RE: XML literals

 

 -----Original Message----- 
 From: ext Dan Connolly [mailto:connolly@w3.org] 
 Sent: Fri 8/1/2003 4:17 PM 
 To: Stickler Patrick (NMP/Tampere) 
 Cc: Brian McBride; Jeremy Carroll; Graham Klyne; ext pat hayes; w3c-rdfcore-wg@w3.org; Peter F. Patel-Schneider 
 Subject: Re: XML literals
 
 

 On Fri, 2003-08-01 at 00:06, Patrick Stickler wrote:
 [...]
 >         
 > Maybe. 
 > 
 >         in 1:1 correspondence with the lexical space.
 > Right.
 
 Hmm... that one gave me pause... but OK.
 
 >         The exact nature of XML values is not specified.
 > 
 > No. This bothers me. Alot.
 
 Really? The exact nature of integers is not specified;
 just various relationships like addition and
 multiplication of them.
 

No. The nature of integers *is* specified, by the relationships such as addition, multiplication, etc.

Perhaps I've misunderstood Pat, but I understood him to be saying that we simply are not going to say anything about the L2V mapping other than it is 1:1 -- i.e. it may very well be that the value space of rdfs:XMLLiteral is rdfs:subClassOf xsd:integer! Since after all, we could posit an L2V mapping that mapped canonical XML fragments onto integers in a 1:1 relationship (theoretically at least).

 

 
 If it bothers you, then feel free to suggest an alternative.
 

 

I did.

 
 We could pick any set that's in 1-1 correspondence
 with the lexical space; e.g. pairs
         (humpty-dumpty, lexical-value)
 
 or perhaps less churlishly...
         (http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral <http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral> , lexical-value)
 
 
 > It is our responsibility to define what the values of XML Literals
 > are.
 
 Only inasmuch as required to get the technology deployed.

 

If you mean deployed in a fully system independent, portable manner that facilitates the interchange of RDF expressed knowledge between arbitrary applications where interoperability is based on our specs, then I agree.

But I'm not sure you're saying that. "Deployed" can mean so many things.

 
 
 > It's *our* datatype, and no'one else should have to define it,  or
 > guess.
 
 They don't have to guess; what Pat wrote tells them everything
 they need to know.

I'm an implementor, and *I* am unsure of what the value space of rdfs:XMLLiteral contains, per Pat's recent recommended change. It is clearly *not* the set of lexical forms -- which is the only thing that looks like XML to me. It could be the set of all pumpkins that have ever existed for all I know. 

I expect to operate on values, not lexical forms, and if the datatype definition does not tell me which value I have, given a particular lexical form, the that definition is incomplete and unnacceptable.

  
  > I've never understood the opposition to having a value space
  > consisting of infosets. I wish someone would tell me what significant
  > problem or issue I'm missing...
  
  How to construct an infoset and how to compare them isn't
  specified.
  
  

Thank you. That seems a good reason for not having infosets in the value space.

Still, we do need to define what the value space of rdfs:XMLLiteral is. At the very least, we can simply make rdfs:XMLLiterals self denoting, no?

Patrick

Received on Monday, 4 August 2003 03:39:39 UTC