> From [log in to unmask] Fri Jul 12 13:09 MET 2002
> User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.1a) Gecko/20020610
> X-Accept-Language: en-us, en
> MIME-Version: 1.0
> X-Enigmail-Version: 0.62.4.0
> X-Enigmail-Supports: pgp-inline, pgp-mime
> Content-Transfer-Encoding: 8bit
> Date:         Fri, 12 Jul 2002 11:28:32 +0000
> From: Dan Brickley <[log in to unmask]>
> Subject: Re: Mixing dc qualifier with other namespace's qualifier allowed ?
> To: [log in to unmask]
>
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
>
> Hi Carl,
>
> Carl Lagoze wrote:
>
>
> > There is currently no DCMI recommended XML schema for qualified Dublin Core.  A group (of which I am a part) is finalizing a recommendation in this are, which will be released for discussion in the next couple of weeks.
>
> Interesting. Are your discussions publically readable anywhere? Is this
> OAI work?
>
>
> > The XML snips you have included below does not conform to this schema.  Furthermore, the proposed schema does not accommodate the notion of additional attributes that you use in your examples.  While it indeed may be useful for individual communities to employ such additions, easy interoperability amongst parties relies on reasonably strict common languages (defined by schema).  This is not easy if each communicating party mixes in elements are attributes that are unique to their application.
>
> Yes, predictability is nice. But the kind of rigid restrictions you seem
> to be advocating (difficult to tell without seeing the schema) come with
> a different associated cost: others can't easily re-use it, so they go
> off and create their own XML Schemas, fragmenting the information space
> still further.
>
> One trick (there are others) for hanging things together in such an
> environment is to make sure that instance data for all such XML Schemas
> also parses as RDF. That way, RDF apps can merge from all sources,
> regardless of the XML Schema that policed them. Is this what you're doing?

Well Recommendation 8 in the dcmi-xml-guidelines will not work with RDF/XML syntactically.
The (unqualified) attribute "scheme" seems to be of simpleType
"string" - though no details are given.



rs

>
> Dan
>
> >
> > Carl
> >
> >
> >>-----Original Message-----
> >>From: Piccand Régis [mailto:[log in to unmask]]
> >>Sent: Friday, July 12, 2002 5:25 AM
> >>To: [log in to unmask]
> >>Subject: Mixing dc qualifier with other namespace's qualifier
> >>allowed ?
> >>
> >>
> >>Hello all,
> >>
> >>I usually use my own DCMI XML schema, which contains dc
> >>qualifiers, to assign metadata to electronic documents.
> >>
> >>Here is an abstract :
> >>
> >><dcmi:relation qualifier="isPartOf">Batch55</dcmi:relation>
> >>
> >>The namespace I used up-to-now is "dcmi" and points to "our"
> >>URI ; we will consider shortly revising this to stick to the
> >>actual recommendation, but it we are not yet ready for this.
> >>
> >>I am now facing a case where I'd like to add a new attribute,
> >>say "positionInBatch" to the dcmi:relation element which would give :
> >>
> >>    <dcmi:relation qualifier="isPartOf"
> >>positionInBatch="12">Batch55</dcmi:relation>
> >>
> >>To achieve this, I have created a new schema which declares
> >>the simple type for "positionInBatch", and imported it in my
> >>usual DCMI schema, with its own namespace ("dcmieav" in the
> >>example below).
> >>
> >>The relation element is declared in the dcmi schema which
> >>targetnamespace is "dcmi" and looks now like follows :
> >>
> >><xs:element name="relation">
> >>          <xs:complexType>
> >>            <xs:simpleContent>
> >>              <xs:extension base="Relation">
> >>                <xs:attribute name="positionInBatch"
> >>type="dcmieav:PositionInBatch"/>
> >>              </xs:extension>
> >>            </xs:simpleContent>
> >>          </xs:complexType>
> >>        </xs:element>
> >>
> >> <xs:complexType name="Relation">
> >>    <xs:simpleContent>
> >>      <xs:extension base="xs:string">
> >>        <xs:attribute name="qualifier" type="QualifRelation"
> >>use="optional"/>
> >>      </xs:extension>
> >>    </xs:simpleContent>
> >>  </xs:complexType>
> >>
> >>
> >>  <xs:simpleType name="QualifRelation">
> >>    <xs:restriction base="xs:string">
> >>      <xs:enumeration value="isVersionOf"/>
> >>      <xs:enumeration value="hasVersion"/>
> >>      <xs:enumeration value="isReplacedBy"/>
> >>      <xs:enumeration value="replaces"/>
> >>      <xs:enumeration value="isRequiredBy"/>
> >>      <xs:enumeration value="requires"/>
> >>      <xs:enumeration value="isPartOf"/>
> >>      <xs:enumeration value="hasPart"/>
> >>      <xs:enumeration value="isReferencedBy"/>
> >>      <xs:enumeration value="references"/>
> >>      <xs:enumeration value="isFormatOf"/>
> >>      <xs:enumeration value="hasFormat"/>
> >>    </xs:restriction>
> >>  </xs:simpleType>
> >>
> >>In my instance document, where the DCMI schema is imported, I
> >>can legally have :
> >>
> >>    <dcmi:relation qualifier="isPartOf"
> >>positionInBatch="12">Batch55</dcmi:relation>
> >>
> >>where "qualifier" is of a DCMI namespace and
> >>"positionInBatch" is originally from a "dcmieav" namespace.
> >>However, is this "legal" in a DCMI point-of-view ? Is the
> >>mixing of attributes from namespace other than DCMI allowed ?
> >>Thanks in advance for any help or pointers, Régis Piccand
> >>
> >
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.6 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>
> iD8DBQE9Lr1hPhXvL3Mij+QRAgt2AJ9AyXw4o4Dt2gPbc+EUUxhpLhxlCQCgprjU
> syZM7cX3pcl3dzAdkVOb45g=
> =+Pxv
> -----END PGP SIGNATURE-----
>
>