Visual programming languages (VPLs) offer the hope of higher reliability at lower labor costs, but only if they can be scaled up for realistically-sized programming projects. This research addresses a critical factor in scaling up declarative VPLs, by developing an interactive and visual yet powerful approach to data abstraction. The visual approach adds expressive power to VPLs through simplicity, rather than the traditional approach of providing them at the expense of simplicity. The concepts developed are being implemented in a research declarative VPL and made available to interested researchers. This work builds upon recent work in data abstraction, event- handling, and the scaling-up problem. Visual design issues are addressed, and the current implementation is being prepared to serve as a base research system. The research seeks the incorporation of a visual and flexible approach to code reuse. At the same time, a simple approach to exception handling based on the earlier work on events is being developed. Additional work regarding visual impact and visual support for the approach to data abstraction is expected to enhance the simplicity and effectiveness of the results.
