Hi,

The results directory now contains a results table for the
1180 submitted extracts - basic sentence overlap numbers, 
sentence precision and sentence recall - against one or 
two model extracts. Many other measures can of course be 
calculated.

There are also two tar files with the model extracts and
submitted extracts in a condensed form, namely as lists of
document-id.sentence-number strings. 

<http://www-nlpir.nist.gov/projects/duc/duc2002/results/>

- Paul