W3C home > Mailing lists > Public > www-xml-schema-comments@w3.org > July to September 2005

[Bug 2309] Clarify identity of values across related simple types

From: <bugzilla@wiggum.w3.org>
Date: Wed, 28 Sep 2005 11:48:55 +0000
To: www-xml-schema-comments@w3.org
Message-Id: <E1EKaQh-0008Lc-NX@wiggum.w3.org>


           Summary: Clarify identity of values across related simple types
           Product: XML Schema
           Version: 1.0
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: XSD Part 1: Structures
        AssignedTo: ht@w3.org
        ReportedBy: cmsmcq@w3.org
         QAContact: www-xml-schema-comments@w3.org

In 3.11.1, Structures says "Values of differing type can only be 
equal if one type is derived [a recent proposal substitutes
'constructed'] from the other, and the value is in the value 
space of both.

This has, perhaps, two problems:  

(1) It suggests that any given value belongs just to one type, 
while Datatypes is more or less elaborately built around the 
notion that simple type restriction creates subsets of value 
spaces, which means the xsd:integer value 37 is identical to
the the xsd:decimal value 37.

(2) Two types can have overlapping value spaces (in the world
according to Datatypes) if each is constructed from the same
primitive type; it is NOT necessary that one be constructed
from the other.  An integer that comes in tagged as being of
type xsd:integer and an integer that comes in tagged as 
being of type (union of xsd:decimal and string, with member
type xsd:decimal) should be identical for the purposes mentioned

A related point is that if singleton lists of integers are
to be identical to atomic integers (some WG members believe this
is entailed by the decisions on Bug 2045 [and Bug 2046 and
Bug 2047] made in May 2005 in North Carolina, others are not
completely sure, see resolution of those bugs for the final 
story), then this sentence and others will almost surely need
further work.

This requires both a corrigendum in 1.0 and a change in 1.1.
Received on Wednesday, 28 September 2005 11:49:00 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:50:04 UTC