Scaling up visual programming languages

TL;DR: The scaling up problem is how to expand applicability without sacrificing the goals of better logic expression and understanding, and nine major subproblems are discussed and emerging solutions from existing VPL systems are described.

Abstract: The directness, immediacy, and simplicity of visual programming languages are appealing. The question is, can VPLs be effectively applied to large scale programming problems while retaining these characteristics. In scaling up, the problem is how to expand applicability without sacrificing the goals of better logic expression and understanding. From a size standpoint, scaling up refers to the programmer's ability to apply VPLs in larger programs. Such programs range from those requiring several days' work by a single programmer to programs requiring months of work, large programming teams, and large data structures. From a problem domain standpoint, scaling up refers to suitability for many kinds of problems. These range from visual application domains-such as user interface design or scientific visualization-to general purpose programming in such diverse areas as financial planning, simulations, and real time applications with explicit timing requirements. To illustrate the scaling up problem, we discuss nine major subproblems and describe emerging solutions from existing VPL systems. First, we examine representation issues, including static representation, screen real estate, and documentation. Next, we examine programming language issues-procedural abstraction, interactive visual data abstraction, type checking, persistence, and efficiency. Finally, we look at issues beyond the coding process. >