TEXT

Consider a biobliography database, comprising under a bib node, several book
nodes, and under each book node, the content of the book, organized by
chapter, section, etc.  I am interested in passages having to with XML.  So I
define a scoring function that counts the number of times the term "XML"
occurs under the element in question, and returns the highest ranked
elements.
What is wrong with this scheme?

Is ftscore a special case of ftcontains in TeXQuery?

I am interested in titles of books by author Smith published in year 2004.
I write an XQuery expression for this as follows:
FOR $b in doc("bib.xml")//book
WHERE $b/author/name="Smith"
AND $b/year=2004
RETURN $b/title

Suppose that my database schema has sub-elements first-name and last-name
under name.  (So the correct predicate should have been
$b/author/name/last-name="Smith").  How would Schema-Free XQuery handle this
case? 

Suppose instead that element author has no sub-elements, and just the
author's name as content.  (So the correct predicate should have been
$b/author="Smith").  How would Schema-Free XQuery handle this case?

Is there a better way of writing the XQuery expression above so that Schema-Free XQuery behaves better?

What is one trivial way to create a system with perfect recall?

Draw a suffix tree for a database with the following words:
waikiki, kick, air.