> DCXMLRevision/Comments

The following 575 words could not be found in the dictionary of 550 words (including 550 LocalSpellingWords) and are highlighted below:

A2   abbreviate   abbreviation   abbreviations   ac   acceptable   accommodate   accompanied   accomplish   accordingly   action   Add   add   adopt   affect   all   allow   allowing   allows   Also   also   always   amemded   Amend   amend   amended   an   An   Anchor   and   andy   Andy   Ann   annotated   annotating   any   applicable   application   applications   applied   approach   appropriate   approved   Apps   Architecture   architecture   architecturewiki   are   Are   aren   as   assumption   at   atomic   attribute   attributes   Attributes   available   avoid   avoided   aware   axml   backwards   Baker   base   based   be   bearing   because   been   being   best   better   between   bin   Board   both   Brickley   But   but   By   by   can   Can   captured   cases   cf   change   changes   child   choices   chose   citable   cite   clarification   clarifying   class   collaborating   com   comment   Comment   Comments   comments   comp   compatible   complete   component   components   confirmed   confusing   confusion   considerably   consideration   considered   consistent   constitutes   constraints   constructs   contain   content   context   contexts   continue   convention   copies   Core   correctly   corresponding   could   cover   creation   Creator   current   currently   Dan   data   datatype   datatypes   date   Date   Davis   dc   dcmitype   decided   decision   decisions   declaration   declarations   default   definition   depending   depends   describes   description   Description   design   designer   dev   difference   different   discussing   discussion   distinction   Do   Doctype   Document   document   documentation   documents   does   Draft   draft   drop   Dublin   dublincore   duplicate   during   each   earlier   Eduserv   eduserv   element   Element   Elements   elements   elsewhere   emphasise   encoded   encoding   ensure   entity   enumerated   enumerations   equally   Essentially   etc   evolve   example   Example   Examples   examples   existing   expanded   experimentation   explicitly   Expressing   Extended   extended   Farrenkopf   features   ffffcc   Finalise   finalised   Finished   first   Fischer   for   form   format   former   Foundation   from   full   general   given   guaranteed   guidance   Guidelines   handle   has   have   helpful   highlight   How   However   human   Ian   Id   Identifier   if   If   illustrate   illustrating   imcompatible   Implementers   imply   in   In   include   included   incompatible   inconsistent   ind0606   ind0607   indicating   infosystems   instances   Instances   intended   intent   internal   interpret   interpretation   interpreted   introduce   introduced   is   Is   Issue   issue   Issued   it   It   itself   jiscmail   johnston   Johnston   know   La   language   Language   languagers   languages   largely   Latest   limited   link   lists   Literal   literal   make   makes   mapped   mapping   marc2   materials   Means   means   mechanism   mechanisms   metadata   might   mind   most   Multiple   multiple   must   name   Name   nameaspace   names   Names   namespace   Namespace   Namespaces   natural   need   new   news   nnnn   No   no   non   Not   not   note   Note   notes   notion   now   object   objects   obviously   occurs   of   omission   on   One   one   only   optional   options   or   order   ordering   oriented   other   part   particular   particularly   Path   period   persistent   pete   Pete   point   pointer   possibilities   possible   potential   Powell   powell   ppp   pre   preferable   prefix   Prefix   presence   preserved   probable   probably   process   processor   properties   property   Property   Proposed   provide   provision   public   published   purely   purl   purposes   qualified   query   quesion   quite   raised   range   ranges   rather   rdfs   read   reader   readers   Received   received   Redundant   Ref   references   referencing   reflect   reflects   relative   remains   Remove   removed   repeated   Replaced   Replaces   represent   representable   representation   representations   Representations   represented   representing   require   required   requirement   requires   requiring   Resolution   resolved   Resource   Response   result   reverse   revising   Revision   Rich   rich   root   same   scenarios   Schema   schema   schemas   Schematron   scheme   section   see   See   seems   select   Self   semantics   sense   Separate   separate   separation   servers   set   Should   should   signal   simplifies   single   site   So   so   some   sometimes   space   specification   specified   specifies   specify   stamped   standard   statement   statements   Stefan   string   String   strings   structure   stylistic   sub   subset   such   Suggest   suggest   suitable   summarises   support   syntax   tablebgcolor   tablewidth   Tag   tag   Te   term   text   than   that   The   the   theaimsgroup   their   then   There   there   these   they   third   this   This   Thomas   though   three   times   Title   title   to   To   Tom   transform   treated   tree   true   two   type   Type   types   Types   uk   understand   unexpectedly   unix   unprefixed   up   Usage   use   Use   used   usefulness   uses   using   validation   Value   value   values   Version   version   versioned   vocabulary   W3   way   we   Web   webadmin   well   were   what   What   when   Where   where   wherever   Whether   whether   which   will   with   within   work   working   Working   Would   would   xerces   xsd   zero  

Clear message

Comments Received on "Expressing Dublin Core metadata using XML" Working Draft 2006-05-29

This document is part of the [Self]DC Architecture Wiki.

IMPORTANT: Do not cite materials in this Wiki other than for the purposes of collaborating on document creation. This Wiki is intended to be used to work on draft copies of documents. Finished documents will be published, in a persistent and citable form, on the dublincore.org Web site (or elsewhere in some cases).

Title: Comments Received on "Expressing Dublin Core metadata using XML" Working Draft 2006-05-29
Creator: Pete Johnston, Eduserv Foundation <pete.johnston@eduserv.org.uk>
Creator: Andy Powell, Eduserv Foundation <andy.powell@eduserv.org.uk>
Date Issued: 2006-06-30
Identifier: http://dublincore.org/architecturewiki/DCXMLRevision/Comments/2006-06-30
Replaces: Not applicable
Is Replaced By: Not applicable
Latest Version: http://dublincore.org/architecturewiki/DCXMLRevision/Comments
Description of Document: This document summarises comments on the 2006-05-29 Working Draft of "Expressing Dublin Core metadata using XML" received during the public comment period from 2006-05-30 to 2006-06-27.

Issue 1: Resource URIs and Value URIs not representable as QNames

Ann Apps, 2006-06-12 1

Comment

If DC-XML allows the representation of other URIs as QNames, it should allow the representation of Resource URIs and Value URIs as QNames.

Response

The DC-XML format is probably inconsistent on allowing QName representations for VES and SES URIs but not for Resource URIs and Value URIs.

Note also that Property URIs can only be represented as QNames (so there are some Property URIs which can not be represented in DC-XML).

Resolution

See general discussion of using XML QNames as abbreviations for URIs.

Issue 2: Use of XML entity references (1)

Ann Apps, 2006-06-12 1

Comment

The example illustrating the use of XML entity references to abbreviate URIs is confusing to a human reader.

Response

The use of XML entity references is part of the XML standard. Examples of their use were included to illustrate that there are mechanisms other than the QName-URI mapping that can be used to abbreviate URIs in XML content.

Resolution

No action required.

Issue 3: Use of XML entity references (2)

Ann Apps, 2006-06-12 1

Comment

The use of XML entity references (where the entity declarations are in the internal subset) may not be acceptable to an XML Schema processor and/or the presence of a DOCTYPE declaration may imply a requirement for DTD validation.

Response

An XML processor must read the internal subset and use (some of) the declarations in it.

Also the presence of a DOCTYPE declaration is not a signal for DTD validation.

See

http://www.xml.com/axml/notes/DoctypeMeans.html

and

http://marc2.theaimsgroup.com/?l=xerces-j-dev&m=111643969503776

The XSV XML Schema processor correctly uses the entity declarations in the internal subset.

Resolution

No action required.

Issue 4: Redundant duplicate example

Ann Apps, 2006-06-12 1

Comment

The content of example 7 and example 8 is the same. One could be removed.

Response

Example 7 is annotated to highlight particular constructs, but there is no need to duplicate the example.

Resolution

Remove one example and amend references in text accordingly.

Issue 5: Separate attributes for URIs and QNames-representing-URIs

Ann Apps, 2006-06-12 1

Thomas Fischer, 2006-07-13 8

Comment

Are separate attributes required for URIs and QNames-representing-URIs?

Response

If DC-XML is to support both URIs in full and QNames-as-abbreviations of URIs, an application has to know the datatype of the text in order to process it correctly.

An attribute value "name" may be interpreted as

attribute value)

In the attribute value "ppp:nnnn", the string "ppp" may be interpreted as

In the attribute value "news:comp.infosystems.www.servers.unix", the string "news" may be interpreted as

Resolution

No action required, though may be helpful to include examples of relative URIs and unprefixed QNames to emphasise point.

See general discussion of using XML QNames as abbreviations for URIs.

Issue 6: Multiple statements v multiple value strings

Ann Apps, 2006-06-12 1

Comment

What is the difference between:

Response

There are two different scenarios, depending on whether the value is

Whether both options are available for a property depends on the range of the property. If the range of dc:title is (rdfs:)Literal, then the first form would not be appropriate; if the range of dc:title is the class of "natural language objects", then it would be.

The Usage Board is working on clarifying the ranges of the DCMI properties; their decisions will be accompanied by guidance documentation.

Resolution

Add note indicating that example of dc:title with multiple value strings reflects an assumption that that is consistent with the range of the property.

Issue 7: Multiple Value URIs

Thomas Fischer, 2006-07-13 8

Comment

How does DC-XML support the provision of multiple value URIs in a statement?

Response

The DCAM specifies that a statement can contain at most one value URI, and DC-XML reflects this.

Resolution

No action required.

Issue 8: Rich Representations (1)

Ann Apps, 2006-06-12 1

Comment

Do rich representations include instances of non-XML text markup? Is a LaTeX text a rich representation or a value string? If the former then DC-XML does not accommodate it.

Response

In part, this requires clarification in the a DCAM to specify what constitutes a rich representation and what constitutes a value string. But it seems probable that the intent is that a LaTeX text should be treated as a rich representation.

If so, there remains the quesion of whether it is a requirement for DC-XML to support that (sub)class of rich representation.

Resolution

(To be decided)

Issue 9: Rich Representations (2)

Thomas Fischer, 2006-07-13 8

Comment

How is the MIME type for a rich representation encoded?

Response

DC-XML does not support the provision of a MIME type for rich representations, because this is not currently specified by the DCAM. However this has been raised as a possible omission from the DCAM.

Resolution

This is a DCAM issue. If the DCAM is amended to support MIME types for rich representations, then DC-XML will be amemded accordingly.

Issue 10: Name of descriptionRef attribute

Ann Apps, 2006-06-12 1

Comment

Would it be better to name the attribute descriptionIdRef, rather than descriptionRef (cf ID and IDREF datatypes)?

Response

Resolution

(To be decided)

Issue 11: URI used for XML Namespace Name (1)

Tom Baker, 2006-06-15 2

Comment

If the URI http://purl.org/dc/xml/ is being used purely as an XML Namespace Name and not as a "base URI" for a set of term URIs used in an RDF vocabulary, then it may be preferable to use a different form of URI, to avoid any potential for confusion on the part of human readers.

Response

The URI is an XML Namespace Name and there is no intent to suggest that there is a corresponding set of URIs with this URI as root. A different URI could be used.

Resolution

Use a http://dublincore.org/... URI as an XML Namespace Name

Issue 12: URI used for XML Namespace Name (2)

Pete Johnston, 2006-06-15 3 4

Comment

Should the XML Namespace Name be "versioned"?

Response

It is quite possible that the XML format will evolve. If a change is backwards compatible (e.g. a new optional XML attribute), then there is probably no need to change the XML Namespace Name, but backwards incompatible changes probably should result in a change.

Implementers should always explicitly select the version they are using, so that backwards-imcompatible changes at the "namespace document" aren't introduced unexpectedly.

Resolution

Use a "date-stamped" http://dublincore.org/... URI as an XML Namespace Name

Issue 13: Use of enumerated lists of in dcmitype.xsd

Stefan Farrenkopf, 2006-06-06 6

Comment

How to make use of the enumerations in dcmitype.xsd?

Response

The enumerations were part of earlier experimentation and are of limited usefulness for validation of attribute values given the current format and schema design, though see section 3.4.1 of [Self]DC-XML and W3C XML Schema for an example of use with the Value String Element.

Resolution

Remove QName and URI enumerations for dcmitype.xsd?

See also general discussion of using XML QNames as abbreviations for URIs.

Issue 14: XML Elements v XML Attributes

Thomas Fischer, 2006-07-13 8

Comment

In general in XML, attributes are used to provide data which is used to interpret or understand the content of the XML element. Is it appropriate to use XML attributes for value URIs and vocabulary encoding scheme URIs?

Response

In the context of "document-oriented" XML markup languages, there is sometimes a convention of using attributes for data which is considered to be "not part of the document content", particularly where the language is used as a markup language in the sense of "annotating" a pre-existing text.

However XML itself makes no distinction between the "semantics" of attributes and elements. Essentially wherever a format designer chose to use an XML attribute, an XML child element could have been used. The reverse is not true, because attributes of the same name can not be repeated on a single XML element and because there is no ordering of XML attributes - and obviously attribute values are "atomic" and if the child element itself has child elements that "sub-tree" structure can't be captured in a single attribute value.

So in DC-XML, the choices between using XML elements and XML attributes have been based only on the constraints on the structure of a DC description set specified by the DCAM. Where a sub-component occurs zero or one times (e.g. value URI or vocabulary encoding scheme URI in statement, syntax encoding scheme or language tag on value string), then an XML attribute has been used; where a sub-component is repeated (e.g. value string or rich representation within statement) then a child XML element has been used.

But the former set of components could equally well have been represented using child XML elements. It is largely a stylistic decision (though the features of schema languagers and query languages may also be a consideration.)

Resolution

No action required.

Issue 15: GRDDL/XSLT

Dan Brickley, 2006-06-01 5

Comment

Can we set up namespace document with pointer to GRDDL transform?

Response

A nameaspace document with a suitable link to a transform will be set up when the DC-XML format is finalised, the DC-RDF mapping is finalised, the XML Namespace Name is confirmed, and the XSLT transform is complete.

Resolution

Finalise XSLT. Extended version(s) of transform for current DC-XML are now available for XSLT/XPath 1.0 and XSLT/XPath 2.0.

http://dublincore.org/architecturewiki/DCXMLRevision/DCXMLXSLT/2006-06-01

Extended set of example instances at

http://dublincore.org/architecturewiki/DCXMLRevision/DCXMLInstances/2006-06-01

Issue 16: XML QNames in content

Pete Johnston, 2006-06-14 7, referencing comments by Ian Davis on use of QNames in RDF/A

Comment

The Namespaces in XML specification describes the use of QNames as names of XML elements and XML attributes. It does not cover the use of XML QNames as XML element content or XML attribute values.

A "non-datatype-aware" XML processor (e.g. an XSLT transform) might handle XML Namespaces in such a way that the interpretation of QNames in XML attribute values or XML element content is not guaranteed to be preserved, and so the use of XML QNames in these contexts is probably best avoided.

Response

There are three possibilities

  1. drop the URI abbreviation mechanism, and always represent URIs in full (bearing in mind that the use of XML entity references is always available

  2. continue to support the use of XML QNames to represent URIs, but ensure that all uses are for the names of XML elements or XML attributes

  3. introduce a namespace/qualified name mechanism for which the semantics is specified by the DC-XML format itself

Resolution

Suggest revising DC-XML to adopt the third approach and drop any use of XML QNames in XML element content and XML attribute values. See

http://dublincore.org/architecturewiki/DCXMLRevision/DCXMLGuidelines/2006-07-04

Note that this also considerably simplifies the use of XML Schema for the format, though it also means that applications requiring extended validation of DC-XML content will require mechanisms other than XML Schema to accomplish that (e.g. XSLT, Schematron etc)

Issue 17: Proposed changes to DCAM

Comment

The DC Architecture WG is discussing some changes to the DCAM, two of which would affect DC-XML

Response

DC-XML is based on DCAM so would need to be amended/extended to reflect DCAM changes (and also to ensure that examples reflect DCMI's intended use of URIs, e.g. for VES v Value Types).

Resolution

(If DCAM changes approved) add new attribute for Value Type; amend Value String description. Amend examples. Amend schemas. Amend XSLT.

[1] http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0606&L=dc-architecture&P=3254

[2] http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0606&L=dc-architecture&P=4300

[3] http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0606&L=dc-architecture&P=4427

[4] http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0606&L=dc-architecture&P=4703

[5] http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0606&L=dc-architecture&P=252

[6] http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0606&L=dc-architecture&P=622

[7] http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0606&L=dc-architecture&P=3878

[8] http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0607&L=dc-architecture&P=303