Re: Questions on Container Proposal

>>>Aaron Swartz said:
> Reading through the container proposal:
> I have several questions.
> Has this proposal been implemented yet? If so, is there an online demo and
> what was the difficulty/effect of this implementation?

Since the proposal is a simplification of the grammar, it has already
been mostly implemented since RDFFilter (Java) and the Repat parser
(C - recoding of same algorithm) both use it.  This means that Jena
has it and also Redland since mid-2000.  All Redland online demos
show it and since repat is the default parser, it has working for
months on many deployed rdf/xml.

Proposal #4 is not something that has been implemented since
aboutEach processing is something often ignored by parsers.

Proposals 1-3 make parsing easier and the grammar more consistent.  I
have also been working on it for my parser Rapier, but also provide
the existing interpretation too.  The new one is a lot easier.

> > The rdf:li processing of sub-elements is independent of the processing of
> > enclosing elements. The selection of an ordinal to replace an rdf:li is not
> > affected by any ordinals encountered in sub-elements of the element. The
> > selection of an ordinal to replace an rdf:li is not affected by ordinals
> > encountered in enclosing elements.
> If this is true, what would be the results of something like this:
> <rdf:Description rdf:about="#a">
>   <rdf:li>
>     <rdf:Description rdf:about="#a">
>       <rdf:li>inner</rdf:li>
>     </rdf:Description>
>   </rdf:li>
> </rdf:Description>
> It seems you'd have something like:
>     <#a> rdf:_1 <#a>.
>     <#a> rdf:_1 "inner".
> is this correct?


Well, you can create examples like that in many ways:
 <rdf:Description rdf:about="#a">
     <foo:bar rdf:resource="" />
 <rdf:Description rdf:ID="a">
     <foo:bar rdf:resource="" />

Since RDF has no processing model, it isn't clear where the
responsibility lies to check either of the above examples.  In the
latter case XML is happy since the same ID is not used in one XML
file, but generates the same RDF statement due to the RDF URI rules.

> > Note that XML states that the ordering of attributes is not significant and
> > that the same attribute name cannot appear more than once on an element. It
> > probably unwise to use rdf:li as an attribute. If it is used in presence of
> > other ordinal property attributes, the ordinal property with which it will 
> > replaced is undefined.
> Err, this isn't very helpful. Could the proposal provide precise guidance on
> this -- like, don't do it? Or is that what is meant by undefined?

We were proposing stuff outside a working group had no mandate or
process then to really decide stuff.  Now we have RDF Core and I can
propose we say it is recommended that rdf:li is not used as an
attribute since the results are undefined in existing
implementations.  A stronger version would be to forbid it.


Received on Tuesday, 8 May 2001 03:43:22 UTC