nHow to formulate such a classification scheme?
n
1.How does such a classification relate to previous work?
2.How are web queries distributed given this scheme?
3.How well do human subjects classify against this scheme?
n
nJust briefly: how to automate such a classification?