Wendell Piez wrote:
> That's the same as saying always return me the
> div[count(ancestor::div)=1] level, so no, it's not an argument for
> numbered divs.
> In fact, given that TEI lets you start with either div0 or div1 when you
> use numbered divs, it's an argument against numbered divs, as unnumbered
> ones can be more precise -- assuming you're not typing them, which you
> should. :-)