1. Explore what kind of sequences are annotated as VP in the CONLL2000 "train" corpus data.

2. Develop a chunk.Regexp grammar to capture the regularities.

3. Use the trace=1 setting of the chunk parser to examine the success of your VP chunking rules.

4. Once you are reasonably happy with your rules, try evaluating your them against the CONLL2000 "test" corpus data (i.e., using the chunk.accuracy() function).

5. Briefly comment on how easy or difficult it was to develop an adequate rule set.