Is Apex-XPath scalable enough?

[JenniferSimonds]

Apex-XPath is based on the Dom.Document class. When this class parses an XML file, it constructs the whole DOM structure in memory, unlike XMLStreamReader. Because of this, there are bound to be practical limits on how large an XML file it can process.

Dom.Document imposes a limit of 50 nodes deep on the XML structure it can handle. (Although, so does XMLStreamReader, so maybe it's no less scalable?)

If there are limits to how big an XML file Dom.Document (and therefore Apex-XPath) can handle, do you know of any ways we could work around them?