct version number (Ted)
Makefile.PL now creates /man/man3 directory - in previous
versions this was not created, sometimes causing an error
during "make install" when that directory is not found (Ted)
Programs
---------
statistic.pl and sort-bigrams.pl were modified by changing
the internal N-gram separating marker # to sequence <||>.
As some users pointed out, use of # as a separator between
the N-grams causes problems when N-gram tokens include #.
In 0.71, we used a more rare sequence <||> to separate
N-grams in the internal hash structure. (Amruta)
Testing
-------
Added test cases to catch an error when internal delimiter <||>
is detected within N-gram tokens.
(Amruta)
Fixed test cases for huge-count, that were relying on the mistaken
idea that count.pl produces output in sorted form. Now an actual
sort is done of count.pl output before comparison with gold
standard results. (Amruta)