Proceedings: GI 1996

An Adaptable Software Architecture for Rapidly Creating Information Visualizations

Rick Kazman, Jeromy Carrière

Proceedings of Graphics Interface '96: Toronto, Ontario, Canada, 22 - 24 May 1996, 17-27

DOI 10.20380/GI1996.03

  • BibTeX

     title = {An Adaptable Software Architecture for Rapidly Creating Information Visualizations},
     author = {Rick Kazman and Jeromy Carri`{e}re},
     booktitle = {Proceedings of the Graphics Interface 1996 Conference, May 22-24, 1996, Toronto, Ontario, Canada},
     year = {1996},
     month = {May},
     isbn = {0-9695338-5-3},
     publisher = {Canadian Human-Computer Communications Society},
     url = {},
     pages = {17--27}


While data visualization is an increasingly important analysis tool, both in research and commercial communities, the process of creating these visualizations is still quite complex. Visualizations tend to be hand-crafted, each one different from the previous. This paper presents VANISH, a system created to ease the rapid creation of arbitrary data visualizations. VANISH simplifies the creation of visualizations in two ways: by providing a special-purpose visual language, called VaPL, which maps semantic domains to visual domains; and by easing the integration of new semantic and visual domains. The software structure of VANISH emphasizes separation of concerns: we follow the Arch/Slinky metamodel of user interface software by not only separating the underlying semantic domain to be visualized from the dialog and presentation components, but by providing virtual semantic domain and presentation layers. In this way, it is simple to port from one visualization domain to another, and from one presentation component to another. We demonstrate that it is simple to create arbitrary visualizations by implementing several well-known visualization styles such as cone-trees, tree-maps, fisheye views.