Precision in vector documents: a spatial approach Despite being scalable, vector images are typically specified and manipulated with fixed-precision numbers. Existing formats therefore are unable to store documents which need greater precision, and existing viewers perform coordinate system transforms at a fixed precision, limiting the amount any document may be "zoomed". Using existing arbitrary-precision numeric types can resolve these issues, but often at a significant or infeasible performance penalty. By taking into account the spatial nature of the data, we can store and view documents to an arbitrary precision with little-to-no performance impact. This is achieved by clipping the document's constituent Bézier paths to nodes in a quadtree. We'll discuss the artefacts caused by low-precision, the theory and implementation of our quadtree-based format and some remaining issues and future possibilities with the technique.