Status of XML library project

The XML library project is stalled due to lack of free time on my  
part, and the following problem:

As a matter of least-authority and consistent value semantics, XML  
fragments gotten from documents must not provide access to their  
parent nodes. For wrapping DOM, we simply don't provide that  
operation; however, attempting to borrow the Java XPath implementation  
means that we have to actually copy the DOM node so that it doesn't  
have a parent. This means that getting a subtree, or the first time  
XPath is used on it, is O(N).

I thought I could work around this by building my own XML tree objects  
instead of wrapping DOM, and providing a DOM-emulation layer for the  
XPath implementation to use, but it seems that the XPath  
implementation expects to see its own known DOM node type (this is  
poorly documented in the Javadoc), not just anything implementing the  
interfaces.

My current plan is to stick with the DOM-wrapping and the O(N) copy  
cost for now (this exists now, but we could make the copy deferred  
until XPath is invoked), and later look into the feasibility of  
reimplementing XPath.

-- 
Kevin Reid                                  <http://switchb.org/kpreid/>