My real problem seems to be that I need a XPATH 1.0 solution since
I want to do this in a browser environment, right?

The real problem is as follows:
- open an arbitrary web page in Firefox browser

- with a bookmarklet do an arbitrary selection in that page
  (http://en.wikipedia.org/wiki/Bookmarklet)

- then the bookmarklet generates eg. the following xpath:
  "//*[contains(normalize-space(.),'xyz')]"
  where xyz is replaced by the actual selection data

- then Mozilla's document.evaluate() is used to determine the
  corresponding node in the DOM

Any hint what can be done to make this work?
(I have no control over the webpage nor control over user selection)

	Martin Honnen
	http://msmvps.com/blogs/martin_honnen/