Re: rdfs:XMLLiteral related syntax changes

[caught in spam trap -rrs]

Date: Mon, 28 Oct 2002 15:46:43 -0500 (EST)
Message-Id: <200210282244.58618.jjc@hpl.hp.com>
From: Jeremy Carroll <jjc@hpl.hp.com>
To: w3c-rdfcore-wg@w3.org
Subject: Re: rdfs:XMLLiteral related syntax changes

 > 2) Remove the xml"<a>foo</a>" form in N-Triples

agree

 > 1) Forbid rdf:datatype="&rdfs;XMLLiteral"

I don't really see this as necessary. If you wanted to say something peculiar
this would be a way to be clear.

e.g.

   <ex:prop rdf:parseType="Literal"><a>blah</a></foo:prop>

and

   <ex:prop rdf:datatype="&rdfs;XMLLiteral">&lt;a&gt;blah&lt;/a&gt;</foo:prop>

entail one another (with knowledge of the datatype).

Yet whether or not they are identical is implementation dependent.

We know that:


   <ex:prop rdf:datatype="&rdfs;XMLLiteral">&lt;a &gt;blah&lt;/a&gt;</foo:prop>

is also entailed by and entails the other two, but is syntactically different
from the second.
(The extra whitespace is not significant).

Admittedly there aren't any reasons I can think of to want to be peculiar, but
it's less work to not define this special case - ummm, maybe not ...

For systems that read n-triple they really need to have an implemnetation of
rdfs:XMLLiteral mapping, then they can process the above XML correctly. For
systems that don't read n-triple, then just writing the SAX rules to
canonicalize the parseType="Literal" is probably the path of easiest
implementation. We make it easier for them by simply banning the datatype.

I am fairly neutral, weak preference agaisnt banning
rdf:datatype="&rdfs;XMLLiteral"

Jeremy


Jeremy

Received on Monday, 28 October 2002 16:30:20 UTC