After 13 RageBuilder projects dominated by canvas-based visualizations and Web Audio sonification, this project completely breaks the pattern. Research uncovered academic work on optimal cooking procedure presentation (MDPI Algorithms 2021) and the Cooklang recipe markup language. The insight: recipes are fundamentally programs — they have variables (ingredients), sequential logic (steps), dependencies (you can't serve before you cook), and parallelism opportunities (boil water while chopping vegetables). Nobody had built an interactive tool that treats recipes as executable code.
Recipe Compiler lets you write cooking recipes in a simple domain-specific language (DSL), then analyzes them like a real compiler: parsing ingredients, building a dependency graph between steps, performing topological sort to determine execution order, calculating the critical path, and generating an optimized parallel schedule. The result is an interactive timeline showing exactly when to do each step, which tasks can run simultaneously, and how much time you save through parallelism.
@ are ingredients (format: @name{quantity unit}). Lines starting with # are steps (format: #description duration: Nmin temp: T°C depends: step_ids).@ followed by the ingredient name in curly braces with quantity and unit: @flour{2 cups}# followed by a description. Add optional metadata: duration: 10min, temp: 200°C, depends: 2 3