User interface support for the integration of software tools: an iconic model of interaction

Reviewer: Hausi Albert Muller

The authors present a model of interaction for user interfaces based on an iconic representation of objects. The paper begins with a description of user interface objects such as icons, windows, interaction boxes, and scroll bars. The interaction model is a description of the semantics of an icon. The semantics define a mapping between user actions on icons and windows, and operations defined on the objects represented by the icons and windows. Both the objects and the interactions described in the paper are now part of the user interface of most modern workstations. One new feature, at least for me, is the notion of an electric icon. A set of selected icons can be dragged onto an electric icon (that is, the dragged icons are moved to the window that belongs to the electric icon). An electric icon provides a visual feedback (for example, turns gray) when the mouse passes over if the drag operation is applicable to that icon; such feedback can be interpreted as dynamic type checking for drag operations. The paper validates and illustrates the interaction model using WISh (Window-Icon Shell), an iconic command interface to the UNIX operating system. This iconic shell is very similar in flavor to the directory browser featured on the NeXT computer. Although applications can be launched from within WISh, they cannot themselves take advantage of the iconic interface provided by WISh due to implementation restrictions. The second part of the paper describes a solution to this problem, a server-based architecture for iconic shells. Server-based window systems like X Windows and NeWS have demonstrated the advantages of such an architecture over traditional centralized architectures. The paper concludes with a discussion of some applications of the server-based architecture to software development environments. Considering the audience (the paper was presented at a conference on practical software development environments), this rather short section should have been expanded. The first part of the paper is terse but well written and can serve as a good introduction to iconic interaction models. Server-based architectures as described in the second part of the paper will be commonplace in future hardware and software systems. Thus, the research ideas and contributions expressed in this paper are very timely and have already been realized in some interface architectures, such as the one featured on the NeXT computer.