FLORIDA STATE UNIVERSITYCollege of Communication and InformationSchool of Library and Information Studies

Representation Patterns for Cultural Heritage ResourcesCreating Linked Data Patterns for Libraries, Archives, and Museums

Dr. Richard J. Urban – rurban@fsu.edu – http://orcid.org/0000-0003-2817-4590

SolutionDesign patterns – optimal solutions to common problems – are useful tools used by developers for software engineering, interface design (Figure 4), ontology development, and Linked Open Data modeling (Figure 3) (Gamma, et al., 1995; Blomqvist, Gangemi, & Presutti, 2009; Dodds & Davis, 2011; Gangemi, 2005; van Harmelen, ten Teije, & Wache, 2011). Although the library, archive, and museum (LAM) domain frequently uses concrete examples in standards documentation, these examples lack important features which make design patterns useful. In addition to providing solutions, design patterns serve an important function by identifying and articulating common problems. By doing so, design patterns create a shared technical lexicon around which designers, developers, and creators can structure their conversations (Dearden & Finlay, 2006). Because design patterns make problems, their contexts, and solutions explicit, they can serve as important educational tools for students and novices (Chatzigeorgiou, Tsantalis, & Deligiannis, 2008). Design pattern languages are also capable of expressing patterns at different scales and in ways that build relationships among patterns (Alexander, 1977).

Methode Linked Open Data for Libraries, Archives, and Museums (LODLAM) Patterns project seeks to establish a pattern library for cultural heritage Linked Open Data. A Mediawiki site has been established as a sandbox to organize "protopatterns" that emerge from a content analysis of contemporary metadata standards. Within a particular standard, we are asking:

• what representation features are present in the standard? • what problem does this feature try to solve? • what contexts/forces make this feature relevant? • how does this feature resemble other features observed in other standards? (does it constitute a new pattern or is it an exemplar of one that has already been de�ned?)

Because the LODLAM community values openness and transparency, this research blends content analysis with a distributed participatory design methodology (Danielsson, Naghsh, Gumm, & Warr, 2008). Members of the LODLAM community are invited to contribute to the identi�cation of problems and solutions, iterative re�nement of found patterns, and the ongoing organization/classi�cation of patterns.

http://lodlampatterns.org/protopattern/Surrogate IdentityProblemHow can I distinguish between metadata about an original resource and metadata about a surrogate that stands in for that resource?

ContextCultural heritage repositories contain surrogate representations of resources they hold in their collections (i.e., a digital image that depicts a painting). Some document-based data management patterns may con�ate these entities, resulting in confusing, incoherent metadata (Hutt & Riley, 2005).

SolutionCultural heritage data models should explicitly include surrogate resource classes that can be identi�ed independently of the resource a surrogate instance represents. Models should specify the relationship between a resource and its surrogate(s).

Related Patterns• One graph per resource (Dodds and Davis, 2011)

Examples• Categories for the Description of Works of Art (CDWA) (Work/Related Textual or Visual Documentation)• Dublin Core Abstract Model (1:1 Principle)• Europeana Data Model (EDM) • Requirement 1: distinction between "provided objects" (painting, book, movie, archeology site, archival �le, etc.) and their digital representations. • Requirement 2: distinction between objects and metadata describing the object. • VRA Record Type (Work/Image)

ContextLinked Open Data (LOD) is an important new paradigm for de�ning representation standards for cultural heritage resources. Unlike methods that require the adoption of a single primary representation standard, LOD principles encourage the reuse and extension of existing models. Rather than building new representation standards from the ground up, leading examples of cultural heritage Linked Data creatively remix general LOD models with existing cultural heritage standards (see Figure 2). Because data modeling is a sociotechnical process as much as it is a set of technical problems, Churchill (2012) has called on researchers to bring HCI approaches and methods to bear on data design problems.

Figure 1: Seeing Standards: A Visualization of the Metadata Universe (Riley & Becker, 2010)

Figure 3: Graph Per Resource Pattern (Dodds & Davis, 2011)

Figure 4: Yahoo! Design Pattern Library (http://developer.yahoo.com/ypatterns/)

Figure 2: British National Bibliography Linked Data Model (http://www.bl.uk/bibliographic/natbib.html)

e ProblemCurrently, librarians, archivists, and museum professionals can choose from a large universe of representation standards (see Figure 1). Each of these standards exhibits various strengths and weaknesses based on the problems they are engineered to address. Unfortunately, standards developers do not always explicitly articulate the problems or the contexts that shaped a particular solution. Although Greenberg (2005) provides a way to classify standards according to their domain, objectives, and architecture, there is no mechanism to identify and organize the features found within a standard.

Future WorkLODLAM Patterns aims to create a new way to organize our debates and discussions about representation standards for cultural heritage resources that is oriented around problems and solutions rather than standards and schemas. As LAM professionals navigate the paradigm shift from traditional representation methods to Linked Data and Semantic Web contexts, LODLAM Patterns can also provide a useful tool to crosswalk current domain knowledge.

ReferencesAlexander, C., Ishikawa, S., & Silverstein, M. (1977). A pattern language: Towns, buildings, construction. New York: Oxford University Press.Blomqvist, E., Gangemi, A., & Presutti, V. (2009). Experiments on pattern-based ontology design. In Proceedings of the �fth international conference on knowledge capture (pp. 41–48). New York, NY: ACM.Chatzigeorgiou, A., Tsantalis, N., & Deligiannis, I. (2008). An empirical study on students’ ability to comprehend design patterns. Computers & Education, 51(3), 1007–1016. Churchill, E. (2012). From data divination to data-aware design. Interactions, 19(5), 10–13.Danielsson, K., Naghsh, A. M., Gumm, D., & Warr, A. (2008). Distributed participatory design. In CHI’08 Extended Abstracts on Human Factors in Computing Systems (pp. 3953–3956). New York, NY: ACM.Dearden, A., & Finlay, J. (2006). Pattern languages in HCI: A critical review. Human–Computer Interaction, 21(1), 49–102.Dodds, L., & Davis, I. (2011). Linked Data patterns: A pattern catalogue for modelling, publishing, and consuming Linked Data. Retrieved from: http://patterns.dataincubator.org/

Gamma, E., Helm, R., Johnson, R., & Vlissides, J. (1995). Design patterns : Elements of reusable object-oriented software. Reading, Mass.: Addison-Wesley.Gangemi, A. (2005). Ontology design patterns for Semantic Web content. In Y. Gil, E. Motta, V. Benjamins, & M. Musen (Eds.), e Semantic Web – ISWC 2005 (Vol. 3729, pp. 262–276). Springer Berlin/Heidelberg. Retrieved from http://www.springerlink.com/content/f513071t4477h2w2Greenberg, J. (2005). Understanding metadata and metadata schemes. Cataloging & Classi�cation Quarterly, 40(3-4), 17–36.Riley, J., & Becker, D. (2010). Seeing Standards: A visualization of the metadata universe. Indiana University Libraries. Retrieved from http://www.dlib.indiana.edu/~jenlrile/metadatamap/van Harmelen, F., ten Teije, A., & Wache, H. (2011). Knowledge engineering rediscovered: Towards reasoning patterns for the semantic web. In D. Fensel (Ed.), Foundations for the web of information and services: A review of 20 years of semantic web research (pp. 57–74). New York: Springer.