On and off for the last couple of weeks I've been working on the Perseus-generated xml that MP will be editing, trying to make it as coherent as possible.
I now have several reasonable P5 data files (apollodorus.xml,characters.xml,groups.xml,events.xml) using a schema generated using the odd-by-example script.
The cruft left over from the P4 original will need to be addressed, though (chunks wrapped in "div1" for example). Ideally, I'd like to enclose each logical chunk (e.g. Apollodorus 1.1.1) in a div with an appropriate xml id, but the problem is that I'd end up with overlapping hierarchies. I assume that's why milestones were used by the Perseus project.