XSLT 2.0 and XPath 2.0 Programmer's Reference, 4th Edition (345 page)

The real driver for introducing these new operators was not really the requirements of XSLT users, but the needs of XQuery, which is a superset of XPath 2.0. XQuery needs to be able to search large databases, and if you want to search a terabyte of data then you need to take advantages of indexes. This means you need to be able to rearrange the query as written by the user into a form that can take advantage of the indexes known to be available, and this rewriting of an expression into a different form is only possible if the operators have very clean mathematical properties. For example, a very useful property that makes rearranging expressions possible is called
transitivity
, which means that if
A=B
and
B=C
are both true, then you know that
A=C
will also be true. Unfortunately, this isn't the case for the
=
operator in XPath 1.0. For example, in XPath 1.0,
1=true()
and
true()=“true”
are both true, but
1=“true”