1) Setting paths of xercesImpl.jar and xalan.jar in the CLASSPATH

No. Xalan library duplicates some code, especially http://xerces.apache.org/xerces2-j/javadocs/api/javax/xml/transform/package-summary.html, as also as some general stuff required for both Xerces and Xalan. Better idea is to look for http://xml.apache.org/crimson/. That would minimize count of Java classes loaded at runtime 
   2) Command that user should use to start the Derby i.e. java org.apache.derby.tools.ij
Do not understand how it is connected to XML data type 
   3) Mention that user needed to create a table which contains XML data type column/s, because users might thought that they can retrieve XML data from an any table.
Rather an explanation of architecture difference needed for retrieving XML data from tables without XML type column(s). And do not forget about transactions; and explanation of a fact that simple locking rows with XML data type records brake data consistency by default (why consistency couldn't be provided in any way here, actually)
   4) Mention the requirements that are needed when inserting values to a XML data typed column, i.e. need of XML keywords document & preserve whitespace, tags etc.
Dealing with whitespace and other stuff is not a problem of SQL parser, it is problem of client library to prepare XML chunks and insert into SQL in previously prepared places + validate (non-text symbols only)    
   5) A small example which is clearly shows the creation, insertion & selection using XML data type.
I think I guess where is the problem. Check out the PDF which I attached to this post. Described features differ in QA & QC conformance levels, while being grouped to same category of standard. That should be a focus.