Abstract :
GUIs are good for any application that requires human control interaction and that can usefully be modelled as interactive graphics (mimicked). This includes, and arises from, applications which would be considered as C3I in the military, scientific or engineering world: command, control and communication information. The goodness or otherwise of such interfaces depends on the control interaction, and on the mimic. Some control interaction design issues are common to all styles of interface. Specifically, all must include a spanning set of commands, which addresses all required functionality. In absence of other criteria, we would initially desire that this spanning set be arranged to be orthogonal. Having resolved such an orthogonal command set in the design, we might then attempt to structure it for useability. Predominantly within GUI design, established theory tells us that we should then subject the command set to task analysis. This involves a third-rework of the command set, for example, grouping commands into setup type tasks and imperative type tasks. Task analysis is well suited to structuring textual command verbs into tree graphs, and the `logical´ next step is to implement this tree structure as a menu and dialog box interface. It is exactly this which causes empty domain space syndrome. The main work area of the interface is empty, because at no point did we try and work out what to do with it. The solution lies in the use of mimics, which are what distinguish real GUIs from text based interfaces. This is what we should put in the domain space: interactive graphical models of the problem space. The decision as to what these models should look like and how they should behave, needs to be taken after completion of task analysis, but before any further design effort. In fact, the design of mimic(s) is best treated as a ongoing task, from the start of the project. Mimic design is an iterative process, with exploratory steps which may need to be abandoned, or modified, as the design evolves
Keywords :
graphical user interfaces; human factors; interactive systems; systems analysis; user modelling; control interaction design issues; empty domain space syndrome; exploratory steps; goodness; graphical user interface design; human control interaction; imperative type tasks; interactive graphical models; interactive graphics; iterative process; ongoing task; orthogonal; orthogonal command set; problem space; setup type tasks; spanning set; task analysis; textual command verbs; tree graphs; useability;