XOM API for Processing XML with Java

[Laurent Bihanic <laurent.bihanic AT atosorigin.com>]

Neat! The Node interface is a *huge* step forward compared to JDOM and not 
relying on Iterators makes the implementation code a lot simpler.

Yet there are some issues XOM does not address (in addition to those listed on 
the site) like:

 - ClassLoader issues: Type-safe enumerations, such as Attribute.Type, do not 
fit well with class loaders, even when implementing equals() (see 
http://www.javaworld.com/javatips/jw-javatip122_p.html).

 - Code compactness: As I already said on the jdom-interest mailing list, 
there are cases (e.g. TreeNode.getChild()) where instanceof tests make the 
code very unclear and thus harder to maintain.
I know Elliotte does not like the idea of a int getNodeType() method but we 
need to find a way to remove these ugly tests.

 - CDATA handling: There are cases where I need a tool to edit XML documents 
that are also edited manually. If a user used a CDATA section, I'd probably 
like to output the document with a CDATA section. Wouldn't 
Text.setCDATA(boolean) / isCDATA()  be acceptable?

 - Document portions: JDOM uses Lists to "model" document portions. Wouldn't 
XOM need a DocumentFragment node class, for example as a receptacle for XSLT 
result trees?

 - Automatic detach: Having to call detach when moving nodes from one place 
to another is realy cumbersome. A typical example is sorting a list of 
children. Why not support an automatic-detach mode so that programmer can move 
nodes freely?

 - Tree walking: XOM should provide classes to ease walking document trees 
(TreeWalker, Visitor...).

 - HTML Serializer, as more and more people use XML input and XSLT to 
generate HTML pages.

Other small remarks:
- Element.getNamespacePrefix() vs Attribute.getPrefix(): to be made 
consistent. getNamespacePrefix is clearer.
- Element.add(): Why not addAttribute() as every other attribute- and 
child-related method names have suffixes (get/read/removeAttribute, 
append/insertChild, ...).
- Node.howManyChildren(): getChildCount() would have been more intuitive to 
Java programmers.

My best wishes of success to XOM!

Laurent