Ink matched is the share of pixels carrying ink - in this renderer's output, the
reference, or both - that agree, averaged over the chapter. It counts only where something was drawn,
deliberately: most of a test canvas is blank, so measuring over the whole image would score a document
that rendered nothing at all at around 93% purely for agreeing about the background.
It moves whenever rendering improves, which a pass count only does once a test crosses its threshold,
so it is the better progress signal of the two - and the worse verdict. Passing still means matching the
reference across the whole image, which is a stricter thing to ask.
Reference images. The pass/fail verdict and ink-matched figure above are computed against a fresh, independently-rendered reference set, not the W3C image shown on each test's own page (that one is kept only because it is safely W3C-hosted, and may no longer visually agree pixel-for-pixel with what was actually compared). See ctgnz/foxglove#200. Generated 2026-09-15T21:09:49.425210596Z with Playwright/Chromium 153.0.8010.12, W3C SVG 1.1 suite sha256 b5f46cca1ad79b670f9179770b2366c57efd5c671d084144090feab4b7ff1030.
Unlike every other category on this page, this is a real, measured result - not a static single frame,
but a live comparison against a headless Chromium reference, seeked to several moments derived from each
test's own real SMIL timing (#208). See
ctgnz/foxglove#208 for methodology.
These results are not meaningful, by design. This test's pass criteria require scripted or user-driven interaction (a click, a DOM call) that foxglove itself never supplies - foxglove renders SVG into JavaFX Nodes, and interaction is the consuming application's responsibility, not this library's. This category is expected to remain near 0% permanently; see ctgnz/foxglove#201.