XOM API for Processing XML with Java

[Wolfgang Hoschek <whoschek AT lbl.gov>]

On Aug 9, 2005, at 4:35 PM, Elliotte Harold wrote:

> Wolfgang Hoschek wrote:
>
> > If one wanted to recreate the exact same XOM document, preserving   
> > additional namespace declarations "as is" is key.
>
> What constitutes the "exact same document" is a matter of some  
> debate. XOM is never going to provide exact handling of white space  
> inside tags, character references, the formatting of the DTD, and  
> so forth; and I don't think you really want it to. Once you've  
> ruled out the idea of byte-per-byte equivalence, I doubt any one  
> model will satisfy everyone perfectly. XOM very closely follows the  
> XPath 1.0 data model, with a few small exceptions. (XOM does  
> provide access to the DOCTYPE declaration that XPath doesn't. It  
> also makes some effort to maintain base URIs.) I agree with Michael  
> Kay that the XPath data model is more sensible than the full  
> Infoset model. However, these are hardly the only two such models;  
> and others are possible.

True, but that's besides the point since I referred to the the "exact  
same XOM document", not the "exact same document". The "exact same  
XOM document" is easy to define: It is the same as "new Document 
(doc)", i.e. a faithful copy; it has identical XOM nodes, values and  
other retrievable properties in the very same node locations as the  
original document (subclasses are considered out of scope).

Wolfgang.