The primary concern of the default values is, it changes the document
infoset. So the parsing result will differ depending on whether you've
used the schema or not.
Also, if we choose to augment the infoset, we can't use the existing
parser; we have to have a parser that recognizes default values of RELAX
NG. Users who want to use RELAX NG have to use that special parser!
If my understanding is correct, Murata-san and James, as well as many
other people, do not like this idea of augmenting infoset.
That's why we don't have default values.
regards,
--
Kohsuke KAWAGUCHI +1 650 786 0721
Sun Microsystems
