W3C home > Mailing lists > Public > www-rdf-interest@w3.org > October 2001

RDF/XML Containers: Syntax & Semantics

From: Vassilis Christophides <christop@ics.forth.gr>
Date: Wed, 24 Oct 2001 17:36:11 +0300 (EEST)
Message-Id: <200110241436.RAA10428@athena.ics.forth.gr>
To: www-rdf-interest@w3.org
CC: "Karsten Tolle" <tolle@dbis.informatik.uni-frankfurt.de>
Hi,

We are currently exploring the 'Refactoring RDF/XML Syntax'
(http://www.w3.org/TR/rdf-syntax-grammar/) and the corresponding test
cases for the 'rdf-containers-syntax-ambiguity'. Our remark concerns
the translation of the rdf:li to the rdf:_nnn elements. There is no
rule given to do so.

a) In the test cases we can find the mixture of using rdf:li and
rdf:_nnn elements. The existence of the rdf:_n elements has no effect
on the translation, as shown in the test cases. Given the proposed
translation we may encountered the following cases:

1) 

<rdf:Seq rdf:ID="myseq" rdf:li="a" rdf:_1="b"/>

which would create the triples: 

[rdf:Type, myseq, rdf:Seq], 
[rdf:_1, myseq, "a"], 
[rdf:_1, myseq, "b"]

2)

<rdf:Seq rdf:ID="myseq">
 <rdf:_1/>
 <rdf:li />
 <rdf:_2/>
</rdf:Seq>

Semantically this would mean for the sequence 'myseq' there are two
elements at the first position! What is the underlying semantics for
sequences?

In order to avoid these inconsistencies it should be better to
disallow the mixture of the rdf:_nnn and rdf:li elements inside one
typedNode. Do we really need both elements or wouldn't be sufficient
to use the rdf:_nnn element?

b) In the test case Nr. 5 the counting goes on even outside the
typedNode element. While in test case Nr. 8 there is a reset for the
counting. The second sounds more reasonable but at least it should be
consistent.

Best regards

Karsten Tolle

Vassilis Christophides
Received on Wednesday, 24 October 2001 10:39:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:52 GMT