XOM API for Processing XML with Java

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

At 3:10 PM -0500 3/26/04, jcowan AT reutershealth.com wrote:

> I think it is too confusing to have one rule for Nodes built by Builder
> and another for Nodes created directly by the application.  One rule
> should apply to all.

But that rule could be, "Nodes have the base URI of the document they 
come from, which is null for nodes that don't come from a document."

> It would be good to add a clearBaseURI() method to remove the base URI.
> I'm not sure what this does if the node has a parent.

It seems unnecessary. One can just call setBaseURI(null) if you 
really need to do this.

> I'm not sure how I/O interacts with the base URI.  Presumably, Builder
> should set the base URI whenever an xml:base attribute is found.

No, the base URI is the one the document actually comes from. This is 
necessary, for example, to resolve relative base URIs encountered in 
xml:base attributes. The public getBaseURI() method takes account of 
xml:base attributes and ancestor elements. However, setBaseURI() 
merely sets the element's actual base URI. The two methods are *not* 
perfectly symmetric because an xml:base attribute, including one on 
ancestor element, can override the element's actual base URI.

>  Is the
> attribute left in the attribute list or removed?

It's left in.

> Likewise, when does
> Serializer insert xml:base attributes?

When preserveBaseURI is true, or when there's an xml:base attribute 
in the element's attribute list.

Hmm, I am thinking that it probably makes sense that the result of 
calling getBaseURI() not change just because an element is detached. 
Therefore, it probably does make sense to set the base URI of all 
detached elements, whether they were initially created with a null 
URI or not. But should this be set to the base URI (that includes the 
result of xml:base) or just to the actual base URI? The base URI 
could no longer be in scope when the element is detached.

This is definitely one of the flakier areas of the XML spec. I don't 
think there's a clear one right answer. At some point I'll just have 
to pick one. But I am leaning toward retaining base URIs on 
detachment of all nodes; or at least I am at 5:26 on Friday. Ask me 
again in half an hour, you may get a different answer. :-)
--

  Elliotte Rusty Harold
  elharo AT metalab.unc.edu
  Effective XML (Addison-Wesley, 2003)
  http://www.cafeconleche.org/books/effectivexml
  http://www.amazon.com/exec/obidos/ISBN%3D0321150406/ref%3Dnosim/cafeaulaitA