2048 2005-09-08 22:42:18 +0000 R-182: clarify equality and identity of union values 2007-03-29 18:48:47 +0000 1 1 1 Unclassified XML Schema Datatypes: XSD Part 2 1.0 only PC Linux ASSIGNED http://www.w3.org/2001/05/xmlschema-rec-comments.html#pfiUnionEquality needsDrafting P2 normal --- 2423 1 cmsmcq cmsmcq www-xml-schema-comments oldest_to_newest 5904 0 cmsmcq 2005-09-08 22:42:18 +0000 [This was originally raised by Stefan Wachter and booked as Rec comment R-182. It's now being transferred to Bugzilla.} A similar question (to that of R-181, see bug 2044) concerning equality arises with union types. Having the two union types: <simpleType name="u1"> <union memberTypes="string"/> <simpleType> <simpleType name="u2"> <union memberTypes="string"/> <simpleType> are these two elements equal? <element xsi:type="u1">abc<element> = <element xsi:type="u2">abc<element> See: http://lists.w3.org/Archives/Public/xmlschema-dev/2002Nov/0081.html Response from Henry: http://lists.w3.org/Archives/Public/www-xml-schema-comments/2002OctDec/0052.html The Working Group discussed this issue at the May 2005 face to face meeting and instructed the editors to prepare wording proposals. [The minutes do not explicitly say, but the editors' recollection is, that the agreement was that two unions may have overlapping value spaces, and that equality of values is based on the identity of the value (viewed as a member of a primitive datatype's value space), not on the additional identity or structural similarity of the datatypes whose value spaces it may be a member of. In other words, the two values in the example offered by Stefan Wachter are identical.] Note that we need both a 1.0 corrigendum and a wording proposal for 1.1. Because identity and equality are distinct relations in 1.1 but not in 1.0, the wording proposals will not necessarily be the same. This record is for the 1.0 corrigendum. 6689 1 davep 2005-10-14 00:42:28 +0000 (In reply to comment #0) > Response from Henry: > http://lists.w3.org/Archives/Public/www-xml-schema-comments/2002OctDec/0052.html > > The Working Group discussed this issue at the May 2005 face > to face meeting and instructed the editors to prepare > wording proposals. [The minutes do not explicitly say, but > the editors' recollection is, that the agreement was that > two unions may have overlapping value spaces, and that > equality of values is based on the identity of the value > (viewed as a member of a primitive datatype's value space), > not on the additional identity or structural similarity of > the datatypes whose value spaces it may be a member of. In > other words, the two values in the example offered by Stefan > Wachter are identical.] > > Note that we need both a 1.0 corrigendum and a wording > proposal for 1.1. Because identity and equality are > distinct relations in 1.1 but not in 1.0, the wording > proposals will not necessarily be the same. This > record is for the 1.0 corrigendum. That is also my understanding of identity and equality for 1.1: IF the two values are in the same primitive datatype's value space, and are equal/identical, then in all derivations and unions they are also equal/identical. 6697 2 sandygao 2005-10-14 16:36:18 +0000 Discussed at 2005-10-14 telecon. Resolution: classified as "clarification with corrigendum". Instruct the editor to draft wording proposal along the following lines: "The identity/equality of values in derived and union datatypes is based on their identity/equality in the primitive datatypes in whose value space they occur. The identity/equality of lists is determined item-by-item: all corresponding items must be identical/equal."