Hi,

The documents for DUC 2002 have been chosen and the creation of
ABSTRACTS
has started. Now we are in the process of preparing materials for the 
creation of EXTRACTS: finding the sentence breaks and adding numbering.

Although we have talked about the extracts as sentence extracts, there
will be sentence fragments in the documents - dates, headlines, bylines,
locations, items from lists, etc. which might be useful information that
a human would include in an extract. 

So, unless someone presents a very strong argument to the contrary, we
plan to allow the human extract creators to include tagged sentence
fragments as well as full sentences in the extracts they will be
creating.
The automatic summarization systems would then be allowed to do the
same.
Scoring would not distinguish between fragments and non-fragments,
anymore
than it will distinguish between long sentences and short.

Comments? Strong objections?

- Paul

-- 
Paul Over - Retrieval Group
            Information Access Division
            Information Technology Laboratory
            National Institute of Standards and Technology
            Bldg. 225  Rm. A211  (Mailstop 8940)
            Gaithersburg, MD  20899-8940   USA
            Voice: 301 975-6784    Fax: 301 975-5287