Specifying Single-user and Collaborative Profiles for Alerting Systems
Jung, D. (2009). Specifying Single-user and Collaborative Profiles for Alerting Systems (Thesis, Doctor of Philosophy (PhD)). The University of Waikato, Hamilton, New Zealand. Retrieved from http://hdl.handle.net/10289/3578
Permanent Research Commons link: http://hdl.handle.net/10289/3578
The 21st century is the age of information overload. Often, humans are incapable of processing all of the information that surrounds them and determining its relevance. The impact of overlooking crucial information ranges from annoying to fatal. Alerting systems help users deal with this vast amount of information by employing a push-based rather than a pull-based approach to information delivery. In this way, users receive the information they require at the appropriate moment. Users specify their alerting needs in a profile that is subscribed to the alerting system. The alerting system is continuously fed with data, and filters this data against all subscribed profiles. Whenever incoming data matches a profile, the subscriber is alerted. Although alerting systems solve the problem of information overload, the potential of these systems has not been fully put into practice. Alerting systems are either realised as dedicated systems that, at best, offer a set of possible profiles to choose from or, at worst, offer a preset profile for one purpose only. Alternatively, they are application frameworks that offer no support for the average user; that is, the specification of profiles is realised using a programming interface. Collaboration between users when specifying profiles is not supported. This thesis verifies the described situation by considering the example application domain of health care. Within this context, a requirements analysis was undertaken involving a patient-based online survey and interviews with health care providers. This analysis revealed the utility of alerting systems but a need for support for profile specification by end-users. It also identified the need for such a system to support the collaborative nature of health care. The shortcomings of alerting systems identified for the health-care area also exist in other domains. Hence, a variety of application areas will benefit from providing universal solutions to eliminate these shortcomings. Based on these findings, this thesis proposes the graphical profile specification language GPDL and an interactive single-user software tool that supports its use (GPDL-UI). The thesis introduces a novel collaborative alerting model for Information Systems. A collaborative extension of GPDL is implemented in the software tool CoastEd, an editor for the graphical specification of collaborative profiles. The developed languages and software tools target average users who have no expertise in specifying profiles involving logics and temporal constraints. The efficacy of the proposed languages and software were evaluated through three user studies. The first study examined interpretation and specification with GPDL. Based on the results of this first study, the single-user system GPDL-UI was designed and implemented and then evaluated in a second study. In turn, the lessons learned from the implementation and user studies for the single-user system influenced the development of the collaborative approach CoastEd; this editor was evaluated in the third study. The studies have shown that GPDL and GPDL-UI are suitable means for average users to effectively specify profiles in single-user alerting systems. High levels of accuracy were reached for specification and interpretation in both studies. GPDL-UI turned out to be a usable and effective software tool. The collaborative approach and CoastEd succeed in conveying the idea of collaborative profile specification to average users. Most types of collaborative profiles were successfully specified by users. For the initiator of the collaborative profile specification process, two types of profiles call for further research. Overall, the approach, languages and software tools developed are shown to be effective and merit future research in that area.
The University of Waikato
All items in Research Commons are provided for private study and research purposes and are protected by copyright with all rights reserved unless otherwise indicated.
- Higher Degree Theses