Search

Understanding the semantics of graphical representations of textual programs can be difficult because the reader must understand both the relationship between the graphic and textual representations as well as the semantics of the underlying textual program. We present a framework for describing the semantics of graphical program representations in graphical terms through transformations, thus reducing the need to understand the textual form and its relation to the graphical form. In addition, our graphical semantics framework potentially offers the same advantages for understanding semantics that visual program representations offer for understanding programs. In order to illustrate our framework, we describe the semantics of control structures in VIPR, a graphical program representation for imperative languages such as C and Tcl.

The Visual Programming Challenge (VPC) provides a framework for exploring visual programming language issues in a quasi-real-time environment and for comparing competing languages. VIPR is an attempt to bring the traditional strengths of textual imperative languages to a visual programming language. It leverages these strengths to make use of well-understood design and abstraction methodologies while providing additional visual features, such as explicit representations, contextual information, and execution animation. VIPR uses these strengths and features to address the quasi-real-time problem posed by the VPC, whose requirements include interaction with the low-level vehicle manipulation code, map exploration and display, as well as additional requirements such as flexibility of solution and performance demands. Our solution does not directly address pedagogical issues or the value of VIPR's simple graphical semantics. It did point out weaknesses with VIPR's approach to solving the scalability problem in visual programming languages.