DOCBOOK-APPS: curious about xerces parser and endorsed classes

["Robert P. J. Day" <rpjday at mindspring dot com>]

  since this involves parsing potentially docbook files, i 
figured i'd throw it out here.

  i'm trying to clarify how one properly installs the xerces-j
parser for use by, for instance, saxon, and i've read conflicting
explanations.

  the xerces-j distribution comes with three jar files:

  xmlParserAPIs.jar	# DOM and SAX interfaces
  xercesImpl.jar	# the implementations themselves
  xercesSamples.jar	# some sample programs using the parser

  ok, it's clear that, if you want to run some of the sample
programs, you should have xercesSamples.jar in the CLASSPATH
somewhere.  but the other two jar files have an added complication.

  sun's "endorsed standards override mechanism" allows you to
override some of the DOM- and SAX-related classes that come with
their JDK by copying the appropriate jar file(s) to the endorsed/
directory in the JDK install hierarchy.  so which of the other
two jar files should be copied there?

  the endorsed standards web page states that the current endorsed
standards include:

	org.wc3.dom
	org.xml.sax
	org.xml.sax.ext
	org.xml.sax.helpers

  most of the docs i've seen suggest setting up xercesImpl.jar
this way, since it contains some org.wc3.dom classes, right?
certainly, the online xerces tutorial at ibm's developerworks
suggests this.

  but the xmlParserAPIs.jar file contains a number of 
org.xml.sax classes as well.  should this be installed as
endorsed as well?  anyone have a definitive answer on that?

  also, showing my ignorance of class paths in java, once a
jar file is installed in endorsed/, are all of its contents
now considered to be on the classpath?

rday