Shiro - A language to represent alternatives

When people solve problems they explore a variety of potential solutions. Parametric systems have been added to the interaction models of design tools to make it easier to change a design. Just as a cell is changed in a spreadsheet, when a parameter is changed all parts of the design that depend on that parameter update. While tools with parametric systems are powerful, users are limited to single state solutions and are forced to use workarounds like using layers and file naming conventions. These improvisations are caused by tools whose user interface and document models only support a single state. Single-state document models are only designed to represent a single artifact. In this work, I describe Shiro, a declarative, dataflow language for expressing alternatives in parametric systems. It provides a multi-state document model for parametric systems. To make this possible, I introduce the concept of subjunctive nodes, nodes that contain options. Options allow users to vary property values and computations. I demonstrate Shiro with a variety of examples from the designs of design and data analysis. Finally, I discuss what I learned while designing and implementing the language and provide a set of recommendations for future research.