Re: XML literals, canonical form, and normal form C problem

From: "Jeremy Carroll" <jjc@hplb.hpl.hp.com>
Subject: RE: XML literals, canonical form, and normal form C problem
Date: Thu, 25 Sep 2003 16:26:28 +0200

> Hi Peter,
> 
> Sorry to vascillate ...
> 
> On further reflection, while I did not support the editorial changes you
> were suggesting, it occurred to me that there were possible improvements
> that I would be happy to make.
> 
> A possibility is:
> 
> 
> At the end of
> section 3.4 Literals (informative)
> http://www.w3.org/TR/2003/WD-rdf-concepts-20030905/#section-Literals
> add the following new para
> 
> [[
> The string in both plain and typed literals is required to
> be in Unicode Normal Form C [NFC]. This requirement is motivated
> by [Charmod] particularly section 4 Early Uniform Normalization.
> The strings of literals can be considered
> <a href=
> "http://www.w3.org/TR/charmod/#def-construct">
> constructs</a>
> in the sense used in that section.
> ]]
> 
> and add charmod into the informative references.
> 
> Dave has indicated support for this change on the RDF Core list, and unless
> I hear objections, I will make this change.
> 
> Do you think that helps?
> Is it orthogonal to your difficulties with the current text?
> Is it positively bad?
> 
> thanks
> 
> Jeremy

While it is generally a good idea to include rationale in one's document I
am having a very hard time trying to understand what the difficulties of
merging Unicode strings that start with combining characters have to do with
RDF literals, which can start with combining characters, at least as far as
I can tell.

Perhaps you meant to point to some other section of the Charmod document? 

Peter F. Patel-Schneider

PS:  I find it difficult to conceive of any way that Charmod could be used
to provide a rationale for prohibiting non-NFC Unicode in typed literals.
For example, this means that RDF cannot use certain IRIs in the XSD
datatype xsd:anyURI.

Received on Thursday, 25 September 2003 11:02:27 UTC