Sharing Epigraphic Information as Linked Data

Volume Editors Salvador Sánchez-Alonso Universidad de Alcalá, Edificio Politécnico despacho O-246, Ctra. Meco s/n, 28871 Alcalá de Henares, Spain E-mail: [email protected] Ioannis N. Athanasiadis IDSIA/USI-SUPSI, Galleria 2, 6928 Manno, Lugano, Switzerland E-mail: [email protected]

Library of Congress Control Number: 2010936639 CR Subject Classification (1998): H.4, H.3, I.2, H.2.8, H.5, D.2.1, C.2 ISSN ISBN-10 ISBN-13

1865-0929 3-642-16551-6 Springer Berlin Heidelberg New York 978-3-642-16551-1 Springer Berlin Heidelberg New York

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting, reproduction on microfilms or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. springer.com © Springer-Verlag Berlin Heidelberg 2010 Printed in Germany Typesetting: Camera-ready by author, data conversion by Scientific Publishing Services, Chennai, India Printed on acid-free paper SPIN: 06/3180 543210

Table of Contents

Bridging Scales and Paradigms in Natural Systems Modeling . . . . . . . . . . Ferdinando Villa

1

Analyzing Hidden Semantics in Social Bookmarking of Open Educational Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Juli` a Minguill´ on

8

Case Studies of Ecological Integrative Information Systems: The Luquillo and Sevilleta Information Management Systems . . . . . . . . . . . . . . Inigo San Gil, Marshall White, Eda Melendez, and Kristin Vanderbilt

18

Agrotags – A Tagging Scheme for Agricultural Digital Objects . . . . . . . . . Venkataraman Balaji, Meeta Bagga Bhatia, Rishi Kumar, Lavanya Kiran Neelam, Sabitha Panja, Tadinada Vankata Prabhakar, Rahul Samaddar, Bharati Soogareddy, Asil Gerard Sylvester, and Vimlesh Yadav

36

Application Profiling for Rural Communities: eGov Services and Training Resources in Rural Inclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pantelis Karamolegkos, Axel Maroudas, and Nikos Manouselis

46

Developing a Diagnosis Aiding Ontology Based on Hysteroscopy Image Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Marios Poulos and Nikolaos Korfiatis

57

Utilizing Embedded Semantics for User-Driven Design of Pervasive Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ahmet Soylu, Felix M¨ odritscher, and Patrick De Causmaecker

63

Utilizing Linked Open Data Sources for Automatic Generation of Semantic Metadata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Antti Nummiaho, Sari Vainikainen, and Magnus Melin

78

Application of Semantic Tagging to Generate Superimposed Information on a Digital Encyclopedia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Piedad Garrido, Jesus Tramullas, and Francisco J. Martinez

84

Mapping of Core Components Based e-Business Standards into Ontology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ivan Magdaleni´c, Boris Vrdoljak, and Markus Schatten

95

Model-Driven Knowledge-Based Development of Expected Answer Type Taxonomies for Restricted Domain Question Answering . . . . . . . . . . Katia Vila, Jose-Norberto Maz´ on, Antonio Ferr´ andez, and Jos´e M. G´ omez

107

XII

Table of Contents

Using a Semantic Wiki for Documentation Management in Very Small Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vincent Ribaud and Philippe Saliou A Short Communication – Meta Data and Semantics the Industry Interface: What Does the Food Industry Think Are Necessary Elements for Exchange? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kathryn A.-M. Donnelly

119

131

Social Ontology Documentation for Knowledge Externalization . . . . . . . . Gonzalo A. Aranda-Corral, Joaqu´ın Borrego-D´ıaz, and Antonio Jim´enez-Mavillard

137

Information Enrichment Using TaToo’s Semantic Framework . . . . . . . . . . Gerald Schimak, Andrea E. Rizzoli, Giuseppe Avellino, Tomas Pariente Lobo, Jos´e Maria Fuentes, and Ioannis N. Athanasiadis

149

Exploiting CReP for Knowledge Retrieval and Use in Complex Domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lorenza Manenti and Fabio Sartori

160

Quality Requirements of Migration Metadata in Long-Term Digital Preservation Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Feng Luan, Thomas Mestl, and Mads Nyg˚ ard

172

A Model for Integration and Interlinking of Idea Management Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adam Westerski, Carlos A. Iglesias, and Fernando Tapia Rico

183

An Enterprise Ontology Building the Bases for Automatic Metadata Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Barbara Th¨ onssen

195

Matching SKOS Thesauri for Spatial Data Infrastructures . . . . . . . . . . . . . Cristiano Fugazza, Soeren Dupke, and Lorenzino Vaccari

211

Sharing Epigraphic Information as Linked Data . . . . . . . . . . . . . . . . . . . . . . ´ Fernando-Luis Alvarez, Elena Garc´ıa-Barriocanal, and Joaqu´ın-L. G´ omez-Pantoja

222

Development Issues on Linked Data Weblog Enrichment . . . . . . . . . . . . . . Iv´ an Ruiz-Rube, Carlos M. Cornejo, Juan Manuel Dodero, and Vicente M. Garc´ıa

235

On Modeling Research Work for Describing and Filtering Scientific Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ´ Miguel-Angel Sicilia Localisation Standards and Metadata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dimitra Anastasiou and Lucia Morado V´ azquez

247 255

Table of Contents

The Design of an Automated Workflow for Metadata Generation . . . . . . . Miguel Manso-Callejo, M´ onica Wachowicz, and Miguel Bernab´e-Poveda Assessing Quality of Data Standards: Framework and Illustration Using XBRL GAAP Taxonomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hongwei Zhu and Harris Wu Brazilian Proposal for Agent-Based Learning Objects Metadata Standard - OBAA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rosa Maria Vicari, Alexandre Ribeiro, J´ ulia Marques Carvalho da Silva, Elder Rizzon Santos, Tiago Primo, and Marta Bez Towards Quality Measures for Evaluating Thesauri . . . . . . . . . . . . . . . . . . . Daniel Kless and Simon Milton Enriching the Description of Learning Resources on Disaster Risk Reduction in the Agricultural Domain: An Ontological Approach . . . . . . Thomas Zschocke, Juan Carlos Villagr´ an de Le´ on, and Jan Beniest Descriptive Analysis of Learning Object Material Types in MERLOT . . . Cristian Cechinel, Salvador S´ anchez-Alonso, ´ Miguel-Angel Sicilia, and Merisandra Cˆ ortes de Mattos Quality in Learning Objects: Evaluating Compliance with Metadata Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Christian Vidal, N. Alejandra Segura, S. Pedro Campos, and Salvador S´ anchez-Alonso

XIII

275

288

300

312

320 331

342

The Benefits and Future of Standards: Metadata and Beyond . . . . . . . . . . Christian M. Stracke

354

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

363

Sharing Epigraphic Information as Linked Data Fernando-Luis Álvarez1, Elena García-Barriocanal1, and Joaquín-L. Gómez-Pantoja2 1

Information Engineering Research Unit Computer Science Dept., University of Alcalá Ctra. Barcelona km. 33.6 – 28871 Alcalá de Henares (Madrid), Spain {fernandol.alvarez,elena.garciab}@uah.es 2 Dep. Of History.- University of Alcalá C/ Colegios 2, E-28801 Alcalá de Henares (Madrid), Spain [email protected]

Abstract. The diffusion of epigraphic data has evolved in the last years from printed catalogues to indexed digital databases shared through the Web. Recently, the open EpiDoc specifications have resulted in an XML-based schema for the interchange of ancient texts that uses XSLT to render typographic representations. However, these schemas and representation systems are still not providing a way to encode computational semantics and semantic relations between pieces of epigraphic data. This paper sketches an approach to bring these semantics into an EpiDoc based schema using the Ontology Web Language (OWL) and following the principles and methods of information sharing known as “linked data”. The paper describes the general principles of the OWL mapping of the EpiDoc schema and how epigraphic data can be shared in RDF format via dereferenceable URIs that can be used to build advanced search, visualization and analysis systems. Keywords: Epigraphy, EpiDoc, OWL, linked data, Semantic Web.

1 Introduction The digital representation of information has become a target for modern society, especially for the preservation and dissemination of cultural heritage objects, even though recent measures show that many digitization projects are still waiting to be realized (Poll, 2010). An epigraph is a text cut or scratched in a hard surface as stone, metal or pottery. Text could go from few letters to the length of an actual booklet. These elements were mostly cut for permanence and durability, hence its use as public records for legal or government regulations as well as cultic altars and private tombstones; at the same time, there are instances of impromptu writing, such as graffiti, ownership labels and the like. Epigraphy is a welcome source when documenting ancient cultures, since they provide actual samples of how long-gone languages as Gaulish or Punic were written and spelled. For better known Greek and Latin, inscriptions are valued as a witness for words, idioms, personal and place names not attested in any other historical sources as well as testimony of everyday life: family relationships shows prominently on S. Sánchez-Alonso and I.N. Athanasiadis (Eds.): MTSR 2010, CCIS 108, pp. 222–234, 2010. © Springer-Verlag Berlin Heidelberg 2010

Sharing Epigraphic Information as Linked Data

223

tombstones as well common people life expectancy; military gravestones attest the deployment of the Army all over the Roman Empire and milestones bear witness on its extended and well cared road network; and the actual wording of most Roman laws is known because copies of them were cut in bronze tables hung on public buildings; Digital epigraphy projects have been for years considered as a key technique for the study and dissemination of ancient texts (Manuelian, 1998). A number of both public and private institutions have sponsored the development and operation of epigraphic databases which will allow scholars or the general public to access the documented inscriptions in Roman provinces from any Web browser, making research easier and quicker for a number of scholars (Pasqualis, 2005). For example, Crowther (2002) describes the need for scanning documents and the creation of large databases, and Alkhoven (2005) stresses the importance of training. Recently, the open EpiDoc specification (Cayless, 2003) has been developed and several systems using this proposed standard have already been deployed (Bodard, 2008). An EpiDoc file is a representation in XML of the edition of one inscription or a group of inscriptions. At a minimum the file will contain a text in Greek or Latin, probably with editorial siglae. It may also contain apparatus, translation, commentary, place of finding, description and dating of the text or object, among several other information elements that are normally published in a scholarly edition. However, EpiDoc is not providing a way to encode computational semantics, as it is relying mostly on structured metadata with text fields. Such computational semantics would provide epigraphic information the added value that could be exploited in Semantic Web applications. Also, they would provide the basis for sharing epigraphical information following the ideas of open linked data (Berners-Lee, 2006), which opens new possibilities for automated analysis, search or visualization applications. It should be noted that the linked data approach has been considered an important element in the agenda of the European digital library Europeana1. This paper reports preliminary work on providing computational semantics to epigraphic information by using ontology representations and sharing these enhanced information representations following the linked data approach. The rest of this paper is structured as follows. Section 2 provides background information on digital epigraphy projects and proposed schemas specific to the domain of ancient texts and museums. Then, Section 3 describes the rationale and design of an EpiDoc-based ontology expressed in W3C OWL. Section 4 sketches the design to sharing RDF-based representations of epigraphic information following the open linked data approach. Finally, conclusions and outlook are provided in Section 5.

2 Background In past years, ancient Epigraphy has jumped wholeheartedly into the digital bandwagon and the number of operative databases and data repositories keeps growing2. Such an enthusiasm arises from historical, economical and convenience reasons. First 1 2

Europeana White Paper No. 1: Knowledge = Information in Context by Prof. Stefan Gradmann, available at http://version1.europeana.eu/web/europeana-project/whitepapers For a selection of those dealing with Latin inscriptions, see: http://eda-bea.es/, http://www.manfredclauss.de/

224

F.-L. Álvarez, E. García-Barriocanal, and J.-L. Gómez-Pantoja

of all, there is a long tradition (going back as farther as the XVIth. century) to collect and edit inscriptions in great corpora or collections put together according geographical, chronological or thematic criteria. But those project are always expensive, both in terms of scholarly work and in the cost of publishing specialized volumes which need to be brought up to date often and target a restricted group of readers; furthermore, data recovery and searching on printed copy are tricky proposition, because compiling indexes by hand is such a cumbersome task that limit the number of search-key allowed. An unexpected benefit of digital databases, even the simpler ones, is that they naturally provided notable searching capabilities, while the volumes from which data was taken often lack of indexes.3 Most current epigraphic repositories are based on relational databases (Harl and Schaller, 2004), and in general searching in these systems becomes a complex task. A first important problem is that they the objects were monuments made for reading and seeing; hence, it is important to provide a detailed description of both its decoration and the place where they were set up. Second, the text could have written in Latin, Greek of any other combination of both or with other scripts like Punic, Celtiberian or Egyptian and there are not established standards for mixing scripts in a regular searching pattern. And, last, any word in an inscription may have different relevance from a scholar’s point of view.

Fig. 1. An example simple Search in the EDH Portal

3

For a detailed account on computers and Epigraphy see Gómez-Pantoja, forthcoming (Gómez-Pantoja, 2010).

Sharing Epigraphic Information as Linked Data

225

To overcome those difficulties, existing databases allow for complex search patterns, as the search mask (Figure 1) of Epigraphische Datenbank Heidelberg (EDH)4 shows. Figure 1 shows a fragment of the data needed to search for a particular inscription or set of inscriptions. Figure 2 shows a query formulated in Clauss database5. As it can be seen, they are completely different, although the results of the query depicted will include the same inscription as the query in Figure 1. Although the output in these cases is the same (as shown in Figure 3), the search is performed using different search fields and with different strings. Relational databases allow linking data from a variety of tables to establish relationships that make it possible to relate the required information. The problem is that in current databases, there is not a normalization of key entities as people or places, but they are provided as free text in metadata fields. These complications hampers efficient data analysis for which a full normalization of main information entities is required.

Fig. 2. An example simple search on Clauss/Slaby Data Base

Fig. 3. Result of queries in Figure 1 and Figure 2 4 5

http://www.uni-heidelberg.de/institute/sonst/adw/edh/index.html http://oracle-vm.ku-eichstaett.de:8888/epigr/epigraphik_en

226

F.-L. Álvarez, E. García-Barriocanal, and J.-L. Gómez-Pantoja

A first step to solve the problems of normalization of digital information in epigraphy was proposed at the end of the 90s by Elliott6. He promoted a new standard (EpiDoc)7, to represent and store epigraphic information. EpiDoc is not a database and has not rules of association; each element (inscription), is described with a XML file formatted following conventions based on the TEI guidelines8. EpiDoc uses XSLT to display stored items. So, through an XSLT template, the browser transforms XML into an HTML page that returns the required information. There are several examples of EpiDoc-based systems available on line, including Inscriptions of Aphrodisias9 (Reynolds, 1982; Bodard, 2008) and Vindolanda Tablets Online10 (Bowman, 1994; Bowman, 1983; Terras, 2006). Some of the advantages of this design are portability – as it is based on XML – ease of management – as each inscription is placed as an independent file. In consequence, it can be said that “the use of structured mark-up, like XML, increases precision in location and retrieval of content and opens up the possibility of reuse.” (Breure, 2006). However, representing information in XML files is only practical for small collections, and by itself it does not provide further entity normalization. Advancing towards increased normalization of information entities can be achieved by engineering an ontology taking as point of departure the schema and guidelines in EpiDoc. There are examples of standarized ontologies of cultural contents, such as CIDOC CRM11, which are valid for a general and wide cultural context, however none of them come close enough to the specificities of epigraphy to be usable without extension. Developing an ontological schema that allows for more normalized information structure has the added benefit of preparing epigraphic data to be shared on the Web via linked data, which opens new possibilities to relating information currently dispersed in several databases.

3 Representing EpiDoc in OWL The semantic representation for epigraphic information reported here is based on a mapping of the EpiDoc schema to an ontological representation expressed in the W3C OWL language. The mapping is not only a change in format (from EpiDoc XML to OWL) that would be rather straightforward, but a re-engineering of the scheme trying to represent all the important entities as OWL classes. Each concrete information element then becomes an instance of an Inscription concept, which relates to other concepts as Individual, Material or Letters. A series of properties are used to formally representing the links, e.g. has_civilization, whose domain will being Inscription and its range Civilization will establish the relationship among the instances of both concepts. It should be noted that this adds possibilities to map with other knowledge organization systems. For example, civilization appears in the 6

http://homepages.nyu.edu/~te20/ http://EpiDoc.sourceforge.net/ 8 http://www.tei-c.org/index.xml 9 http://insaph.kcl.ac.uk/iaph2007/index.html 10 http://vindolanda.csad.ox.ac.uk/ 11 http://www.cidoc-crm.org/ 7

Sharing Epigraphic Information as Linked Data

227

Art&Architecture Thesaurus12 of the Getty foundation associated with the facet . Even though the AAT does not provide a terminology for civilizations many art styles and periods can be easily mapped to them, providing a way to relate epigraphic information to them. EpiDoc has adopted some controlled vocabularies, e.g. the “EAGLE/EpiDoc Object Type Vocabulary” describing types of monuments and objects bearing inscribed texts. These vocabularies have been translated to ontology modules reusable separately, enriching them with subsumption and mereological predicates when appropriate. For example, the following vocabulary elements are mereologically related: columna (column) has as parts columna, basis (column base), columna, imoscapus (column shaft, lower) and columna, summoscapus (column shaft, upper). Also, in the EpiDoc header section , is intended to provide classificatory and contextual information about the text, such as its subject matter, the situation in which it was produced or the individuals described by or participating in producing it. This opens many possibilities for encoding semantically rich information as the Civilization mentioned above, or the key persons referred in the inscription. Endowing to each inscription, a series of properties and inference rules allows more precise search and relating information. This way, for example, if we look for “M Messio Abascanto”, we can carry out a textual search and return only those inscriptions containing that character string. But if we use a search in our ontology, we can differentiate the division of the selection “tria nomina”, and to look for the nomen, praenomen, cognomen or filiation. Also, genealogical trees or geographical distribution of names can be used to infer related information. Continuing the same example, roman citizens, that is, those who possessed all the rights such as voting in the city of Rome elections, received three names. The praenomen was given at the birth to the child in something like the actual name, which coincided with his father's name or a ancestor’s one. The nomen referred to the gens or family. The cognomen was a sort of nickname by which the individual was known. In out representation, each part of the nomen is a data-type associated to the inscription, as well as the filiation. This way, each one will have assigned a series of labels as attributes and properties, which will give them useful semantics for the recovery of the information. The following rule infers gens affiliations between citizens named in the inscriptions. Thus, if the reasoner finds two individuals with the same nomen, it can establish a relationship between them. Inscription(?i1) ^ has_person(?i1, ?p1) ^ Inscription(?i2) ^ has_person(?i2, ?p2) ^ has_nomen(?p1,?n1) ^ has_nomen(?p2,?n2) ^ sameAs(?n1, ?n2) → has_same_gens(?i1, ?i2)

As an example we will sketch in what follows the OWL mapping for the inscription AE 1987,63513 to illustrate the differences between an EpiDoc XML file and the corresponsing OWL representation. The inscription is showed in Figure 4.

12 13

http://www.getty.edu/research/conducting_research/vocabularies/aat/ http://edh-www.adw.uni-heidelberg.de/EDH/inschrift/012116

228

F.-L. Álvarez, E. García-Barriocanal, and J.-L. Gómez-Pantoja

Fig. 4. The AE 1987,00635 inscription The text that can be read on it is: “DIS MAN M MESSIO ABASCANTO SEGONTINO IVLIA SCINTILA MARITO PIENTISSIMO ET SIBI”

The Transcription (following an Standard notation) is: Dis Man(ibus) / M(arco) Messio Abascanto Seguntino Iulia Scinti[l]la marito / pientissimo et sibi

Which can be translated as: “To the sacred spirits of the departed. Marco Messio Abascanto of Segontia. Iulia Scintilla for her loving husband and for herself”. The following is an extract of a corresponding description in a XML / EpiDoc file . . . UAH - 2009 . . .