5.2. HEURISTICS AND METAHEURISTICS: GUIDANCE IN DECISION MAKING
User interface design heuristics, as guiding principles, have been accepted as one way to assess and shape systems to make them more usable (Nielsen, 1994). In this usage heuristics are prescriptive rules of thumb that correct for biases rather than descriptive of the biases that produce errors. User interface design heuristics typically concern the decisions about how the representing world should be structured to be more usable. For example, there are heuristics concerned with ensuring the user is informed of what is going on in a system. The analysis presented in this paper suggests there may be a larger context for describing heuristics in user interface design. This larger context has to do with metaheuristics that are the rules of thumb governing decisions about the transformations required of designers. Such metaheuristics pertain to two broad areas in creation of a representational system: the choice of design processes and the choice of representational models. The next three sections describe three groups of rules of thumb. These metaheuristics reflect the experience described in this paper and echo the perceptions of authors within this volume, as well as others.
5.2.1. Defining the Represented World: Start Off on the Right Foot
The thrust of this metaheuristic is that it is critical to obtain the information needed to define the represented world and that this information needs to be cast into forms amenable to the larger goal of user-centered design (a point argued as well by Monk in this volume). In particular, the work on this project suggests the following rules:
- 1. Make the represented world sharable One of the strengths of the scenario approach is that it retains the overall information and structure of end user experience. This enables multiple diverse audiences to share this representation and work with it throughout the course of the project. The scenarios serve as the context for storyboard walkthroughs, usability goal definition, and other design activities.
- 2. Preserve the end-user view in the represented world Though usually an intrinsic part of the scenario form, it is worthwhile to isolate this particular idea. Specific words and end user experiences are extremely important in selecting aspects of the represented world to be modeled and for choosing details in the representing world. Essentially, key data must be kept alive through the course of the design cycle.
- 3. Define the represented world at an appropriate granularity One size does not necessarily fit all for design. This project used both scenarios and use cases as complementary forms for the represented world. Under other circumstances, simply using scenarios may be adequate and because they are so robust they may be the preferred option if a single representation must be chosen. Alternatively other representations of end user work may work well for specific tasks. This application was complex enough to require the disciplined decomposition provided by use cases, while others may not require it. In any event, it is useful to appreciate the relative strengths and weaknesses of alternative representations as they are brought to bear during a design cycle.
5.2.2. Minimizing the Gap: Help the Magic Happen
Part of the pleasure of design is watching the magic happen. Nevertheless, part of the goal of a skilled designer or design team should be to create the conditions for the magic to happen, and not to leave it to chance (hence a notion like Systematic Creativity, Scholtz and Salvador, Chapter 9, this volume). One way to accomplish this is to build a process that supports the coding and recoding of information necessary for success.
- 1. Consciously create mediating abstractions As work progresses objects, models and metaphors should be the blueprint guiding the use of the bricks and mortar determined by the user interface syntax. This principle is based on the belief that it is difficult to go from user-centered data to an implementation without some mediating ideas; thus, explicitly defining them makes the process more efficient. In effect, part of solving the problem of design is creating different representations.
- 2. Make iteration work for you This and the next principle go together. Often iteration is looked upon as the failure to get it right the first time. Here the opposite is advocated. Do not expect to get it right the first time. Early progress can be measured by the number of questions asked, rather than the number answered. Use rapid cycles in iterative design to test alternative views and designs, knowing the gap is bridged incrementally. However, be disciplined enough to get the information needed to resolve the questions and refine the design by creating paper prototypes, quick high-fidelity prototypes and other approximations of the representing world.
- 3. Design is a bootstrap operation If the answer was clear, then the design work would be finished. Expect early iterations to be fragmented and imperfect. The goal of early iterations is create mediating representations that lead to the finished product, not create the finished product itself.
|