@Perkin - thanks for the suggestion, that is indeed a crackingly good one if it is practical. Unfortunately I know from writing the FF metadata plugin that their particular pages are a bit of a pain. Primarily because the underlying html of the page does not contain the book data, instead it is dynamically written out using javascript (do view source on a FF web page and you will see what I mean). I'm sure there is a workaround for this using webkit or something, Kovid will know.
I am guessing that such a rendering approach would not be desirable as a default (for speed reasons) but if I slapped another option checkbox on the screen there it may cater for it assuming it would indeed work... There is still no guarantees that the html would be structured sufficiently nicely to pull the books out of it but worth a look anyways.