Re: [obix] Resolution procedure + schema issue - back to contract

Well, I guess I shouldn't have gone to
      Singapore (working) and missed the meeting last week...

      

      I now understand this is to avoid breaking applications that
      expect this behavior - but I have conformance and description
      questions.

      

      The answers should inform editing of the text, and also some must
      appear in the conformance section.

      

      (1) the wording is clumsy and not very clear (I think others
      mentioned that)

      (2) what is the conformance? An empty string is prohibited? so
      it's 1 or more contracts/URIs?

      (3) if there's more than one space, is it as good as a single
      space (the typical XML whitespace-doesn't-count rule)?

      (4) how do implementations enforce the one space rule? Do they
      accept any whitespace (newline, tab, space and multiples)?

      

      Thanks!

      

      bill

      --

      

        
        
        
        
        
        
        
William Cox 

        
Email:  

        
Web: http://www.CoxSoftwareArchitects.com 

        
+1 862 485 3696 mobile

        

        

      

      On 11/3/14 6:35 PM, William Cox wrote:

    

    
      
      
One more thing.

        

        "contract" was plural per discussion (and I missed last week).

        

        "contractList" appears to be a space-separated list of anyURI;
        is it really whitespace-separated? 

        

        Line 1018 says "It is a string with special structure regarding
        the space-separated group of URIs." I would suggest "It is a set
        of space-separated URIs with the components being of XML Schema
        type xs:anyURI."

        

        But that seems really strange. Why not have an actual list of
        anyURI that normal parsers and object models can handle instead
        of a parse a string?  Per line 832 it's an obix::list. Would it
        make sense for contracts (note the "S") to be such a list
        instead of a must-be-separately-parsed string?

        

        Thanks!

        

        bill

        --

        

          
          
          
          
          
          
          
William Cox 

          
Email:  

          
Web: http://www.CoxSoftwareArchitects.com 

          
+1 862 485 3696 mobile

          

          

        

        On 11/3/14 6:27 PM, William Cox wrote:

      

      
        
        All --

        

        I've finished my Jira items; several are now assigned to Craig
        for minor editorial fixes.

        

        In my Jira review, I noticed the changes around contract (still
        in the WD38 schema at lines 41-43) as

        

        <xs:simpleType name="contract">

                  <xs:list itemType="xs:anyURI"/>

              </xs:simpleType>
        

        The text related to "ContractList" (line 1020-1027 in WD38) is
        troublesome - but I could be convinced it's acceptable.

        

        Note that "contract" is in the schema and is not used.

        

        I'm thinking that since ContractList is really a conformed
        string that it would be best to have it as a simpleType with
        constraints, so

        <xs:documentation>Names the
          Contracts the Object implements. The value must conform to the
          rules for the contractList as described the
          specification.</xs:documentation>
        would more appropriately be in the description a new simpleType 
        so the constrained string is described in one place :

        <xs:simpleType
          name="contractList">

                  <xs:list itemType="xs:stringI"/>

              </xs:simpleType>
        

        Would this break anything? It jars that "contract" would mean
        "contractList" by the normative comments...Removing an unused
        type definition is IMO non-substantive so we could remove it
        later.

        

        Thoughts?

        

        Thanks!

        

        bill

        

        

        
On 11/2/14 11:47 AM, Considine,
          Toby wrote:

        

        
          
          
          
          
          

            
Please Resolve tickets rather than
              closing them.

            
 

            
In the world of standards, closing
              first says “We blew off your comment”

            
 

            
Thanks

            
 

            
tc