Cultural heritage portals have the goal of providing users with seamless access to all their resources. This paper introduces initial e orts for a user-oriented restructuring of the German Digital Library (DDB). At present, cultural heritage objects (CHOs) in the DDB are modeled using an extended version of the Europeana Data Model (DDBEDM), which negatively impacts usability and exploration. These challenges can be addressed by leveraging ontologies, and building a knowledge graph from the DDB's voluminous collection. Towards this goal, an alignment of bibliographic metadata from DDB-EDM to FRBR-Aligned Bibliographic Ontology (FaBiO) is presented.
As a Europeana aggregator, the DDB uses the European Data Model (EDM)
to represent CHOs. EDM de nes non-specialized classes and properties to model
the metadata of CHOs. Therefore, diverse types of CHOs ranging from the
conventional (manuscripts) to the highly specialized (wooden skewers), tangible
(bird cage) to the intangible (bird songs), movable (porcelain) to the immovable
(monuments) are modeled without distinction. This limits the portal's
usability. Another drawback is that EDM allows data providers to use their preferred
metadata element sets, unnormalized literals, varied vocabularies and linked
datasets, further adding to the di culty and confusion of the users. By favoring
simplicity and exibility, EDM has failed to model the inherent complexity of
the CHOs and negatively impacted usability [
DDB-EDM. As in EDM, all CHOs are instances of the edm:ProvidedCHO class. For each object, its digitized versions are represented by the class edm:WebResource. In the DDB, digitized items from the libraries are highly granular. A book may be composed of several objects or instances of edm:ProvidedCHO, such as, cover, title page, chapter, section, etc. The distinction of these objects is re ected in the properties dc:type and edm:hasType, from hereon referred to as object type terms. FaBiO. FaBiO is intended \for recording and publishing bibliographic records of scholarly endeavors," and is based on the Functional Requirements for Bibliographic Records (FRBR). In FRBR, a book is not just a single concept but abstracted in several levels using entities connected by a particular relationship. For example, a book containing Friedrich Schiller's rst drama, \Die Rauber " (The Robbers), starts out as a \distinct intellectual or artistic" creation, from hereon referred to as Work (fabio:Work ). The work is linked by fabio:hasRealization to a fabio:Expression. An Expression is a distinct combination of forms to convey the content, in this case, words. When the words are embodied in a physical object, indicated by fabio:hasEmbodiment, the book becomes a Manifestation (fabio:Manifestation), identi ed by its ISBN. The copy of the book one nds on the shelf of the library, identi ed by its call number, is the Item (fabio:Item). A manifestation is linked to its items
DDB-EDM to FaBiO: The Case of the German Digital Library through fabio:hasExemplar. Work, Expression, Manifestation, and Item (WEMI) are collectively referred to as Endeavors. They are expected to improve and simplify retrieval by enabling users to search on speci c levels, e.g. information at expression or manifestation level, rather than the highly granular book parts and copies. This prevents information overload while exploring. FaBiO also provides new relationships and their inverse: (1) fabio:hasManifestation between Work and Manifestation, (2) fabio:hasPortrayal between Work and Item, and (3) fabio:hasRepresentation between Expression and Item.
These relationships simplify representations by allowing omissions, e.g. a rare
book may be described with only Item and Work instances since it has neither
translations (Expression) nor editions (Manifestation). FaBiO de nes extensive
hierarchical sub-classes under each of the endeavors, which allows for
specialization of heterogeneous objects. Moreover, FaBiO may be used in conjunction
with Document Component Ontology (DoCO) to model object granularity.
Alignment. In line with the recommendation of the Europeana working group
on library metadata alignment [
In order to model the heterogeneity of the objects, the object type terms are used as reference for mapping to FaBiO. All terms in the library dataset are assigned to FaBiO endeavor sub-classes manually when applicable. There is no 1to-1 correspondence between a term and a FaBiO class; this is because a term has several conceptual equivalences: document type, document structure, production process, purpose, manifestation, and subject headings. Terms that refer to a document type are matched with existing FaBiO sub-classes on the Work and Expression levels. Those that are not assigned to a speci c Work or Expression sub-class use the default superclass (See \E-Die Rauber" in Figure 2). Terms that refer to production process and manifestation determine the sub-classes under fabio:Manifestation. It is possible for terms to be assigned to multiple sub-classes, for instance, Audiovisuelles Material (Audiovisual Material), which is both a fabio:AudioDocument and fabio:Movie (sub-classes of Expression). DDB CHOs in FaBiO. As an example, the partial mapping of \Dritter Aufzug "7 is shown in Figure 2. The visual representation, including Manifes
tation instances, is hosted on GitHub8. \Dritter Aufzug " is the 3rd act of \Die Rauber ", which is part of \TvFS." The dc:type value of \Dritter Aufzug " is Abschnitt !doco:Section, while \TvFS " is Monogra e!fabio:Book. The dc:type for \Die Rauber " is \Beigefugtes oder enthaltenes Werk " (Contained Work). A fabio:Work instance for \Die Rauber " is created since the title is considered a Work (Werk ) in the German Integrated Authority File aka Gemeinsame Normdatei (GND)9. On the Work level, \TvFS " is designated as a fabio:WorkCollection since it includes a Work (\Die Rauber "). 3
The task of building a KG from the rich and complex collection of the DDB is fraught with challenges that can been addressed by adapting sector-speci c ontologies. This paper introduces the alignment of bibliographic metadata from DDB-EDM to FaBiO to make retrieval simpler and more e cient through the Endeavors, their sub-classes, and additional relationships. This alignment may serve as a reference for other CH portals that are currently using EDM. Rigorous evaluation will be conducted to assess scalability and e ciency of the DDB-KG implementation. Finally, further investigations will be conducted to formulate a top-level ontology enabling interoperability across objects from di erent sectors, making access and exploration of German cultural heritage e ortless for both experts and the general public.