Talking of naming, can we get a consensus on what we call the thing that results from processing a TEI ODD with teitoodd ? It's analogous to the p5subset.xml we know and love, so maybe we should call it a subset, not an ODD (which is what teitoodd implies : give it foo.odd and you will receive foo.odd.odd) . It isn't an ODD as we know it, obviously. But it isn't necessarily a subset either. I think of it as a "compiled ODD" but maybe that has the wrong connotations. Its distinguishing feature is that any *Ref elements such as elementRef classRef etc. in the input will have been resolved so maybe its a "resolved ODD" or an "explicit ODD" ...