Implementing DOM Interface with existing Java classes

Question

I had planned on using some tree-like Java classes as a Document Object Model, which would give me access to all sorts of XML and DOM tools like parsers and XSLT transformers. Initially, I thought all that would be neccessary is to implement all the DOM Interfaces in org.w3c.dom and then I would have a set of classes that conformed to DOM Level 1. It was my understanding that interfaces such as DOMImplementation and Document would interface with various XML tools, allowing creation of a class that implements Document and then Document would have its various factory methods that know how to create the various DOM nodes such as Element, Attr, Text, NamedNodeMap, NodeList, etc.
The problem I'm seeing now is that the JAXP specification (which is what the latest Xerces and Xalan tools conform to) has something called a DocumentBuilder and DocumentBuilderFactory that appear to be necessary to tell the framework what type of class to instantiate that implements the Document DOM interface. Those appear to have a lot of methods that deal with parsing of XML documents and I didn't really want to write or even subclass any existing Parsers in order to get the functionality of traversing and transforming a set of classes that implement the DOM interface.
Am I missing something here? Is it possible to plug in any (set of classes for) DOMImplementation and get them to work with the various DOM and XML tools out there?
Is there an easier way to allow parts of an application access to internal data structures but have the more generic tools or APIs, such as XSL transformers, access that same set of classes as a DOM with the generic DOM interface methods?
Can someone provide me with some guidance here? I'm in the process of finalizing some design on a system and need to know if this is possible or if I need to alter my design.
Thanks.

Answer 1

If I understand you correctly, I think I am working on a similar issue. I am unhappy with the methods given by the DOM for retrieving data from the XML file and for building a file. Our software has a bunch of code that uses these classes and it is extremely ugly. My solution was to create a facade on top of the DOM model. Essentially I have some simple classes that store all the pertinent info (for me) about the XML structure. Essentially that is the element or attribute name, its values and in the case of the element, it's children. This makes it easier for me to build and retreive the data. What I then built was a loader class and a builder class. The loader takes an XML file and parses it in using the DOM classes and builds a structure using my classes and returns the root element. The builder takes a root element and creates a DOM object out of it. This frees me of having to code around the DOM classes all over the place and makes it simple to upgrade our XML code if the DOM changes or a better DOM is released. I am using factories to facilitate this and allow me to have loaders for specific types of XML documents so that I can have a class for each which further simplifies the XML related tasks of other developers on my team.