Look inside the sequence.
Scrub through real accepted-stroke records, isolate a phase, and test where and how the drawing develops.
Sequence
Reading the instrument
Each record contains a mark ID, dark/light flag, sampled points, and pressure. Path length is the sum of straight segments between recorded points. The density view counts stroke starting positions in a 20 × 20 grid. The fifth view divides the record into five equal ranges by stroke count.
The bundled examples are generated drawing-machine records. Traces are centerline approximations: they do not reproduce brush width, pressure, erasure or media physics. The finished rendering may also include a pre-existing canvas. Order is accepted-stroke order, not clock time.
Built from generated drawing-machine records. scikit-image and OpenCV offer further skeletonization and registration methods for future image-based experiments.
Works offline. Imported JSON is processed locally. CSV exports include every stroke in the current window and its calculated geometry.