Listening to the 14th LIBREAS podcast about EuropeanaConnect (and reading the transcript here – my apologies if you don’t read or speak German, I’ll summarize some important points as I go along), I couldn’t help but make the connection to Topic Maps as a viable solution to what Europeana is trying to achieve (the site is currently in beta status and the project phase is scheduled to end in October 2011).
But let’s start at the beginning; Europeana is a networked platform and portal providing European cultural institutions (museums, libraries, archives, galleries and audiovisual archives) with the possibility to present their documents. It’s important to note that Europeana is really only a metadata aggregation overlay, pointing the user to the respective institution’s website to view the item; the digital / digitized objects stay with the institutions and are not stored in Europeana itself.
In his white paper, Stefan Gradmann explains the Europeana infrastructure, using terms like matching and merging which also occur fairly often in Topic Maps speak. The “Paris” example (distinguishing between the city and the mythical figure and offering the respective resources to the user) is one that could also be realized using Topic Maps by typing the topics in question (type “city” and type “mythical figure”, for instance, thus solving the homonym problem) and by leveraging the rich associations between topics that create a web of knowledge and discovery.
There are several work packages dealing with different aspects, EuropeanaConnect being in charge of the technical implementation. Work packages 1 and 2, Creating the Europeana Semantic Layer and Multilingual Access to Content, are particularly interesting for everyone concerned with semantics and metadata. As described in the podcast, part of the process is linking up the different vocabularies of the original data using SKOS. So, content in the form of heterogeneous data from heterogeneous sources in heterogeneous formats about heterogeneous objects (museum, library and archival material) – feeling dizzy yet? – is pulled in and processed to form the information layer. Topic Maps could be used to create the knowledge layer, organizing the gathered information into topics and associations and linking it back to the content / information layer with the help of occurrences. To aggregate data in varying schemas, Topic Maps don’t have to squeeze them into one “unified” schema, rather the formats and models can sit side by side without necessarily having to be migrated.
There’s another level where topic maps could play out their potential: the front-end, the interface the users get to see and where they query the site. As reported in the podcast, one of the goals of Europeana is to create an interface where users can select their language and also do keyword searches in that language. This requires an underlying multilingual thesaurus with merged subject headings that can cater to various linguistic communities. With the concept of scope, topic maps have a powerful means to provide information by using the query language to determine in which language to display subject headings and results to the user.
One of the most significant strengths of topic maps is the fact that many pieces of information from different sources can be clustered around a topic, and that linguistic contexts can be accounted for by scope on language. Isn’t that a perfect fit for a project accommodating such a multitude of digital cultural heritage objects and linguistic backgrounds?