The scoring software in this tarball includes two Perl programs.
One program scores a single output file. The other program reads
the Test Definition file and scores all the “samples," by repeatedly
calling the program that scores a single file.

There is a set of mock system output files and a corresponding set
of mock “answer key” files that they can be scored against.  One
should be able to score those mock files as follows:

    scoreDataset_SAD.pl -d fake_testDef.xml

If you wish to score a single file, the command looks like

    scoreFile_SAD.pl -r F/54321_67890_eng_F_annot.txt -h F/54321_67890_eng_F.txt -s 2 -e 3 -g 4 -t 5 -f 6 -u 7

Note that both Perl programs begin with a line that read
#!/usr/bin/perl
which you may need to modify to point to the location of
your Perl installation.


