XOM API for Processing XML with Java

[Elliotte Rusty Harold <elharo AT metalab.unc.edu>]

At 4:59 PM -0700 9/29/03, Wolfgang Hoschek wrote:
> In order to have XML docs easily human readable, I'm trying to 
> remove as much namespace clutter as possible from third-party XML 
> documents autogenerated and merged from multiple sources.
>
> Does anyone know rules and/or code that removes redundant namespace 
> declarations in XML documents, either by strictly retaining 
> semantics or by slighly relaxing them?
>
> Possibly by reusing prefixes already declared "further up", by 
> moving common namespace declarations into parent elements, by 
> renaming prefixes that refer to the same namespace, by removing 
> prefixes that refer to the default namespace, etc. This seems to be 
> not so trivial.

This is hard but doable. However, what moves it from hard to not 
completely possible is that you may encounter namespace prefixes in 
element content and attribute values. Then again you may not. (Think 
XSLT or schemas). This severely limits how much generic rewriting you 
can do without a deep knowledge of the document's semantics. This is 
really a flaw in the design of namespaces in XML. :-(
--

  Elliotte Rusty Harold
  elharo AT metalab.unc.edu
  Processing XML with Java (Addison-Wesley, 2002)
  http://www.cafeconleche.org/books/xmljava
  http://www.amazon.com/exec/obidos/ISBN%3D0201771861/cafeaulaitA