Re: New syntax spec

>>>Rick Jelliffe said:

> B.t.w. why is the production "literal" needed?  It says it is non-empty 
> but then parseLiteral allows list().  Why just make literal be 
> any string including the empty string?

The precise definiton of RDF literals is currently under discussion,
however we are distinguishing between what can be put in XML
attribute values in the literal production, and well-formed XML
content of things inside elements matching the
rdf:parseType="Literal" construct.

In notation terms, list(literal) is used since [children] requires a
list.  So this means, the element contains a single literal.

The parseLiteral production has [children]=list() as content since I
wanted a way to be more precise and say any well-formed XML etc. but
couldn't find an Infoset term to define that.  I couldn't re-use
'any' since that was defined for infoset property values.  I think
this is another ISSUE that I should address for that tem - thanks for
picking this up.

> Also, I am confused. I thought the RDF WG had made rdf:_n elements,
> but even the 6.31 attributes are removed. Is that correct?

The rdf:_<n> etc. attribute forms match the propAttr production and
generate properties just like any other propAttr property.  They
don't need to be a special case.  rdf:li as an *attribute* is a
special case, which will need to be considered, and we have a

and test cases for this.

Not yet folded in.

[ ... Schematron schema elided..

Interesting - note we are also thinking about the top-level <rdf:RDF>
and making it required (it isn't actually yet) , scoping RDF embedded
in general XML such as is done in SVG:

and also allow it to appear multiple times inside containing XML documents.


Received on Thursday, 13 September 2001 11:30:42 UTC