E Pluribus Unum: a Uniform DL Solution for Historical Data Management, Archiving and Exploitation of Opera

Among the performing arts, opera is undoubtedly the one that presents the greatest complexity, since it is made up of a variety of 'texts' that differ in nature (literary, musical, choreographic and visual), but which are closely coordinated.
As a result, documents relating to opera come in multiple forms (verbal, musical, iconographic, sound, and visual documents) and cover the entire life cycle of an opera: the verbal texts (librettos), the musical score, materials for the opera’s performance in the theatre (contracts, scene sketches, etc.), and documents that relate to the performance itself (audio and video recordings, stage photos, reviews) (Bianconi et al., 2004). The conservation of these heterogeneous sources is usually entrusted to different institutions (libraries, archives, museums) according to the type of material (books, administrative documents, artistic objects) and catalogued using methodologies that vary according to the type of material. The documentation relating to an opera is thus to be found in different institutions and is described in different information systems which are often not compatible. The material is therefore not always easily accessible for consultation.
Our initial approach was to adopt Functional Requirements for Bibliographic Records (FRBR) (IFLA Study Group, 1998) as a conceptual solution for a unified management of documents and their contents (Pompilio et al., 2005). The result is a highly specialized model in which information is collated and analysed in depth. But, to fulfil the functional perspective prescribed by FRBR, cataloguing must take account of knowledge of the domain, and the interests and expertise of different categories of users.
One of the aims of the Corago LOD

The project web site address is: http://corago.unibo.it/lod
project was thus to demonstrate that by adopting DL it would be possible to express domain knowledge at different descriptive levels within a single coherent semantic model (Bonora and Pompilio, 2019).

The FRBR’s conceptual hierarchy is defined by the three core classes: Work, Expression and Manifestation. In the domain of opera, these three classes are represented by the headings Opera, Event and Document, which become entry points for accessing information about repertoire, chronology and documentary sources.

Diagram 1- From Works to Performances and Librettos
The path that leads from the conceptual content of operas to its two expressions, namely performances and librettos, can be expressed using the FRBR Object Oriented (FRBRoo) model. Moreover, by transposing knowledge from the original relational model to RDF, using CIDOC Conceptual Reference Model (CIDOC CRM) and FRBRoo as main reference ontologies (Bekiari et al., 2015), it was possible to overcome constraints deriving from the original implementation of FRBR by using a proprietary model such as data interoperability (Le Boef, 2012).
Access to information contained in catalogues, archives or directories has often been limited by a lack of interoperability provided by the information system used. From the users’ point of view, a knowledge based facility such as Linked Open Data (LOD) lowers the access threshold because alignment is at a basic semantic level (Bizer et al., 2009). Increased interoperability also leads to a wider dissemination and circulation of information. As this information becomes part of the knowledge held in other systems, its availability would be guaranteed even beyond the capabilities or aims of the original project. As a result, LOD interoperability both facilitates "universal" access to digital resources and is a viable strategy for preserving information in the long term. Conversely, the adoption of formal ontologies often produces data structures that are complex and cannot be browsed in a user-friendly way.
The project faced two major factors of complexity introduced by RDF data representation: the deeper level of analysis and the morphological intricacy of the graph. The analyticity is directly linked to the fine-grained accuracy of the ontology adopted for the definition of the dataset, while the complexity of the graph structure derives both from the fine-grained conceptual model and the amount of information that the repository contains. As a result, providing synthetic and effective representations of data to end users becomes challenging.
The solution presented exploits some basic mechanisms of the DL such as generalization and inheritance. Through generalization it is possible to build progressive levels of abstraction on top of basic RDF triples. These intermediate layers act as semantic lenses which are capable of focusing the functional requirements expressed by users.
While the traditional design of cataloguing applications based on RDBMS requires that data extraction and representation tasks follow a business logic, the transition to semantic models allows intermediation to take place on a conceptual level. An intermediate layer can be formally introduced within the semantic model of the repository to provide a user-friendly representation of the domain entities. This layer would be designed by the domain expert according to user needs.

Diagram 2 - Multi layer semantic architecture
The same approach has been adopted with regard to access to domain knowledge. The goal in this case is to make information from diverse sources accessible through search criteria which start go from plain text level up to criteria specific to the field of music. The aim of the project was also to identify semantic references between results and search criteria. We realised this goal by introducing a full-text semantic index within the dataset. Indexing is based on a specific class of properties that identifies the relationships between the core entities of the domain and terms that would be used to search each class of entity. The system can thus provide the reference context that produced the results displayed. This allows users to navigate the results in a more controlled and accurate way. This kind of semantic indexing becomes a tool for managing both the complexity of the domain, and for making with information more accessible in a distributed environment such as the LOD ecosystem.
Finally, in our experience, identifying the layers of abstraction used when searching and navigating knowledge constitutes a kind of conceptual modelling which, since it is part of the semantic model of the knowledge base, also becomes part of the competence of the domain. In other words, modelling the way opera history is being accessed and represented becomes a possible way of reducing the intrinsic complexity of the domain as experienced by users.

Bibliography

Bartolini, I., Luzzi, C., Patella, M. and Pompilio, A. (2012). MIMESIS: A Semantic-based Approach for a Digital Library of.
Poetry in Music. in Proceedings ECLAP 2012: Conference on Information Technologies for Performing Arts. Media Access and Entertainment: Florence 7-9. Firenze University Press.

Bekiari, C., Doerr, M. and Le Boef, P. (2015).
FRBRobject-Oriented Definition and Mapping from FRBRER. FRAD and FRSAD (Version.

Bellini, P. and Nesi, P. (2014).
Modeling Performing Arts Metadata and Relationships in Content Service for Institutions. Multimedia Systems. Springer doi:10.1007/s00530-014-0366-0.

Berners-Lee, T., Hendler, J., Lassila, O. and others (2001). The semantic web.
Scientific American, 284: 28–37.

Bianconi, L., Pompilio, A. and Pagannone, G. (2004). RADAMES: prototipo d’un repertorio e archivio digitale per il melodramma.
Il Saggiatore Musicale. Firenze, Olschki edition.

Bizer, Heath and Berners-Lee (2009). Linked Data - The Story So Far.
International Journal on Semantic Web and Information Systems. Special.

Bonora, P., Ossicini, C. and Raffa, G. (2006). From Relational Metadata Standards to CRM Ontology: a Case Study in Performing.
Arts. CIDOC2006 Proceedings.

Bonora, P. and Pompilio, A. (2019). ‘Osservate, leggete con me’. Risorse LOD per la storia del melodramma: una prospettiva funzionale di rappresentazione. Udine http://hdl.handle.net/11585/677025.

Crawford, T., Fields, B., Lewis, D. and Page, K. (2014). Explorations in Linked Data practice for early music corpora.
Proceedings of the 14th ACM/IEEE-CS Joint Conference on Digital Libraries.IEEE. Press.

Doerr, M. (2003). The CIDOC conceptual reference module: an ontological approach to semantic interoperability of metadata.
AAAI. AI Magazine, 24 doi:DOI: http://dx.doi.org/10.1609/aimag.v24i3.1720.

Doerr, M. and Le Boef, P. (2007). Modelling Intellectual Processes: The FRBR-CRM Harmonization.
Digital Libraries: Research & Development. First International DELOS Conference. Springer Verlag, pp. 114–123.

Doty, C. (2013). The Difficulty of An Ontology of Live Performance. InterActions.
UCLA Journal of Education and Information Studies.

Guerrini, M. and Possemato, T. (2015).
Linked data per biblioteche. archive e musei. Editrice Bibliografica.

IFLA Study Group (1998).
Functional Requirements for Bibliographic Records. (IFLA Section on Cataloguing). K . G. Saur München.

Kobilarov, G., Scott, T., Raimond, Y., Oliver, S., Sizemore, C., Smethurst, M., Bizer, C. and Lee, R. (2009). Media Meets Semantic Web – How the BBC Uses DBpedia and Linked Data to Make Connections. In Aroyo, L., Traverso, P., Ciravegna, F., Cimiano, P., Heath, T., Hyvönen, E., Mizoguchi, R., Oren, E., Sabou, M. and Simperl, E. (eds),
The Semantic Web: Research and Applications. Springer Berlin Heidelberg, pp. 723–37.

Le Boef, P. (2012).
Towards Performing Arts Information As Linked Data?. In SIBMAS 2012 Conference: Best Practice Innovative Techniques for Performing Arts Collections. Libraries and Museums. France.

Lembo, D., Rosati, R., Ruzzi, M., Savo, D. F. and Tocci, E. (2014). Visualization and Management of Mappings in Ontology-based Data Access.
DL 2014 27th International Workshop on Description Logics. CEUR Workshop Proceedings, vol. Vol. p. 1193.

Pompilio, A., Bianconi, L., Regazzi, F. and Bonora, P. (2005). RADAMES: A new management approach to opera: Repertory. archives and related documents.
Proceedings - First International Conference on Automated Production of Cross Media Content for Multi-Channel Distribution doi:DOI: 10.1109/AXMEDIS.2005.42.

Weigl, D. M., Page, K. R., Lewis, D., Crawford, T. and Knopke, I. (2015). Unifiedaccess to media industry and academic datasets: a case study in Early Music.