W3C home > Mailing lists > Public > public-rdf-wg@w3.org > November 2011

Re: ISSUE-13: a proposal for rdf:XMLLiteral

From: Ivan Herman <ivan@w3.org>
Date: Thu, 17 Nov 2011 09:51:37 +0100
Cc: <public-rdf-wg@w3.org>
Message-Id: <A9E89606-9AA5-4CF9-AC7F-79820CFF7763@w3.org>
To: Peter Frederick Patel-Schneider <pfps@research.bell-labs.com>

On Nov 16, 2011, at 21:51 , Peter Frederick Patel-Schneider wrote:

> This would still require any tool that does inference to correctly
> create XML infosets.  

That is correct. As Jeremy says, if XML Literal becomes optional, then this may still be o.k.

Let me also add: we should realize that there *are* legitimate usages of XML Literals, even if they are in the long tail, so to say. We always referred to I18N usages, but that is not the only one. For example, if I want a literal that expresses a mathematical formula, then using the content markup of MathML is still one of the best solutions (though not the only one, one can use LaTeX formulae). Similarly, if I want to describe some 2D geometry in a literal, using SVG may be the most obvious approach. In all those cases I may want to compare, say, two math formulae whether they are identical, and referring to the infoset is a viable solution.

I think that Peter's proposal goes a bit too far. Sorry Peter...

Ivan



> My proposal would eliminate that.
> 
> peter
> 
> 
> From: Ivan Herman <ivan@w3.org>
> Subject: Re: ISSUE-13: a proposal for rdf:XMLLiteral
> Date: Wed, 16 Nov 2011 13:43:36 -0600
> 
>> Well, I try to keep a bit closer to what we have now. Here is my view... (To make things simpler at this point I do as if the xml literals had a root element. As Jerremy said this is not exactly true, but that can be folded in later)
>> 
>> 1. The lexical space for XML literals are valid XML strings
>> 2. The value space for XML literals are XML Infosets.[1]
>> 3. The lexical to value mapping means parsing the XML string and build (some internal representation of) the infoset.
>> 4. RDF serialization SHOULD attempt to parse XML Literals without chaning the input strings. If this is not possible, the transformation they perform should have the same value, ie, yield the same XML Infosets.
>> 
>> That is it. Of course, one possible way of implementing vale space check is to use canonicalization. But one can also imagine a comparison of some internal representation of a DOM tree, for example; the details are not covered by the RDF standards.
>> 
>> Ivan
>> 
>> [1] http://www.w3.org/TR/xml-infoset/
>> 
>> ----
>> Ivan Herman
>> Tel:+31 641044153
>> http://www.ivan-herman.net
>> 
>> On 16 Nov 2011, at 18:39, Peter Frederick Patel-Schneider <pfps@research.bell-labs.com> wrote:
>> 
>>> Here is my my proposal for rdf:XMLLiteral
>>> 1/ The datatype rdf:XMLLiteral is a sub-datatype of xsd:string.
>>> 2/ The lexical space of rdf:XMLLiteral is the set of finite-length
>>> sequences of characters.
>>> 3/ The value space of rdf:XMLLiteral is the set of finite-length
>>> sequences of characters.
>>> 4/ The lexical to value mapping for rdf:XMLLiteral is the identity
>>> mapping. 
>>> 5/ RDF syntaxes may define a special syntax for rdf:XMLLiteral and the
>>> parse process for that RDF syntax may modify the surface form into some
>>> other lexical form.
>>> 6a/ In RDF/XML, parseType literal is one of these special syntaxes and
>>> the modification produces the canonical form.
>>> 6b/ In RDF/XML, parseType literal is one of these special syntaxes and
>>> the modification does not change the canonical form.
>>> 
>>> 
>>> I prefer 6a.
>>> 
>>> 
>>> peter
>>> 
>> 
> 


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
FOAF: http://www.ivan-herman.net/foaf.rdf
Received on Thursday, 17 November 2011 08:48:56 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:46 GMT