Multi-Layer Active Documents for the Semantic Web

Lanfranchi, Dr. Vita and Ciravegna, Prof. Fabio and Petrelli, Dr. Daniella (2004) Multi-Layer Active Documents for the Semantic Web. In Proceedings Workshop in Interaction Design at the XIII International World Wide Web Conference, New York.

Full text available as:

PDF - Requires Adobe Acrobat Reader or other PDF viewer.

Users’ disorientation and cognitive overload are well known phenomena intrinsic to the idea of hypertext and studied since the early days [3]. The Semantic Web (SWeb in the following) with its layers of annotations can increase the cognitive overload when a document is accessed. As a matter of fact, in a SWeb framework, annotations can be added at different stages in a document lifetime. Initially annotations are added at editing time by the author [11] [12] [13]. Annotations can span from tagging portions of documents with concept labels, to identifying instances or concept mentions, to connect information (e.g. a telephone number and its owner). Different users can annotate the same document in different ways, e.g. using different ontologies, creating different views of the same document. Other annotations can be composed at reading time, i.e. when the document is displayed, for example added by automatic semantic harvesters that extract and integrate information from different repositories [14] [15], or systems which link entities to additional information or services [4]. The many times a document can be annotated and the many pieces of knowledge potentially connected can easily transform a SWeb document into an intricate set of connections. Moreover, the semantic consistence of the annotations (e.g. outgoing links) cannot be guaranteed when different heterogeneous schemas can be applied to the same document. For example, from the author of a document it would be possible to add a hyperlink to reach her diary; from her contact address it is possible to reach the weather forecast for that region. Though both annotations are perfectly acceptable in principle, it is likely that such different navigation choices would distract and disorient the user. This paper proposes to organize the annotations into layers to offer functionalities specific to the user and the context of use as a way to limit the cognitive overload. Managing layers of annotations requires the document to be active. An active document is aware of its own content and can flexibly change the way it presents itself to the actual user, e.g. by allowing the user to read only up to a predefined level of detail. However an active document should not be limited to the presentation phase: we extend the activity to the annotation layer as well. A global framework for editing and accessing multi-layered semantically active documents is proposed.

Contact the site administrator at: hg@ecs.soton.ac.uk