Share on facebook
Save on delicious
RSS Feed
Usability & Facilitating Access
We are investigating how to make Semantic Grid tools accessible to social scientists by creating natural language interfaces for query formulation and information presentation. As users require no training to understand their native languages (unlike a complex graphical interface), a natural language interface makes a tool accessible not only for experts, but for a range of other users including policy stakeholders. Natural language also has very sophisticated ways to convey information about time, quantification etc., which can be hard to present graphically. The short-term aim is to build a WYSIWYM (What You See Is What You Meant: Power et al., 1998) interface that supports both the formulation of queries and the editing of ontologies, using natural language. In a WYSIWYM application, the user edits on the semantic level by means of an intermediate feedback text, which is filled with expandable anchors. Attached to the anchors are menus; when a menu-option is chosen, the semantic representation of the query is updated and a new feedback text is generated. WYSIWYM interfaces have a number of advantages: the user receives feedback in natural language, which affords the advantages listed above; in addition, the interface can process all formulated queries, as it is impossible to specify a malformed query.
To determine the type of metadata that social scientists might use to describe their publications and data, we have completed an analysis of the resources managed by the UK Data Archive. An ontology (see above) has been created based on the UKDA metadata model, to allow social science data sets and other resources (e.g. publications) to be characterised; you can access the ontology here: [Policy Grid Ontologies]. This ontology will be used to drive the formulation of queries in the prototype WYSIWYM interface currently under development.