Declarative Dragging: Drag-and-Drop Without the Fiddly Code
Original: Declarative Dragging
Why This Matters
Lowering the barrier to drag-and-drop could meaningfully expand interactive education and creative tools.
Researchers Horowitz, Blinn, and Heer won a Best Paper Award at UIST 2026 for 'declarative dragging,' a model-driven approach letting developers specify target states instead of coding drag-and-drop geometry manually.
Building drag-and-drop interactions typically means writing a lot of low-level code—geometry calculations, visual feedback, animations—work that often discourages educators, game designers, and tool builders from attempting it at all. The 'declarative dragging' system, developed by Joshua Horowitz, Andrew Blinn, and Jeffrey Heer, flips the approach: developers specify which future application states should be reachable by dragging an element, and the engine handles the rest.
The system speculatively renders those possible states, extracts element positions from each rendering, then synthesizes a smooth drag interaction automatically. The team implemented this as a library called Dragology, built for SVG web interfaces. It supports composable behaviors—simple drag interactions can be combined, with modifiers adding snapping, floating, and transitions.
The project grew from a 2022 prototype Andrew Blinn built for a math game he wanted to feel more tactile, and a 2023 Seattle walk where he shared the ideas with Horowitz. Two years later, Horowitz built the library after wanting to interactively manipulate order-preserving tree maps rather than just draw them. The paper won a Best Paper Award at UIST 2026. Demo applications include algebra tools, Bézier curve editors, braid visualizations, and mathematical explorables.