How much web information could be captured: just the overall address or a list of each document viewed?