Hi Bryan,
> By using the id/idref pair you know
> that any <pc> with an id is essentially <sc>.
> And any <pc> with and idref is essentially <ec>.
I see. The thing I would not like at all with such solution is that it make the content of <pc> impossible: in one case it could contain translatable text, in the other it must be empty. When you have <pc id='1'/> how do you know which case it is?
It simply feels wrong to use the same element for very different semantic: mark a span in one case, mark a boundary in the other. If <sc>/<ec> cannot be used I'd rather use <ph> for marking non-well-formed spans.
I see your point: one could *argue* to use <pc> instead of <sc>/<ec> :) We could discuss this for some time, but I just wanted to know how you were seeing this. Let's keep the focus on the ballot proposal: whether or not <pc> should be offered in addition to another mechanism for non-well-formed spans.
Cheers,
-yves
