Of course, in the PS case, a dummy application outside the UA could
parse the document and do the right thing.
As I understood Ned's comments, just specifying Level 2 would not be enough,
and it sounded like the attribute list would get pretty long. A dummy
application seems preferable. I agree with making information available
to the UA and I especially don't want the UA parsing the document format.
But in this case, it seems that not much is gained by having Level=2 if
you still have to go through the document to figure out WHICH level 2.