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

Re: Action-48 text: a New Plan for plain literals

From: Pat Hayes <phayes@ihmc.us>
Date: Mon, 23 May 2011 19:54:11 -0500
Cc: public-rdf-wg@w3.org
Message-Id: <9DA36A90-F906-4B4D-B8F3-EB60DE7AA085@ihmc.us>
To: antoine.zimmermann@insa-lyon.fr

On May 23, 2011, at 8:02 AM, Antoine Zimmermann wrote:

> Good. I tried to stretch this proposal in every direction to see if it's causing troubles to other specs but I couldn't. Pretty well formulated.
> 
> I have a remark though: by imposing that all concrete syntaxes have syntactic sugar for xs:string-typed literals, we make xs:string a special datatype which becomes an integral part of RDF. This was not the case before.

True, but this was in part because the XSD spec itself was being written when the RDF WG was chartered, and it was too soon to tell what was going to happen. There seems to now be a consensus that XSD, and certainly xsd:string, is ubiquitous and widely accepted. 

> In RDF 2004, XSDs were merely mentioned as a possible example of datatype map, nothing more. "foo"^^xs:string had no specific meaning under the RDF-semantics, nor under the RDFS-semantics nor even under the D-semantics when D does not map xs:string to the corresponding XML Schema datatype.
> 
> I am happy that xs:string be tightly integrated in RDF. However, I think we could go one step further by making XSDs---at least the ones mentioned in Sect.5.1 of RDF Semantics---an integral part of RDF.  This means that there should be an entailment regime that includes the XSD semantics at a "low level" (maybe at the level of RDF-entailment), rather than as an extension of RDFS-entailment.

I am sympathetic to this idea. The reason for putting datatype entailment 'after' RDFS was because in order to fully take advantage of datatype names, it is necessary to have classes (the datatype name is also a class name for the class of datatype values, and this is needed in order to state the rules regarding ill-formed typed literal strings.) 

What I would now like to do is to redefine RDF entailment not as a series of rather monolithic 'layers' as at present, but rather have all the semantic conditions triggered by the vocabulary in use, so that if you use, say, rdf:type and rdfs:Class and xsd:string, then you are committed to accepting the validity of all, but all and only, the semantic conditions that mention these vocabulary items. The current 'layers' can then be defined by the vocabulary sets they use, rather as they are at present, but other combinations and selections can be made, and the use of one or two vocabulary items does not mean that one is obliged to commit to an entire entailment regime including items which one might not be interested in.  I have not yet written up this proposal in detail, but it is on its way. I hope it will deal with the issue you raise in a satisfactory way. 

>  This would also mean that XSDs should be presented in RDF Concept, rather than loosely referenced.
> 
> In the current spec, there is no restrictions on datatype maps, such that one can define a D-entailment where the URI "xs:string" is mapped to numbers or dates or whatever.

That would be illegal in the new proposal: xsd:string would be considered an RDF 'reserved word' .

Pat

> Considering the importance of XSDs, we could impose a stricter relationship between the xsd: URIs and the XML Schema datatypes.
> 
> 
> Regards,
> AZ
> 
> Le 23/05/2011 05:01, Pat Hayes a écrit :
>> The proposal outlined in the wiki here
>> 
>> http://www.w3.org/2011/rdf-wg/wiki/StringLiterals/LanguageTaggedLiteralDatatypeProposal
>> 
>> completes Richard and my action item 48 from the last telecon.
>> 
>> Pat
>> 
>> 
>> ------------------------------------------------------------
>> IHMC                                     (850)434 8903 or (650)494 3973
>> 40 South Alcaniz St.           (850)202 4416   office
>> Pensacola                            (850)202 4440   fax
>> FL 32502                              (850)291 0667   mobile
>> phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
>> 
>> 
>> 
>> 
>> 
>> 
> 
> 
> -- 
> Antoine Zimmermann
> Researcher at:
> Laboratoire d'InfoRmatique en Image et Systèmes d'information
> Database Group
> 7 Avenue Jean Capelle
> 69621 Villeurbanne Cedex
> France
> Tel: +33(0)4 72 43 61 74 - Fax: +33(0)4 72 43 87 13
> Lecturer at:
> Institut National des Sciences Appliquées de Lyon
> 20 Avenue Albert Einstein
> 69621 Villeurbanne Cedex
> France
> antoine.zimmermann@insa-lyon.fr
> http://zimmer.aprilfoolsreview.com/
> 
> 

------------------------------------------------------------
IHMC                                     (850)434 8903 or (650)494 3973   
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
Received on Tuesday, 24 May 2011 00:54:44 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:43 GMT