RE: [xml-dev] extending enumerated lists with xsd:union

["Priscilla Walmsley" <pwalmsley-lists@datypic.com>]

Hi Stephen,

I can see that it is tempting to use substitution groups (or some other
schema extension mechanism) for versioning, but I have found that it just
adds too much complexity.  

At a former client of mine (GS1 formerly known as EAN/UCC) we debated about
this too (both in the context of versioning and extending code lists for
different contexts).  We finally ended up putting every code list in a
separate schema document (just like UBL does) and creating an entirely new
type definition (with the same name) for each version of a code list. If you
wanted to change to a different version of a code list you would just swap
out the schema document.  Minor versions are backward compatible (also like
UBL, I think) but that is just enforced by the designers, not any particular
schema feature.

As far as extensibilty, one approach I have seen at another client is to
make the base type a union of some specific values and xs:token, then
restrict that:

Base Type:
<xs:simpleType name="TransactType">
 <xs:union memberTypes="xs:token">
  <xs:simpleType>
   <xs:restriction base="xs:token"> 
    <xs:enumeration value="A"/>
    <xs:enumeration value="B"/>
   </xs:restriction>
  </xs:simpleType>
 </xs:union>
</xs:simpleType>  


<xs:simpleType name="MyTransactType">
 <xs:restriction base="TransactType"> 
  <xs:enumeration value="A"/>
  <xs:enumeration value="B"/>
  <xs:enumeration value="C"/>
 </xs:restriction>
</xs:simpleType>

Of course, you have very little control, and the values in the base type are
suggestions more than anything.  

Hope that helps,
Priscilla

----------------------------------------------------------------- 
Priscilla Walmsley                 http://www.datypic.com 
Author, Definitive XML Schema     (Prentice Hall PTR) 
        Definitive XQuery (coming in 2006)
-----------------------------------------------------------------
 

> -----Original Message-----
> From: Stephen Green [mailto:stephen_green@seventhproject.co.uk] 
> Sent: Tuesday, August 09, 2005 4:38 AM
> To: xml-dev@lists.xml.org
> Subject: Re: [xml-dev] extending enumerated lists with xsd:union
> 
> Priscilla
> 
> Many thanks for this. It looks unlikely then that
> we will get codelist schemas which are extensible
> using derivation. As Marty Burns has asked, we'd
> appreciate any ideas for alternatives but we have
> of course considered that redeclaration of an
> entire new set of schemas without a trail, as it
> were, of derivation (polymorphism-like) may be
> necessary for codelists (as has been our fall-back
> strategy for UBL 1.0).
> 
> I personally would like to see substitution groups
> used to create minor version schemas from previous
> sets of major or minor schemas to provide the 'audit
> trail' as it were, though I accept that this complicates
> things somewhat. We have looked at the mechanism
> (substitution groups without using abstracts) but not
> made a decision yet and part of the discussion now
> is how it might be applied to codelists. As an interim
> plan we aim to make it possible for anyone to extend
> our next version schemas (2.0) in this way, though we
> remain to decide whether to do so ourselves for minor
> versions. Others may wish to use this inheritance
> mechanism to customise our schemas but, though
> we'd like to ensure this is possible (with all global
> elements as well as types), we haven't decided whether
> to recommend it.
> 
> Thanks again
> 
> All the best
> 
> Stephen Green
> 
> 
> ----- Original Message ----- 
> From: "Priscilla Walmsley" <pwalmsley-lists@datypic.com>
> To: "'Stephen Green'" <stephen_green@seventhproject.co.uk>;
> <xml-dev@lists.xml.org>
> Sent: Monday, August 08, 2005 2:22 PM
> Subject: RE: [xml-dev] extending enumerated lists with xsd:union
> 
> 
> > Hi Stephen,
> >
> > Union is not considered "derivation" for the purposes of 
> substitution
> > groups.  But your example is more complex than that.  You've got (at
> least)
> > two simple times: S1 (the original currency code list) and 
> S2 (your union
> of
> > the original with new values.) Then you've got (at least) 
> two complex
> types:
> > C1 that uses S1 for its simple content, and C2 that uses S2 
> for its simple
> > content.  C2 would not be considered to be derived from C1 
> just by virtue
> of
> > the fact that its content type is "derived by union" from 
> C1's content
> type.
> > This would be true whether you were using union or even simple
> restriction.
> >
> >
> > If your goal is to version the UBL schemas, have you considered just
> > creating an entirely new set of schemas with no relation to 
> the previous
> > version?  (And attempting to control backward compatibility in other
> ways.)
> >
> > Hope that helps,
> > Priscilla
> >
> > ----------------------------------------------------------------- 
> > Priscilla Walmsley                 http://www.datypic.com
> > Author, Definitive XML Schema     (Prentice Hall PTR)
> >         Definitive XQuery (coming in 2006)
> > -----------------------------------------------------------------
> >
> >
> > > -----Original Message-----
> > > From: Stephen Green [mailto:stephen_green@seventhproject.co.uk]
> > > Sent: Monday, August 08, 2005 8:37 AM
> > > To: xml-dev@lists.xml.org
> > > Subject: [xml-dev] extending enumerated lists with xsd:union
> > >
> > > Greetings xml-dev
> > >
> > > I've been looking at a way to use xsd:union to extend
> > > a list of enumerated values (a codelist in this case) as
> > > follows:
> > >
> > > <xsd:schema
> > > xmlns="urn:oasis:names:prototype:ubl:schema:xsd:CurrencyCode-2.1"
> > > xmlns:xsd="http://www.w3.org/2001/XMLSchema";
> > > targetNamespace="urn:oasis:names:prototype:ubl:schema:xsd:Curr
> > > encyCode-2.1"
> > > xmlns:cur="urn:un:unece:uncefact:codelist:draft:54217:2001"
> > > xmlns:ccts="urn:oasis:names:draft:ubl:schema:xsd:CoreComponent
> > > Parameters-1.1
> > > "
> > > elementFormDefault="qualified" attributeFormDefault="unqualified"
> > > version="Ext">
> > >
> > > <!-- ===== Imports ===== -->
> > > <xsd:import
> > > namespace="urn:un:unece:uncefact:codelist:draft:54217:2001"
> > > schemaLocation="../../../LatestDraft-UBL-2.0/xsdrt/common/Code
> > > List_CurrencyC
> > > ode_ISO_7_04.xsd"/>
> > >
> > > <!-- ===== Type Definitions ===== -->
> > > <xsd:simpleType name="CurrencyCodeContentType">
> > > <xsd:union memberTypes="cur:CurrencyCodeContentType
> > > ExtCurrencyCodeContentType" />
> > > </xsd:simpleType>
> > > <xsd:simpleType name="ExtCurrencyCodeContentType">
> > > <xsd:restriction base="xsd:normalizedString">
> > > <xsd:enumeration value="FQD" />
> > > </xsd:restriction>
> > > </xsd:simpleType>
> > >
> > > </xsd:schema>
> > >
> > > where the imported schema contains the original 
> enumerated list I'm
> > > extending
> > >
> > > <xsd:schema
> > > targetNamespace="urn:un:unece:uncefact:codelist:draft:54217:2001"
> > > xmlns:clm54217="urn:un:unece:uncefact:codelist:draft:54217:2001"
> > > xmlns:xsd="http://www.w3.org/2001/XMLSchema";
> > > elementFormDefault="qualified"
> > > attributeFormDefault="unqualified">
> > > <xsd:simpleType name="CurrencyCodeContentType">
> > > <xsd:restriction base="xsd:token">
> > > <xsd:enumeration value="AED"/>
> > > <xsd:enumeration value="AFN"/>
> > > ...
> > > <xsd:enumeration value="ZMK"/>
> > > <xsd:enumeration value="ZWD"/>
> > > </xsd:restriction>
> > > </xsd:simpleType>
> > > </xsd:schema>
> > >
> > > Is this a valid derivation from the original 
> CurrencyCodeContentType ?
> > >
> > > I ask because when I try to create a substitutionGroup based
> > > in the derived
> > > type
> > > substituting an element based on the original type I find
> > > that the Microsoft
> > > Development Environment 2003 (for VS.NET 2003) throws an
> > > error claiming
> > > that the derivation isn't valid.
> > >
> > > I'm aware that there is variation still in how parsers support
> > > substitutionGroups
> > > and maybe that is the problem here, but equally I'd like any
> > > confirmation
> > > folk
> > > can give:
> > > that this is a tool error and/or
> > > that the derivation using union is sufficient for use in a
> > > substitution
> > > group
> > >
> > >
> > > My substitution looks like this:
> > >
> > > <xsd:element name="PricingCurrencyCode"
> > > substitutionGroup="cbc:PricingCurrencyCode"
> > > type="PricingCurrencyCodeType"
> > > />
> > >
> > > ...
> > > <xsd:complexType name="PricingCurrencyCodeType">
> > > <xsd:simpleContent>
> > > <xsd:extension base="sdt2-1:CurrencyCodeType"/>
> > > </xsd:simpleContent>
> > > </xsd:complexType>
> > >
> > > where the sdt21:CurrencyCodeType looks like
> > >
> > > <xsd:complexType name="CurrencyCodeType">
> > > <xsd:simpleContent>
> > > <xsd:extension base="cur:CurrencyCodeContentType">
> > > <xsd:attribute name="listID" type="xsd:normalizedString" 
> fixed="4217"
> > > use="optional"/>
> > > ...
> > > <xsd:attribute name="listSchemeURI" type="xsd:anyURI"
> > > fixed="urn:oasis:names:prototype:ubl:schema:xsd:CurrencyCode-Ext"
> > > use="optional"/>
> > > </xsd:extension>
> > > </xsd:simpleContent>
> > > </xsd:complexType>
> > >
> > >
> > > The error message I get tells me:
> > >
> > > C:\Documents and
> > > Settings\admin\Desktop\SubstitutionGroups\Prototype-UBL-2.1\xs
> > > drt\common\UBL
> > > -CommonBasicComponents-2.1.xsd(21):
> > > 'urn:oasis:names:prototype:ubl:schema:xsd:CommonBasicComponent
> > > s-2.1:PricingC
> > > urrencyCode' cannot be a member of substitution group with
> > > head element
> > > 'urn:oasis:names:draft:ubl:schema:xsd:CommonBasicComponents-2.
> > > 0:PricingCurre
> > > ncyCode'. An error occurred at C:\Documents and
> > > Settings\admin\Desktop\SubstitutionGroups\Prototype-UBL-2.1\xs
> > > drt\common\UBL
> > > -CommonBasicComponents-2.1.xsd, (21, 3).
> > >
> > > C:\Documents and
> > > Settings\admin\Desktop\SubstitutionGroups\Prototype-UBL-2.1\xs
> > > drt\common\UBL
> > > -CommonBasicComponents-2.1.xsd(40): Invalid attribute
> > > restriction. An error
> > > occurred at file:///C:/Documents and
> > > Settings/admin/Desktop/SubstitutionGroups/Prototype-UBL-2.1/xs
> > > drt/common/UBL
> > > -SpecializedDatatypes-2.1.xsd, (40, 10).
> > >
> > >
> > > XML Spy 2005 gives no such error and valids an instance with
> > >
> > > ...
> > > <cbc:Note>sample</cbc:Note>
> > > <cbc2-1:PricingCurrencyCode>FQD</cbc2-1:PricingCurrencyCode>
> > > <cbc:LineExtensionTotalAmount
> > > currencyID="GBP">100.00</cbc:LineExtensionTotalAmount>
> > > ...
> > >
> > > as valid.
> > >
> > > I attach a set of schemas which I've been using (zip file
> > > with extension
> > > renamed to .zzz).
> > >
> > > Many thanks
> > >
> > > Stephen Green
> > >
> >
> >
> >
> > -----------------------------------------------------------------
> > The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> > initiative of OASIS <http://www.oasis-open.org>
> >
> > The list archives are at http://lists.xml.org/archives/xml-dev/
> >
> > To subscribe or unsubscribe from this list use the subscription
> > manager: <http://www.oasis-open.org/mlmanage/index.php>
> >
> 
> 
> -----------------------------------------------------------------
> The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> initiative of OASIS <http://www.oasis-open.org>
> 
> The list archives are at http://lists.xml.org/archives/xml-dev/
> 
> To subscribe or unsubscribe from this list use the subscription
> manager: <http://www.oasis-open.org/mlmanage/index.php>
>