Article Preview
TopEnd-User Development (Eud)
Familiarity with software applications has become an essential requirement for professionals in a variety of complex domains: architects, doctors, engineers, biochemists, statisticians, and film directors (among many others) all depend on the mastery of various collections of applications (Eisenberg & Fischer, 1994) in their areas of expertise. These applications, to be at all useful, must provide domain professionals with complex and powerful functionality. However, in doing so, these systems likewise increase the cognitive cost of mastering the new capabilities and resources that they offer. Moreover, the users of these applications will notice that “software is not soft”—that is, that the behavior of a given application cannot be changed or meaningfully extended without substantial reprogramming effort.
The need for end-user development is not a luxury but a necessity: computational systems modeling some particular “world” are never complete; they must evolve over time because (1) the world changes and new requirements emerge; and (2) skilled domain professionals change their work practices over time—their understanding and use of a system will be very different after a month and certainly after several years. If systems cannot be modified to support new practices, users will be locked into existing patterns of use.
These problems were recognized early in the context of expert systems and domain-oriented environments as illustrated by the following two examples:
- ⁃
Expert systems: The Teiresias system (Davis, 1984) was a module to support domain professionals to augment the existing knowledge base of a medical expert system; the objective of this component was to establish and support interaction at a discourse level that would allow domain professionals to articulate their knowledge without having to program in Lisp.
- ⁃
Domain-oriented environments: The Janus-Modifier system (Fischer & Girgensohn, 1990; Girgensohn, 1992) supported not just human-computer interaction but human problem-domain interaction to allow kitchen designers to introduce new components and new critiquing rules into design environments in support of kitchen design.