Dragology demos illustrate the range of visualizations possible with declarative dragging.
*A new model‑driven approach lets developers create drag‑and‑drop interactions by declaring future states. The breakthrough, presented as the paper “Declarative Dragging,” earned the UIST 2026 Best Paper award.*
A trio of researchers—Joshua Horowitz, Andrew Blinn and Jeffrey Heer—just turned a two‑year‑old prototype into a prize‑winning paper. Their system, dubbed Declarative Dragging, won the Best Paper Award at the 2026 ACM Symposium on User Interface Software and Technology (UIST). The core claim is simple: make a diagram draggable by stating what it can become, not by hand‑coding geometry, feedback and animation. The result is a library called Dragology that renders speculative future states, extracts positions, and synthesizes a smooth drag experience with virtually no low‑level code. The breakthrough promises to lower the barrier for educators, game designers, mathematicians and any tool builder who needs high‑quality drag interactions.
Declarative dragging inverts the traditional drag‑and‑drop workflow. Instead of programming every visual response, the interface author lists the future model states that a draggable element may reach. An engine then speculatively renders each state to a view tree, reads the element’s position in each rendering, and stitches those positions into a continuous, responsive interaction. The approach is model‑driven: the drag behavior emerges from the underlying data model rather than from imperative UI code. By abstracting geometry and animation, the system lets creators stay within the semantic world of their application while the engine handles the tactile translation.
Dragology is a JavaScript library built for SVG interfaces on the web. It implements the declarative dragging concept by exposing a composable API. Simple model‑driven drag behaviors can be combined, and modifiers such as floating, snapping and transitions can be layered on top. The library speculatively renders each declared future state, extracts the dragged element’s coordinates, and synthesizes a drag‑and‑drop interaction without requiring the developer to write geometry or animation code. A demo gallery showcases applications ranging from algebraic visualizations to twisted trees, Bézier curves, braid diagrams, card piles, face trominoes and sprouting trees.
Designing bespoke drag interactions traditionally demands fiddly code to manage geometry, feedback and animation. That overhead forces many educators, game designers and mathematicians to abandon drag‑based interfaces altogether. Declarative dragging eliminates that friction. By declaring permissible future states, developers can instantly generate high‑quality, tactile interactions. The approach shines in parallel‑representation tools, as demonstrated by the permutation lab where dragging in one view updates all others in real time. In the Hazel programming environment, refactorings become tangible and discoverable, turning abstract code edits into concrete, manipulable objects.
Since the paper’s release, the authors have expanded the technique into more intricate mathematical explorables. Andrew Blinn is integrating declarative dragging into existing applications, including the Hazel academic programming environment where every edit preserves code meaning. He is also reviving the original math‑game concept, adapting the interaction to gamepads and moving from a 2D canvas into a 3D world. The composable nature of Dragology suggests it could become a standard building block for any web‑based UI that needs drag‑and‑drop, from educational tools to complex data visualizations.
Declarative Dragging earned the Best Paper Award at the 2026 ACM Symposium on User Interface Software and Technology (UIST).
The initial prototype of the math‑game concept that led to model‑driven dragging was built in 2022.
The library is called Dragology and is built for SVG interfaces on the web.
Declarative Dragging flips the script on UI development. By letting the data model dictate motion, it strips away the tedious scaffolding that has kept drag‑and‑drop out of many applications. The Best Paper win at UIST 2026 validates the approach, and the open‑source Dragology library offers a ready‑to‑use path forward. As more tools adopt the model‑driven paradigm, the next generation of interfaces will feel more natural, more responsive, and far easier to build.
Sources: Hacker News – Declarative Dragging (https://declarative-dragging.github.io/)