- From: Winfried Mühl <muehl@mail.sub.uni-goettingen.de>
- Date: Wed Feb 23 17:51:59 2000
- To: www-rdf-interest@w3.org
(Appologies for crossposting!) the Goettingen Digitization Center (GDZ) of the Lower Saxony State and University Library, a national supply Center for Digitization in German libraries, defined and applies an XML/RDF format for their digitized documents (http://www.sub.uni-goettingen.de/gdz/en/gdz_main_en.html#dms) first in July 1998. At present we are finishing a final proposal for the "edoc" format of the european project DIEPER (http://www.sub.uni-goettingen.de/gdz/dieper/) which aims to build a virtual network and a central access point to make accessible periodicals that have been retrospectively digitised in Europe. This format extends the older XML/RDF metadata format with fulltext in TEI/XML. As we want at least the RDF part to be compatible with the "DC world" we are trying to take into account the latest recommendation, especially that of the "Guidance on expressing the Dublin Core within the Resource Description Framework (RDF)" [GDCRDF] (http://www.ukoln.ac.uk/metadata/resources/dc/datamodel/WD-dc- rdf/). But we are not quite shure in some questions concerning the appropriate mapping of DC elements in RDF. 1) The DIEPER project will use a document identifier for each document structure (e.g. volume, issue, article). Using basic RDF syntax the about attribute in the <rdf:description> element should be used to identify a document structure to be described: ------------------------------- <dieper:docstrct> <rdf:Description about="dieper_identifier"> . . . </rdf:Description> <dieper:docstrct> ------------------------------- Looking through the DC spectacles we instead could introduce a subelement <dc:identifier> within the <rdf:description> element as we would do it for other DC elements: ------------------------------- <dieper:docstrct> <rdf:Description> <dc:identifier>dieper_identifier</dc:identifier> . . . </rdf:Description> <dieper:docstrct> ------------------------------- What do you think would be appropriate? 2) We would like to introduce DC element qualifiers for the role of a creator or contributor. The Guidance (GDCRDF) is proposing dcq:creatorType and dcq:contributorType for the element qualifiers. According to the RDF example 10 (GDCRDF) we would propose the syntax recommended by you using a DIEPER list for qualifier values, e.g. for a creator: ---------------------------------------------------- <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:dc = "http://purl.org/dc/elements/1.0/" xmlns:dcq = "http://purl.org/dc/qualifiers/1.0/" xmlns:dct = "http://purl.org/dc/terms/1.0/creator/type/"> xmlns:DIEPER = "http://www.sub.uni- goettingen.de/gdz/dieper/edocformat/"> xmlns:DIEPERQS = "http://www.sub.uni- goettingen.de/gdz/dieper/edocformat/qualifierschema/"> . . . <dc:creator> <rdf:Description> <rdf:value>Newton, Isaac</rdf:value> <dcq:creatorType> <DIEPER:Author /> </dcq:creatorType> <dcq:creatorScheme> <DIEPERQS:fnf /> </dcq:creatorScheme> </rdf:Description> </dc:creator> . . . ---------------------------------------------------- (The <DIEPERQS:fnf> scheme qualifier would be optional. ) Is this o.k.? The DC working group on agent qualifiers is proposing qualifiers like Agent Role. Using RDF this would result in substituting the <dcq:creatorType> element with <dcq:agentRole> in the above example. What would be up to date here? Would it be allowed to use the abbreviated typed node RDF syntax by implicit introducing <rdf:type>dc:creator</rdf:type> in the above <rdf:description> element? E.g. ---------------------------------------------------- . . . <rdf:Description> <rdf:type>dc:creator</rdf:type> <rdf:value>Newton, Isaac</rdf:value> <dcq:creatorType> <DIEPER:Author /> </dcq:creatorType> </rdf:Description> . . . ---------------------------------------------------- would turn into ---------------------------------------------------- . . . <dc:creator> <rdf:value>Newton, Isaac</rdf:value> <dcq:creatorType> <DIEPER:Author /> </dcq:creatorType> </dc:creator> . . . ---------------------------------------------------- And what would be the dumbed down version for it without qualifiers? Simply kicking off some elements like this?: ---------------------------------------------------- . . . <dc:creator> <rdf:Description> <rdf:value>Newton, Isaac</rdf:value> </rdf:Description> </dc:creator> . . . ---------------------------------------------------- or a version again using typed node syntax like: ---------------------------------------------------- <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" . . . <dc:creator>Newton, Isaac</dc:creator> . . . ---------------------------------------------------- ?? Following is a more elaborated example describing document structures by mixing elements from different namespaces as well as basic with abbreviated typed node syntax. We use for element qualifier values items from a DIEPER list, because we have the feeling that the proposals from the differnet DC working groups are still under development. Further on we use the dc:title element for the main title, the subtitle and a special main title format for sorting, distinguishing them all by element qualifiers. We specify the relation of parent document structures and their children just by nesting the children. The <DIEPER.docstrct> element is a typed node element. Its description may contain the related child document structures as an structure "describing" element. We think building up the hierarchical book structure as an element tree in this way is better human readable then declaring all parent child relationships by dc:relation or some other elements in distinct doc elements, but is this still conform to RDF and DC ?? Please have a look at the details: ---------------------------------------------------- <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:dc = "http://purl.org/dc/elements/1.0/" xmlns:dcq = "http://purl.org/dc/qualifiers/1.0/" xmlns:dct = "http://purl.org/dc/terms/1.0/creator/type/"> xmlns:DIEPER = "http://www.sub.uni- goettingen.de/gdz/dieper/edocformat/"> xmlns:DIEPERQS = "http://www.sub.uni-goettingen.de/gdz/dieper/edocformat/qualifierschema/"> . . . <DIEPER:docstrct DIEPER:type="PeriodicalVolume"> <dc:identifier>dieper_identifier_for_the_volume</dc:identifier> <dc:creator> <rdf:Description> <rdf:value>lastname, firstname</rdf:value> <dcq:creatorType><DIEPER:Author /></dcq:creatorType> </rdf:Description> </dc:creator> <dc:contributor> <rdf:description> <rdf:value>the_editor_if_present</rdf:value> <dcq:contributorType><DIEPER:Editor/></dcq:contributorType> </rdf:description> </dc:contributor> <dc:title> <rdf:Description> <rdf:value>any main title </rdf:value> <dcq:titleType><DIEPER:MainTitle/></dcq:titleType> </rdf:Description> </dc:title> <dc:title> <rdf:Description> <rdf:value>any subtitle</rdf:value> <dcq:titleType><DIEPER:SubTitle/></dcq:titleType> </rdf:Description> </dc:title> <dc:date> <rdf:Description> <rdf:value>1920</rdf_value> <dcq:dateType><DIEPER:PublicationYear/></dcq:dateType> </rdf:Description> </dc:date> <dc:publisher>the publisher if known</dc:publisher> <dc:language> <rdf:Description> <rdf:value>any language</rdf:value> <dcq:languageScheme>RFC1766</dcq:languageScheme> </rdf:Description> </dc:language> <DIEPER:PlaceOfPublication>the_place_of_publication</DIEPER:PlaceOfPublication> . . . <!-- the first child document structure is following --> <DIEPER:docstrct DIEPER:type="Titlepage"> <dc:identifier>dieper_identifier_for_the_volume</dc:identifier> . . . </DIEPER:docstrct DIEPER:type="Titlepage"> . . . <!-- the first document substructure is following --> <DIEPER:docstrct DIEPER:type="Issue"> <dc:identifier>dieper_identifier_for_the_issue</dc:identifier> . . . <!-- the first article is following --> <DIEPER:docstrct DIEPER:type="Article"> <dc:identifier>dieper_identifier_for_the_article</dc:identifier> . . . </DIEPER:docstrct> <!-- end of first article --> . . . </DIEPER:docstrct> <!-- end of first issue --> . . . </DIEPER:docstrct> ---------------------------------------------------- So, that are more then enough lines and questions! Hopefully interesting enough that you have had a look at some details. Of course we appreciate any helpfull hints on it! Thanks from Winfried Muehl ============================================== ========================================================================== Winfried Mühl Tel: +49 551/395686 Niedersächsische Staats- und Fax: +49 551/392361 Universitätsbibliothek Göttingen email: muehl@mail.sub.uni-goettingen.de Göttinger DigitalisierungsZentrum GDZ (Altbau) Papendiek 14, 37073 Göttingen ==========================================================================
Received on Wednesday, 23 February 2000 17:51:59 UTC