This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.

Bug 4912 - Allow empty unions
Summary: Allow empty unions
Status: CLOSED FIXED
Alias: None
Product: XML Schema
Classification: Unclassified
Component: Datatypes: XSD Part 2 (show other bugs)
Version: 1.1 only
Hardware: Macintosh All
: P1 minor
Target Milestone: ---
Assignee: C. M. Sperberg-McQueen
QA Contact: XML Schema comments list
URL:
Whiteboard: cluster: absence
Keywords: resolved
Depends on:
Blocks:
 
Reported: 2007-08-03 18:56 UTC by C. M. Sperberg-McQueen
Modified: 2008-03-08 15:08 UTC (History)
0 users

See Also:


Attachments

Description C. M. Sperberg-McQueen 2007-08-03 18:56:53 UTC
XSDL 1.0 required unions to have at least two member types.

Revisions for XSDL 1.0 changed that to require only one; it
won't be a common idiom, but there might be reasons to need
to do it.

In connection with the definition of the xsd:error type, which
is defined as an empty group (so that it has no valid instances),
the Working Group decided on 27 July 2007 to remove the requirement
that groups have at least one member type.  That change has been
made in Structures, and a corresponding change must be made in
Datatypes.

Since the WG has already decided essentially what to do, I'm setting
the keyword of this issue to 'needsDrafting'.
Comment 1 C. M. Sperberg-McQueen 2007-10-30 01:56:10 UTC
On 21 September 2007 the WG accepted a wording proposal 
(http://www.w3.org/Bugs/Public/attachment.cgi?id=489)
which allows for empty unions.

The WG determined to leave this issue open until an issue of
editorial consistency was addressed.  I think this related
to the difference between saying that a given sequence-valued
property was empty and saying that it was absent, but the
minutes of 21 September 
(http://lists.w3.org/Archives/Member/w3c-xml-schema-ig/2007Sep/0004.html)
are slightly hazy on the subject.  If my memory is correct, the consistency
issue is related to that mentioned in bug 3869 relating to the treatment
of annotation components, and the two should be dealt with in tandem.
Comment 2 C. M. Sperberg-McQueen 2007-10-30 03:21:01 UTC
Having examined the relevant parts of Structures and Datatypes, I 
conclude that no further action is needed here (unless to align
the description of {member type definitions} in Structures with
that in Datatypes).  I therefore make the following wording proposal
to resolve this issue:

  Make no change to Datatypes.  Optionally align the wording of
  {member type definitions} in Structures with that in Datatypes.

I'm setting the status of the issue to needsReview to reflect the
existence of this proposal.
Comment 3 C. M. Sperberg-McQueen 2008-02-05 03:36:40 UTC
A wording proposal for this issue (among others) was placed on the
server on 4 February 2008 at 
http://www.w3.org/XML/Group/2004/06/xmlschema-2/datatypes.omnibus.200801.html (member-only link).
Comment 4 C. M. Sperberg-McQueen 2008-02-08 18:46:56 UTC
The wording proposal mentioned in an earlier comment was considered
and adopted today by the XML Schema Working Group.  Accordingly, I'm
marking this issue resolved.

Since the originator of the issue is a member of the WG, the adoption 
of the proposal by the WG is probably sufficient evidence that the
originator is content with the WG's resolution of the issue.  But if
the editors don't get around to it, it would be convenient if the 
originator could take the time to shift the status of the issue
from RESOLVED to CLOSED.  Thanks.