As per bug http://issues.apache.org/jira/browse/AXIS2-243, I'm
attempting to resolve Javadoc issues, in part so that I can understand
how Axis works, and I started with OM.
I find myself wondering about a general point of design as I'm
proceeding. Is OM meant to do "fail-fast" validation? Some of the
examples I can think of, where enforcement might matter:
* Are namespace name parameters valid URIs?
* Are local element names valid NCNames?
* Can the contents of a text element be serialized (character
�x1;, for example, is not allowed in XML 1.0)?
* Can the contents of an attribute be serialized?
* Can the namespaces associated with the prefix "xml" and "xmlns" be
associated with other prefixes?
* Can the prefixes "xml" and "xmlns" be reassigned to different
namespaces?
* Do you want to log a warning when redeclaring a namespace that is
already in scope (performance issue)?
Should these be enforced? If so, is enforcement conditional? Or is it
determined by code-path, namely factory creation done by the parsers
assumes validity (because it is starting from a valid state), whereas
direct manipulation of OM objects implies validation?
I have not been able figure out the intended design by looking at the
code, so I thought I'd ask.
-Eric.
