Port from JDOM to iDOM

Quick Method Mapping

JDOM	iDOM
Element.getName	Element.getName JDOM use the local name, while iDOM uses the tag name.
new SAXBuilder()	new SAXBuilder(true, false)
Element.getContent, getChildren	Element.children JDOM have a 'seperate' list of elements called children and all child items called content. In iDOM, there is only one list called children. Also, String is no longer used as a child of Element. Instead, use Text. Notice that the content of an element is a unique feature of iDOM that is nothing to do with JDOM's content.
Element.getChildren(String name...)	Element.getElements(...)
Element.getChildText, getChildTextTrim	Element.getElementValue(.., boolean trim)

Major Differences

Supports W3C DOM. Each iDOM tree is also a W3C/DOM tree. Thus, all utilities of W3C/DOM is appliable to iDOM.
The Item and Group interfaces are introduced to make API more consistent, and it also simplify utilities that handle iDOM, because they can be abstracted by Item (rather than dirty switch).
getName returns tag name rather than local name. Similarily, constructors requires tag name as argument. Reason: to be compatible with W3C/DOM (i.e., Attr.getName).
Using Text, rather than String -- to be compatible with W3C/DOM.
According to spec (and like crimson), we put the expanding result under EntityReference if SAXBuilder.isExpandEntityReferences is false. Also, we the standard entity references, e.g., ">", are also kept. Moreover, EntityReference doesn't contain public/system Id to have the code running faster.
Support read-only.
Each item contains a Locator object, which simplifies error reporting.
Powerful methods (getElements, getAttributes...) to search by name, regular expression...
API of Text, CDATA and Comment are made more consistent by extending the same interface, Textual.
A new class, Binary, is introduced to store any kind of Object, not just String.
Document now accepts EntityReference.
Attribute supplies more constructors with string-type namespace like Element.
getDocument now becomes depending the root, rather than "spreading aroun".
DocType is now a child of Document. It makes it easily to be compatbile with Node.
Power searching methods for children and attributes.
Better implementation to reuse codes thru Item and others. Also, better performance (though not measured yet) to use HashMap for children.