New syntax spec

The draft for the refactored syntax looks much clearer.

One point that seems a little uncertain to me is whether the
new productions are supposed to be interpreted
open or closed.  

For example, take 4.12 propertyAttr: it seems
that if a typed-node element has an rdf:type attribute, it cannot have
any attributes from any other namespace.  If it does not have
an rdf:type then the element can only have one attribute in
a non-RDF namespace.

Is that correct?

Cheers
Rick Jelliffe