XOM API for Processing XML with Java

[Wolfgang Hoschek <whoschek AT lbl.gov>]

> Hallvard Trætteberg wrote:
>
> > If you're only interested in img elements, you needn't build a full
> > document and traverse it. Using Jaxen/XPath is definitely overkill for
> > this case. XOM's node factory (don't remember the class name) supports
> > creating only the nodes you're really interested in, to save space and
> > time. Implement a custom NodeFactory (?) that only creates and Element
> > if the name is img.
>
>
> Now why didn't I think of that? You wouldn't even have to build most of
> the document, and the parse would do the traversal for you. This should
> be much faster and more efficient than using XPath.
>
> --
> Elliotte

And if one would have the document already in memory (e.g. because it 
has been given to us from some other library) and wanted to do such 
node factory filtering, one might want a way to pipe a document through 
a builder, using the appropriate node factory. BTW, it appears that 
this is the same underlying mechanism that would also enable in-memory 
validation in a correct, convenient, effective way, requiring minimum 
API complexity.