PUB - Publikationen an der Universität Bielefeld

Solutions for efficient querying of long-term human-robot interaction data require in-depth knowledge of the involved domains and represents a very difficult and error prone task due to the inherent (system) complexity. Developers require detailed knowledge with respect to the different underlying data schemata, semantic mappings, and, most importantly, the query language used by the storage system (e.g. SPARQL, SQL, or general-purpose language interfaces/APIs). While for instance database developers are familiar with technical aspects of query languages, application developers of interactive scenarios typically lack the specific knowledge to efficiently work with complex database management systems. Addressing this gap, in this paper we describe a model-driven software development based approach to create a long-term storage system to be employed in the domain of embodied interaction in smart environments (EISE). To support this, we created multiple domain specific languages using Jetbrains MPS to model the high level EISE domain, to represent the employed graph query language Cypher and to perform necessary model-to-model transformations. As main result, we present the EISE Query-Designer, a fully integrated workbench to facilitate data storage and retrieval by supporting and guiding developers in the query design process and allowing direct query execution without the need to have prior in-depth knowledge of the domain at hand. In this paper we report in detail on the study design, execution, first knowledge gained from our experiments, and lastly the lessons learned from the development process up to this point.