XOM API for Processing XML with Java

["O'Brien-Strain, Eamonn" <eob AT exch.hpl.hp.com>]

Thanks Elliotte for all your feedback.  It is very useful.  I will reply to 
the other messages separately.

> -----Original Message-----
> From: Elliotte Rusty Harold [mailto:elharo AT metalab.unc.edu]
[...] 
> 1. XOM handles external entities. Sparta doesn't. This is a big 
> difference for some files. 

Yes you are right.  This should go in the list of XOM-Sparta differences.  I 
plan to implement this feature in a future release because it is a pretty 
simple low-impact change that fixes a significant Sparta non-compliance.

> Looking at the API, it's not clear what 
> happens when you do hit an unresolved entity. Is there any 
> representation of this in the tree? or do you just throw it away? or 
> throw an exception?

In general, entities are not represented in the Sparta DOM tree -- they are 
resolved at compile time.  If the compiler hits an unresolved entity it puts 
in a blank and puts a warning in the log -- it does not throw an exception.

> 2. According to your web site., "Sparta ignores namespaces. <a:Foo> 
> and <b:Foo> are treated as if ":" was a normal tag character. These 
> will be considered different tags even though strictly speaking there 
> is a possibility that the a: and b: namespace have the same URI. (The 
> Thermopylae wrapper does however support namespaces.)"
> 
> This is so wrong. It rules Sparta out for so many applications. In 
> 2003, without namespace support, I have to say an API is DOA. It's 
> unclear how you can claim to support any of XPath, since XPath is 
> completely dependent on namespaces. XPath, XSLT, XInclude, SVG, RDF, 
> Schemas, XHTML, XLink, XQuery, and so many more all depend on 
> namespaces.

Granted there are some applications which will rely on strict namespace 
compliance, to for example deduce the precise semantics of a particular 
element or attribute.  But I still think that using the simple rule of 
treating the ":" as a normal tag character allow you to parse the vast 
majority of actual XML files out there in a way that is useful for a large 
number of applications.  For example, in pretty much every XML files that I 
have seen the same prefixes are used for the same URIs throughout the file.

Having said that however, I think that adding namespace support is a prime 
candidate for a future enhancement of Sparta, but I want to see if I can do 
it without significantly increasing the code size, parsing speed, or API 
complexity.

__
Eamonn O'Brien-Strain
HP Labs
eob AT hpl.hp.com