Variety of XML questions;

[James Ankrom]

Hi all, sorry I couldn't be more specific on the subject. I have a few
questions that affect the design of objects I am building, and was wondering
if anyone had any input or experiences to relate. Apologies if any of these
answers truly should be RTFM, I'm just spooling the current questions I have
off of my brain and am not sure where to look for these answers.
When you grab a reference to an XMLNode, will changes to the siblings or
parents affect the reference?
Are pointers / references to individual nodes created upon parsing, or upon
demand? (Does parsing create the node instances, or just parses the XML into
another format?)
Will deleting the parent node make the child node reference invalid or
orphaned?
Bearing encapsulation and concurrency in mind, should an object that makes
changes to a node copy the node, edit, then replace the node in the original
document? Or is it a better practice to work with the node in the document?
For example, an Inventory object would have an XML property that is set to a
node in a larger document, and it's properties are XPath queries against the
node's attributes, children, and sometimes parents & siblings. A property
might return Item objects which are similar objects relating to assigned
child nodes. I'm told that encapsulation dictates that the object should
encapsulate it's own data to protect it against changes, but that seems to
me to be contrary to the spirit of an XML document (and/or message) - that
since the document is basically a datastore, one should be able to bind to a
location in the store and edit freely while other objects may also be able
to edit the data at the same time.
Thanks for your thoughts.

-jim