
Okay... Take 2 for design of the XPath parser stuff.

Main public interface points are:

	org.jdom.contrib.xpath.Context
		// Knows how to walk a ?DOM tree for certain
		// small atomic actions.

	org.jdom.contrib.xpath.XPathExpr
		// Knows all about small atomic actions, and
		// requests the Context to perform them.

They will work together to walk a DOM/JDOM tree and Do The Right Thing.

Presumable, there will be a JDOMContext which implements Context.

The org.jdom.contrib.xpath.parser.* stuff creates an XPathExpr
from the classes in org.jdom.contrib.xpath.impl.*.

Still much work to do, but we're closer to having an object-tree
from an textual string xpath.

