Hmmm... that term sounds vaguely familiar.   What does “MS” stand for, again?

;-)

From the look of things, it seems like XML::XPath might do such a good job of ripping into the XML structure that a few short scripts might just do it for us.