I'm monitoring a few and I didnt give it a second thought until recently I happened to notice that the PDF file had actually changed...

If direct PDF files arent compared, does anyone know of a free service that renders a PDF file as a webpage, so I can route the file through the service to diffbot? (Akin to what viewing a PDF file through Google is like...)

@mike 's been looking in this actually.

for now, have you tried searching for the pdf through google, and using their pdf->html cached output?

for example, you could use this page to watch for changes to the 2010 w-4 pdf document: http://webcache.googleusercontent.com/search?q=cache%3AD8hHb4MTkS4J%3Awww.irs.gov%2Fpub%2Firs-pdf%2Ffw4.pdf+pdf+file&hl=en&gl=us (though it is unlikely this example will ever change)