Re: comments on http://www.w3.org/TR/2002/WD-rdf-schema-20021112/

My comments are interspersed below, prefixed with "#####"
============ 
Dick McCullough 
knowledge := man do identify od existent done
knowledge haspart list of proposition

  ----- Original Message ----- 
  From: Brian McBride 
  To: Richard H. McCullough ; www-rdf-comments@w3.org 
  Sent: Wednesday, November 27, 2002 12:09 AM
  Subject: Re: comments on http://www.w3.org/TR/2002/WD-rdf-schema-20021112/ 


  Hi Richard,

  At 12:33 26/11/2002 -0800, Richard H. McCullough wrote:
  >There are a number of domain & range errors in your descriptions of the 
  >RDFS properties.  You should check your descriptions against your table 
  >"RDF Properties".

  Thanks for drawing our attention to this.  Please not however that 
  references to specific errors are much more helpful.  Consider the 
  difference between:

     Your spec is full of mistakes.  You should check it over.

  and

     The range for property foobar is missing on page 5.

  The latter is more helpful to us.
  ##### I agree.  I just got lazy.  I didn't have a printout of the document, and I was trying to manage by flipping between screens.  BTW, the document has no section numbers or page numbers.

  >
  >Since you provide no definitions, there is confusion about the distinction 
  >between "resource" and "instance" and "member".  For example, in the 
  >description of "rdfs:type", does the domain of type include individuals, 
  >or classes, or both?

  Right.  We have some text in progress to clarify that.  The answer to your 
  question is that domain of type is rdfs:Resource.  RDF Schema does not 
  define the term individual.  Please note also that a class is an instance 
  (member) of rdfs:Resource.

  >
  >In your description of the property "rdfs:object", you imply that Literal 
  >is not a subclass of Resource.  That contradicts the definition of Resource.

  Right.  That is a hangover from when we were being coy about whether 
  Literals were resources or not.  Will fix.  Thanks.

  >
  >In many places, you say that x "represents" y.  You should say "denotes" 
  >or "means".

  We are in process of reviewing use of the term "represents".  Sometimes we 
  might replace it with denotes, others some variant of the verb to be.


  >The ranges in the "RDF Properties" table encourage the continuing 
  >confusion between "Class" and "Resource".
  >  With the exception of "type", the ranges should be "Resource" instead of 
  > "Class".

  What properties do you mean.  For example the range of rdfs:domain is 
  definitely rdfs:Class.
  ##### I mean EVERY property in the table.
  ##### For the rest of my discussion, I'm going to drop the qualifiers "rdf:" and "rdfs",
  ##### because I frankly can't remember which one is used, and I'm too lazy to look it up now.
  ##### For the property "domain", its domain is "Property" and its range is "Resource", i.e.,
  #####        Property has domain = Resource
  ##### which means (when combined with other info.) that
  #####        Resource has Property = Resource
  ##### which means that every resource (individual or class) has properties,
  ##### and that the value of every property is a resource (individual or class).

  >The only consistent definition of "Class" that I can come up with is: 
  >"Class" is the set of class names.

  I have shown you a description of class in discussion on rdf interest that 
  was different to that.  To the best of my knowledge, you have found no 
  inconsistencies in it.
  ##### OK, here's the proof that your definition of Class is contradictory.
  ##### I don't remember exactly what you said in the rdf interest discussion.
  ##### I think the basic idea was that "Class" is a class instead of a set of class names.
  ##### The problem with your definition comes out when you consider
  ##### the PROPER subclass relations between Class and Resource.
  ##### By the definition of Resource
  #####        every class except Resource is a proper subclass of Resource
  ##### Therefore
  #####        Class is a proper subclass of Resource
  ##### On the other hand, you have defined Class to include all classes, 
  ##### and Resource is a class, so
  #####        Resource is a proper subclass of Class
  ##### which is a contradiction.  The only logical alternative to this conclusion is that
  #####        Class is identical to Resource
  ##### and I don't think that's what you want Class to be.
  ##### Using my definition
  #####        Class is the set of all class names
  #####        the set of all class names is an individual of the class Set
  #####        the class Set is a proper subclass of Resource
  ##### and there is no contradiction.
  ##### Note that "Resource" is a member of "Class", because it is a class name,
  ##### but "Class" is not a member of "Class", because it is an individual name.

  >  Given that definition, the range of "type" is "Class"

  Given the one we are using also, the range of type is rdfs:Class.

  >, i.e., a class name.

  No, its not the name of a class.  Classes are named by URIrefs.  The range 
  of type is not uriref.

  Brian