Intuitively, keyword searches in text documents
should find the documents that contain all the key-
words. The results should contain related meaningful
information. Similarly, for relational databases, the
results should contain the interrelated tuples connected
by the primary foreign key relationships and contain-
ing all or some of the keywords. The results could
contain other relevant, meaningful data besides the
elements that contain some keywords in the tuples.
However, for XML documents, the subtrees connected
by the content nodes that directly contain some key-
words are not easily retrieved. Accordingly, it is much