XOM API for Processing XML with Java

["Bradley S. Huffman" <hip AT cs.okstate.edu>]

Elliotte Harold writes:

> Michael Kay wrote:
> 
> > If the user controls how many text node objects there are, and if XOM 
> > gives
> > different results depending how many there are, then I'm happy for XPath 
> > to
> > do likewise. I think this is better than ./text()[contains(., "o")]
> > returning 4 text nodes of which only one contains an "o".
> 
> But it doesn't return 4 text nodes. It returns 4 Text objects, which 
> together make up 1 text node. That's a crucial distinction.

Hmmm, what if the context is "makes ", "a", " good", " point.", 
<someElement/>,
"  Yes", " he", " does!". Now don't you have the expression returning 7 Text
objects representing 2 text nodes? But since all Text objects have the
same parent you can't distinguish which ones belong to which text node.

Brad