Contextual Informatics

The “Contextual Informatics” workgroup of the Heinz Nixdorf Institute supports the research discourse of both art and economic historians by developing a web based working environment. The main challenge in this endeavor is to support the input process into the emerging historical database as soon as possible, while at the same time providing means to establish semantic contexts, and then also allowing search queries based on the latter in a most flexible way. The developed formal database scheme should allow, but not enforce object relations without restricting definitions.

The preparation of the diverse sources with their various formats and representations is followed by the actual database entry and storage, and subsequently by an ongoing analysis together with a corresponding research discourse. While supporting all that, the system enables hermeneutic precision work in the sense of a “Close Reading” process, which possibly leads to a corresponding increase of understanding, that would not have been achievable by structured data entry and full-text search alone. Finally, the gained research findings should be concentrated and visualized, to present them to the interested public online.

At first, a classical information scientific approach was taken, with a strong focus on the work object. The goal was the implementation of a project specific database from scratch, based on an object hierarchy that had been determined analytically beforehand. However, this strategy turned out to be impractical, as less obvious structures and semantic contexts sometimes unfolded only while the project progressed. Resulting from the interdisciplinary research discourse, repeated alterations to the object model became necessary. Ultimately, this led to a change of perspective, now foregrounding the work process: The database was then realized based on an initially generic object structure, while specific object properties and relations between different object types were added subsequently.

To realize this approach, choice fell on the Open Source Software platform Omeka, which had previously proven to be a reliable tool within the cultural scientific milieu. Its unique properties distinctly differ from the usual web based content management systems: Besides providing the capability to present the input content in a public web frontend, Omeka allows the structured storage of different, immanently generic object types and object relations between them. Additionally, it is possible to perform in-depth semantic analyses of the data pool, and ultimately presenting the result thereof on the web.

Rooting on the sometimes fuzzy or even contradictory sources and likewise open references, a data model was developed as the result of the interdisciplinary dialog between the Cultural Sciences and the Contextual Informatics. It defines objects that are meaningful within the research process (e.g. buildings, persons, means of transportation, etc.), which can then be enriched with digitized sources. This multi-stage process revealed a fundamental problem: While the principal object hierarchy was easy enough to comprehend, the need for a finer object granularity with a deeper level of semantic context often arose only as late as during the test entry of example data. More than once these cases derived from immanently indistinct cultural-scientific research-theoretical aspects (e.g. the comparative inspection of imprecisely defined calendar dates and time spans, given either in the Gregorian or the Julian calendar). While their significance was indeed implicitly known, explicit observation unveiled that they were difficult to model inside a usually precisely specified information scientific canon. 

Because of the use of common web technologies within an Open Source context, Omeka can be expanded with reasonable effort by additional functions after an in-depth analysis of any problem that might arise. This especially applies to extended and refined search functions that consequentially exceed simple full-text search on a semantic level in many ways. Based on the structurally edited data and search results, subsequently it is possible to verify existing and as well newly formulated research hypotheses by systematically evaluating the search results towards the identified objects and their relations among them.

So the project contribution of the Contextual Informatics goes significantly beyond being merely the “software supplier”: As part of the interdisciplinary dialog with the cultural sciences, a refined structured view of the topic area arose that had not been obvious before. Apart from that, innovative input methods and analytical tools emerged, without which the conception and validation of novel research hypotheses would not have been feasible.