> If someone wants to write a regular expression client that 
> just rips out the contents of <recordData>,


Then they are treating the XML document as a textual document and if
they hit problems due to definitions (namespaces, entities, etc.)
elsewhere in the document and referenced in the bit they've extract then
that is their look-out (the XML worlds response to anyone who complains
being DDTT - T in this case being extracting via a regular expression
client ;-)  )

Matthew