Files beginning with	are
--------------------	---

EN			the English text of the gold standard
FR			the French text of the gold standard
frq			words in the focus set, by frequency
A			directories for the data of each annotator


Within the A?/ directories, files are named as follows:

part[12] refers to parts 1 and 2 of the gold standard, where part 1 is
verse pairs 1 to 100 and part 2 is verse pairs 101 to 250

.jnorm .lnorm and .rnorm refer to the 3 kinds of link normalizations:
joint, left and right.  "left" generally refers to the English side
and "right" to the French side.  .lnorm means that the links are
normalized so that the weights of the links attached to any word on
the English side sum to less than or equal to 1.  Likewise for .rnorm
and French.  .jnorm means the weights of links on both sides were
normalized to sum to less than or equal to 1.

.open means links involving a closed-class word on either side were
ignored.

.complete means the complete ordered set of links for either part1 or
part2 of the gold standard.  The 2nd column in these files specifies
the verse pair number; the 1st column specifies link weight.

.sub files are subsets of the .complete files, generated for the
purpose of computing inter-annotator agreements rates and their
standard deviations (see the paper).

The files samp*.SentPair?.* contain the links for particular verse
pairs, where the verse pair number = (samp# - 1) * 10 + SentPair# + 1.