Independent and Collaborative Visualization Tool Development

Visualization is thriving as an academic discipline. However, the development of visualization heavily relies on applications in other base sciences. We examine the visualization development process, which includes both collaborative development with domain scientists and independent development by visualization tool developers, and tell the behind-the-scene stories of FluoRender.

Visualization is a burgeoning branch of scientific studies, nurtured by experts from originally diverse backgrounds and disciplines of computer graphics, software engineering, natural sciences, social sciences, and arts. Sitting at the juncture of multidisciplinary interactions, visualization has demonstrated its independence as a self-sustained academic discipline. The excitement, as well as growing pains, has never been felt more profoundly by researchers in recent years. The dependence of visualization research on other base sciences has been even more accentuated, as evidenced in the growth of application-concentrated conferences including IEEE VAST and BioVis, the increasing attention to application papers, and the emphasis on user studies. The argument between the dependence and independence of visualization as a science can be an overly simplified discussion considering the trend in scientific research, while the same question may be asked in many interdisciplinary fields. After all, the advance of human knowledge as a whole is the common goal of all scientific disciplines, the merge, division, and collaboration among which are merely an instrument for achieving the goal. Using the words of visualization researchers, the development of visualization can be described as a function of high-dimensional space; a decomposition depending on discrete multidisciplinary bases already distorts it; a simple projection to one-dimensional (1-D) is undoubtedly misleading. Although a comprehensive assessment of all visualization research is understandably impossible, here we attempt to slice a profile of the high-dimensional space by examining the developmental process of a successful visualization tool for biomedical research, FluoRender. Especially, we pay attention to the relationship between domain experts and visualization developers in the collaborative development of FluoRender, as well as in the mutual influence between FluoRender's value in data visualization as an independent tool and its service to advance biomedical research. Some of the viewpoints, insights, and behind-the-scene stories are less likely to be discussed in either technical and application publications by visualization researchers, as well as publications by researchers in biomedical sciences. Although our sample space is far from sufficient, we hope this article serves as an inspiration to researchers by providing fresh examples from our viewpoint.