What practitioners really want: requirements for visual notations in conceptual modeling

摘要

This research was aimed at eliciting the requirements of practitioners who use conceptual modeling in their professional work for the visual notations of modeling languages. While the use of conceptual modeling in practice has been addressed, what practitioners in fact require of the visual notation of the modeling languages they use has received little attention. This work was thus motivated by the need to understand to what extent practitioners' requirements are acknowledged and accommodated by visual notation research efforts. A mixed-method study was conducted, with a survey being offered over the course of several months to LinkedIn professional groups. The requirements included in the survey were based on a leading design theory for visual notations, the Physics of Notations. After preprocessing, 104 participant responses were analyzed. Data analysis included descriptive coding and qualitative analysis of purposes for modeling and additional requirements beyond the scope of visual design. Statistical and factorial analysis was used to explore potential correlations between the importance of different requirements as perceived by practitioners and the demographic factors (e.g., domain, purpose, topics). The results indicate several correlations between demographic factors and the perceived importance of visual notation requirements, as well as differences in the perceived relative importance of different requirements for models used to communicate with modeling experts as compared to non-experts. Furthermore, the results show an evolution from trends identified in studies conducted in the previous decade. The identified correlations with practitioners' demographics reveal several research challenges that should be addressed, as well as the potential benefits of more purpose-specific tailoring of visual notation design. Furthermore, the shift in practitioner demographics as compared to those found in earlier work indicates that the research and development of conceptual modeling efforts needs to stay up-to-date with the way practitioners employ conceptual modeling.