[Bug 5519] Relationship between SML model validity and XSD validity assessment needs to be precisely defined

http://www.w3.org/Bugs/Public/show_bug.cgi?id=5519

           Summary: Relationship between SML model validity and XSD validity
                    assessment needs to be precisely defined
           Product: SML
           Version: LC
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core
        AssignedTo: cmsmcq@w3.org
        ReportedBy: pratul.dublish@microsoft.com
         QAContact: public-sml@w3.org


Comment #2 from
http://lists.w3.org/Archives/Member/w3c-xml-schema-ig/2008Mar/0001.html

The relationship between (SML) model validity and (XSD) validity
assessment is imperfectly expressed at best.  Some aspects of model
validity (e.g. SML indentity constraints) are parasitic on XSD
validity assessment, but that is defined in terms of its effect on the
PSVI.  Other aspects simply use RFC 2119 MUSTs wrt properties of the
model instance.  Still others say "The model MUST be declared
invalid".  A single coherent story about model validation and its
integration with (XSD) validity assessment is badly needed.  Section 7
does not really achieve this, as it uses language which is
inconsistent with that used in the various 'instance validity'
sections which precede it, as noted above.

The fact that some of the model validation language is declarative
(about properties of documents) and some procedural (about actions of
model validators) contributes to the above problem, as well as being a
bad thing in itself.  Can the procedural stuff be eliminated?

Received on Tuesday, 4 March 2008 16:42:32 UTC