Visual dataflow languages (VDFLs), which include commercial and research systems, have had a substantial impact on end-user programming. Like any other programming languages, whether visual or textual, VDFLs often contain faults. A desire to provide programmers of these languages with some of the benefits of traditional testing methodologies has been the driving force behind our effort in this work. In this article we introduce, in the context of prograph, a testing methodology for VDFLs based on structural test adequacy criteria and coverage. This article also reports on the results of two empirical studies. The first study was conducted to obtain meaningful information about, in particular, the effectiveness of our all-Dus criteria in detecting a reasonable percentage of faults in VDFLs. The second study was conducted to evaluate, under the same criterion, the effectiveness of our methodology in assisting users to visually localize faults by reducing their search space. Both studies were conducted using a testing system that we have implemented in Prograph's IDE.

We introduce a visual scripting language and environment for handheld computers which we are developing, and discuss our research into adapting it for use with devices enabled with wireless communication, such as Bluetooth and IrDA. The environment is event-driven, using dataflow diagrams for programming. The language has its roots in the general purpose visual language, Prograph, but is significantly modified for end-user programming on a handheld computer, and supplemented with functionality for monitoring and controlling wireless devices.

This research studies usability problems in ReactoGraph, a visual programming language designed to simplify the development of graphical user interfaces. Its design incorporates concrete representations of interface controls and behavior but suffers from complexity in defining message passing, inter-component data flow, and event handling. We attempt to further simplify the use of ReactoGraph by considering effects of tool design on cognitive abilities and introduce three high-level semantic additions to the language. We describe these in the context of cognitive load and discuss results of a small user study that shows promise in their effectiveness.

PdaGraph is an end-user component-based programming language for the PDA. In the first stage of our research a data-flow visual programming language was designed, but this was lacking an intuitive user interface and the connections between different parts of the program were not visible. To make PdaGraph suitable for end-users, we designed a high-level component-based message-flow programming language and environment. To verify that the design is end-user friendly we conducted a usability study using a low fidelity paper prototype, and feedback was obtained by videotaping the subject-system interactions and by using a cognitive dimensions questionnaire