Question:

Is there any way to do a incremental save using XMLDOMDocument (MSXML parser) and also loading part of the Document tree?

Answer:

A SaveInto or a LoadFrom command? As desirable as such a feature is, it doesn’t currently exist in the MSXML parser, although that may change with the next technology preview release.

With the upcoming SAX parser, I can see an application that will at least support the LoadFrom command?a fast stream to retrieve an element that matches a given search pattern, which in turn retrieves all of its subordinate elements. This will make XPointer possible, for starters. Currently, you have to load the entire DOM into memory, perform an XPath search, then output the node as XML, which can get costly when dealing with large XML files.