Sequence
The kvg:type values in -sN order. A suffix stays in the string. ㇐a is not ㇐.
Edit distance
On the list of types, as tokens. Distance 0 is the same list. Distance 1 is the same length with one token different, or lengths that differ by one where deleting one token from the longer list leaves the shorter list. A swap of two tokens is distance 2, and it is not a candidate.
The 109-unit square
Same stroke, clear stroke
Kept pair
Early, middle, late
For stroke i in a file of N strokes, N at least 2, the place is early when (i−1)×3 is less than (N−1), middle when (i−1)×3 is less than 2(N−1), and late otherwise. i starts at 1. N is the stroke count of the longer file. A pair with more than one differing stroke is counted, in the summary, under the earliest of them. The table lists every differing index.

The kept pairs are the ones these tests leave. The count is a count in one KanjiVG snapshot. It is not a rule that a character must differ from another by one stroke.

Each plate draws itself in writing order. Cinnabar marks the stroke or strokes that differ. On an added stroke, only the longer character has that stroke to mark. Click a plate to replay. With reduced motion the finished character stays still, and the highlight stays.

On this page

Counted over the jōyō list

Step Pairs

Plates

Every kept pair

One row is one kept pair. The stroke index is on the longer file. Distance 0 is the same type list. Distance 1 is one substitution or one added or missing stroke.

Pair Strokes Distance Test Differing stroke

What this shows

Stroke diagrams: KanjiVG, Ulrich Apel et al., CC BY-SA 3.0, from commit 70a0b7ae0c18ceb5cb358274b029cce0234a43bc of github.com/KanjiVG/kanjivg. Plate files are unmodified copies in site/assets/kanjivg/. Jōyō index: the list used on the Gallery.