Ontopoetry Ontology facilitates a set of concepts for the description of poetic works (poems, poetic drama / plays written in verse and songs). It is the product of a homogenization effort that considered different literary traditions, periods, poetic genres and authorship. Additionally, this will enable in the future the comparison of the characteristics and data in this poetry, and thus to carry out invaluable research in Comparative Literature and Comparative Metrical Studies in a quantitative fashion. 

Given the excellent coverage of the Postdata Ontology V1.0 and after a critical analysis of Postdata Ontology V1.0, made by experts on ontology engineering and poetry, it was concluded that it was cumbersome to manage a single artefact as the result of the development tasks. Therefore, a redefinition of the subdomains would improve the semantic description of the ontology, its interoperability, and its ease of use. To solve this weakness, we opt for re-modularisation with the objective to ease the maintenance and publication of the ontology [1]. Modularisation helps during the re-engineering and alignment tasks as engineers are able to work in parallel and not being overwhelmed with big structures or great number of ontological elements. It also helps users to better understand the ontology and to have a more flexibility deciding which modules to reuse [2]. 

While we can find some algorithms that exploit the structural characteristics of ontologies to divide them into modules, we based our approach on the identification of relevant use cases for the potential users of OntoPoetry Ontology. This is achieved after a previous analysis of the relevant applications or objectives of the ontology, and by enforcing encapsulation and independence as modularisation criteria. Encapsulation involves the identification of elements that are related to a given sub-domain so that a module could be easily exchanged by another, while independence seeks the identification of modules that are self-contained and of potential reuse [2]. 

This analysis yielded the definition of two potential cases of use used to define the main OntoPoetry Ontology subdomains (i.e. modules):

• Bibliographic information search and indexing: tackling the requirements of the poetry scholar community researches involves the intensive usage of bibliographic information, mainly coming from the librarian domain. This information is used to differentiate the different conceptual and physical properties of the works yielding to the definition of two modules:
  • OntoPoetry Core module: to represent the abstract or conceptual side of the bibliographic information. Abstract entities of works are required to differentiate between the original idea of a work and its different expressions. repositories work with abstract notions of works that can be used to better differentiate between different editions or versions, and to know more about the intention of the authors. Relevant classes tackled by this module are Poeticwork, Ensemble and Redaction. Therefore, this module includes all the basic information that characterizes works, irrespective of their physical materializations.
  • OntoPoetry Transmission module: to represent the physical side of bibliographic information related to poetic works. It extends the descriptions of Postdata Core module by relating the different physical editions and items to the abstract domain. Some relevant classes at this level include Bibliographic Source and Apparatus. Hence, it provides ontological elements to describe physical editions or manifestations (like a published book or a manuscript).
• Poetic information annotation and searching
  • OntoPoetry Poetic Analysis module: this module is oriented to the representation of different phenomena associated to metrics and prosody, including the textual elements or parts of a poem, and the different metrical patterns that analyse those elements. Relevant classes related to this sub-domain include Line, Stanza, Rhyme, Foot and Syllable. The ontology allows for metrical information annotation using scansion symbols for each line, but also by describing each element’s analysis at metrical syllable, foot or morae level. In addition, we also include means to describe literary elements such as the presence of figures of speech.

Fig. 1 . Encapsulation of OntoPoetry Ontology

Therefore, OntoPoetry Ontology comprised of three ontology modules, each of them developed and serialized independently following OWL and RDF standards, Fig. 1. 

Therefore, each module covers a specific sub-domain as much as possible, so we expect the ontology modules to exhibit a high module cohesion (degree of relatedness between the ontological elements defined in the same module). Another desired property of the ontology modules is related to the logical quality. In a modularisation scenario, we must ensure two properties for each module so that the complete set of ontologies is safe (“they do not produce unexpected results such as new inconsistencies or subsumptions between imported symbols” [3]). 

• Local Correctness of the module: Ensure that any module does not entail any additional fact for the elements defined by the other modules than those entailed by the element’s module. 
• Local Completeness of the module: ensure that all the entailments of the global ontology about the elements of interest of a module are also preserved in this module’s entailments.  

Modules Core and Transmission are founded on FRBRoo model, while Poetic Analysis is based on the selection of appropriate Ontology Design Patterns (ODPs). It also includes a set of controlled vocabularies or concept schemes that help users to better categorise or classify each of the elements described by the ontological modules at a deeper level than the one provided by the defined classes. This is a very useful step towards the consolidation of different poetic traditions as each tradition may consider different classifications (i.e. types of stanzas and lines, types of stress and scansion technique used, objective of the poem, main theme of the poem, etc.). While a complete analysis of all the possible concepts relevant for each tradition has not been thoroughly conducted, we have identified those aspects that vary across traditions [4]. Then, the main contribution in this sense is the identification of discrepancies and the provision of initial concepts that serve as example to guide the inclusion of further concepts as long as the project evolves. This means that most of the vocabularies are open to editions and subjected to modifications after deployment of the Postdata Knowledge Graph. In his work for sake of space ,we will introduce OntoPoetry Core module. 

[1] M. d’Aquin, A. Schlicht, H. Stuckenschmidt, and M. Sabou, “Criteria and Evaluation for Ontology Modularization Techniques,” in Modular Ontologies, vol. 5445, H. Stuckenschmidt, C. Parent, and S. Spaccapietra, Eds. Berlin, Heidelberg: Springer Berlin Heidelberg, 2009, pp. 67–89. doi: 10.1007/978-3-642-01907-4_4. 

[2] P. Doran, V. Tamma, and L. Iannone, “Ontology Module Extraction for Ontology Reuse: An Ontology Engineering Perspective,” in Proceedings of the Sixteenth ACM Conference on Conference on Information and Knowledge Management, New York, NY, USA, 2007, pp. 61–70. doi: 10.1145/1321440.1321451. 

[3] B. C. Grau, “A Logical Framework for Modularity of Ontologies.,” pp. 298–303, 2007. 

[4] H. Bermúdez Sabel, “Towards interoperability in the European poetry community,” presented at the Shaping Data in Digital Humanities, Denmark, Apr. 20, 2018. doi: 10.5281/zenodo.2020302.