W3C home > Mailing lists > Public > www-rdf-comments@w3.org > October to December 2001

Re: Elimination of container production

From: Dave Beckett <dave.beckett@bristol.ac.uk>
Date: Wed, 14 Nov 2001 11:46:05 +0000
To: Autumn Cuellar <a.cuellar@auckland.ac.nz>
cc: www-rdf-comments@w3.org
Message-ID: <6092.1005738365@tatooine.ilrt.bris.ac.uk>

I'm replying as the Editor of the 
  Refactoring RDF/XML Syntax W3C WD, 6 Sep 2001
to which you refer.

>>>Autumn Cuellar said:
> We're using RDF to reference papers that we are describing in
> an XML-based language in hopes of one day creating an easily searchable
> database of biological models.  One of the biggest appeals of using RDF
> was the ability to
> present an ordered sequence with the use of containers.  We use this
> primarily in author listings because, as any published scientist will
> tell you, the order in which the authors are listed *does* matter.
> My understanding is that part of the reason the containers production
> was removed was that parsers should not be expected to recognize that
> the <li> elements indicated an order (i.e. <_1>, <_2>, and so on).  But
> in Sect. 3.2 of the Sept. 6 Refactoring RDF/XML Syntax, there's the
> sentence "rdf:li elements will be translated to rdf:_nnn elements when
> they are found matching either a prpertElt or a typedNode."  Is that
> sentence just meant to clear up any ambiguity that previously existed?

There was no ambiguity, but the rdf:li properties had a special place
in the grammar which stuck out as rather an un-necessary special
case.  We merged the rdf:li handling into the handling of all other
properties, but had to describe the rdf:li -> rdf:_nnn rule in order
to generate the properties for containers in the same was as before.

So; this really is a refactoring of the grammar, not changing any way
of using rdf:li that previously was legal, but allowing it to be used
some new ways.

> So for an author listing, would the following indicate an ordered
> sequence?  If not, is there currently a way to do so within RDF?
> <rdf:RDF
>     xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>     xmlns:dc="http://purl.org/dc/elements/1.0/">
>   <rdf:Description rdf:about="#some_article">
>     <dc:creator>
>       <rdf:li rdf:parseType="Resource">
> 	<name>Fred Flintstone</name>
>       </rdf:li>
>       <rdf:li rdf:parseType="Resource">
> 	<name>Barney Rubble</name>
>       </rdf:li>
>       <rdf:li rdf:parseType="Resource">
> 	<name>Dino</name>
>       </rdf:li>
>     </dc:creator>
>   </rdf:Description>
> </rdf:RDF>
> Thanks in advance for clearing this up for me.

This is a (mostly) legal way to use rdf:li after the refactoring
document was published; since here you using rdf:li off a node which
is not of type rdf:Seq, rdf:Bag or rdf:Alt.

I say mostly legal since the <name> element has no declared
namespace.  I assume that is a typo.  You might want to try out the
RDF Validator at http://www.w3.org/RDF/Validator/ with your examples
which makes pretty graphs of the results.

Although this is allowed, If you wanted to say in a way that other
RDF applications can understand that these are a sequence, you will
have to use rdf:Seq inside the dc:creator property like this:

   <rdf:Description rdf:about="#some_article">
	 <rdf:li rdf:parseType="Resource">
	  <name>Fred Flintstone</name>
	 <rdf:li rdf:parseType="Resource">
	  <name>Barney Rubble</name>
	 <rdf:li rdf:parseType="Resource">

You might be interested in this document about modelling Dublin Core
in RDF/XML:  http://dublincore.org/documents/2001/08/29/dcq-rdf-xml/

Note this is *not* an official RDF Core WG document.

I hope this answers your question.


Received on Wednesday, 14 November 2001 06:46:12 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:15:17 UTC