Peter Robinson wrote:
"We have, for ages now, related occurrences of <lb/> with xy coordinates,
then processed them so we can say "this piece of text, occurring after <lb
n="1"/> and before <lb n="2"/> occurs within these xy coordinates in this
image of this page". However, this won't work for the case Lou cites, of
associating text with a random space in the manuscript (though one might,
with some ingenuity, use milestones to mark out the range of text, and then
link the milestones to co-ordinates)."
Sounds like a viable way to help yourself within <text>. Still, since none
of those elements is a member of the att.coordinated class (@start, @ulx,
@uly, @lrx, @lry), one has to go a long way processing the textual markup in
order to gain similar results.