W3C home > Mailing lists > Public > www-rdf-comments@w3.org > April to June 2005

Re: Error in RDF/XML Syntax Specification?

From: Dave Beckett <dave.beckett@bristol.ac.uk>
Date: Wed, 6 Apr 2005 10:34:50 +0100
To: Patrick Stickler <patrick.stickler@nokia.com>
Cc: "ext Pat Hayes" <phayes@ihmc.us>, www-rdf-comments@w3.org, Arjohn Kampman <arjohn.kampman@aduna.biz>
Message-ID: <20050406103450.09ffcfb4@hoth.ilrt.bris.ac.uk>

(re-editing to remove top-quoting so time goes forward)

> On Apr 5, 2005, at 20:02, ext Pat Hayes wrote:
> >
> > Im presuming that "" means the empty string, right? I don't think  
> > there are any serious issues if that is true. For example,  
> > ""^^xsd:string makes sense (I think. Its rather hard to discover if  
> > the XML schema spec allows empty strings) but ""^^xsd:number doesn't  
> > because the empty string isn't a legal lexical form for xsd:number.

xsd:string ( http://www.w3.org/TR/xmlschema-2/#string ) does not
define any lexical representation (!) but assuming it is the value
space, [[the set of finite-length sequences of characters]] then
"" is an allowed lexical form

xsd:number I assume is a typo, there's no such type.

xsd:integer ( http://www.w3.org/TR/xmlschema-2/#integer ) does not
seem to forbid "" either, as a [[finite-length sequence of decimal
digits]] can also be 0-length, although what value has lexical form
"" is not defined.

xsd:decimal ( http://www.w3.org/TR/xmlschema-2/#decimal ) is similar.
The canonical represetnation requires a '.' so cannot be empty but
that does not prevent the lexical form "".

and so on


> > There might be some issues if we needed to have 'empty values' in  
> > value domains, but I don't see that arising here.

I think such datatype interpretation work is outside of RDF; as long
as the lexical forms (always unicode strings) are encodable, it is
ok.  No reasons somebody could not have "NULL"^^my:sql or ""^^my:sql


On Wed, 6 Apr 2005 09:48:14 +0300, Patrick Stickler <patrick.stickler@nokia.com> wrote:

> OK. Thanks, Pat. So then for datatypes which do not
> have null lexical forms we treat a null lexical form the

'null' isn't a good term to use here, it's a 0-length lexical form
or unicode string.  null would more be in terms of datatype values?

> same as any other invalid lexical form. I.e. both
> 
>     ""^^xsd:number
>     "abc"^^xsd:number
> 
> are both invalid, for the same reason, that the specified
> lexical forms are not in fact in the lexical space of the
> specified datatype.

They are only datatype-invalid under some datatype interpretation
(here XSD).  They are still valid RDF Typed Literals:

  [[If the lexical form is not in the lexical space of the datatype
   associated with the datatype URI, then no literal value can be
   associated with the typed literal. Such a case, while in error, is
   not syntactically ill-formed.]]  
 -- http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#section-Literal-Value

i.e. it's not an illegal RDF triple

but it is an "ill-typed literal", see "The third condition" of
http://www.w3.org/TR/2004/REC-rdf-mt-20040210/#DTYPEINTERP

> We then leave it up to the "owners" of a given datatype
> to clarify whether or not the null lexical form is or
> is not a member of the lexical space of the datatype.

Yes, which is outside RDF, per RDF Core's datatype design.
  http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#section-Datatypes-intro

Dave
Received on Wednesday, 6 April 2005 09:36:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 21 September 2012 14:16:34 GMT