This paper discusses the architecture and implementation of /OnQ/, a web-based survey database system. The /OnQ/ system is an outgrowth of the US National Science Foundation funded project, Survey2001. Survey2001, the second, large-scale survey project hosted by the National Geographic Society, was a complex questionnaire designed to measure the impact of information technology on conservation, community and culture. It also served as the development platform for /OnQ/. Delivered in four different languages, Survey2001 illustrates many of the features of the /OnQ/ system, including complex skip patterns, non-text question prompts, calls to outside databases and the robustness of a Java based implementation.
In this paper we specifically focus on several behind the scenes features of the /OnQ/ system. First, we detail the database architecture, which is built around the logic of questions and answers implemented in a MySql database. The database architecture was deliberately designed to be content-free, such that it is suitable for any survey topic and is, in fact, easily adaptable to other interactive, question-answer application areas including educational testing or medical diagnostics. Second, the paper considers the /OnQ/ presentation manager which delivers questions and collects answers exploiting the power and flexibility of client browsers as well as that of the server's database. Third this paper discusses the /OnQ/ authoring tool, which allows a survey author to quickly design and deploy new survey content. Here, we pay particular attention to the authoring tool's ability to picture survey logic as a directed graph as an effective means to represent survey initiation, filtering and completion as a deterministic finite state automaton.
| Back to: Top | Programme | Page last updated on 31 August, 2003 |