筆順 · length, place, closure
Order as meaning
Near-identical stroke inventories that close into different characters, because a stroke is longer, sits elsewhere, closes further, or is written at another index.
KanjiVG keeps two structures in one file, as Strokes sets out. Nested groups are the hierarchical containment of components. The -sN order on each path is how those strokes are realized, and that order is part of the form. Components are the arguments of the composition. Valence is how many slots the frame opens — the arity — and not the stroke count. Below, the type sequence is held still, or moved at one index, and the character still changes. The change is length, place, how far a stroke closes, or which index a class occupies.
Each plate draws itself in writing order. Click a plate to replay. Gray digits are that file’s own stroke numbers. With reduced motion the finished character stays still.
On this page
Where the kvg:type string matches at every index, the character can still change with the length of a stroke.
What this difference does
土, 士, and 工 carry one sequence: ㇐, ㇑a, ㇐. In 土 the upper horizontal is the shorter bar and the lower one runs wider. In 士 those lengths are reversed. In both, the vertical begins above the upper bar. In 工 the vertical begins on that bar and does not continue above it, and the lower bar is the long one. Same types, same order. The length of the two bars, and how far the vertical reaches, is what separates the characters.
What this difference does
Both files are ㇐, ㇐, ㇑, ㇒, ㇏. In 未 the first horizontal is the short one and the second is longer. In 末 the first is the long one and the second is shorter. The last three types do not change. The character is the relative length of those two bars.
What this difference does
Both files are ㇐, ㇐, ㇒, ㇏. In 天 the top bar is the longer of the two horizontals, and the left-falling stroke begins under that bar. In 夫 the second horizontal is the longer one, and the left-falling stroke begins above the top bar. The sequence is the same. The lengths, and where the third stroke starts, are not.
What this difference does
Four strokes, the same classes: a vertical, a bend, and two horizontals. KanjiVG types 日 as ㇑, ㇕a, ㇐a, ㇐a and 曰 as ㇑, ㇕, ㇐, ㇐, so the full strings are not the same. The frame of 日 is taller than it is wide. The frame of 曰 is wider than it is tall. 曰 is not in the jōyō list, and it is not in the count below. The proportion of that frame is the difference.
Where the type sequence matches, the character can still change with where a stroke begins.
What this difference does
The sequence is ㇒ then ㇏ in all three. In 人 the left-falling stroke is the long one and starts at the top; the right-falling stroke is shorter and starts lower. In 入 those roles are reversed. In 八 the two strokes start apart: the left one lower and further left, the right one higher, and they do not meet in a single peak. One sequence, three characters, distinguished by length and by where each stroke begins.
What this difference does
Both are ㇒, ㇐, ㇐, ㇑. In 牛 the vertical, s4, begins above the short diagonal and above both bars. In 午 that vertical begins under the upper bar, and the short diagonal is what reaches above it. The types match. The reach of the vertical does not.
What this difference does
Both are ㇆ then ㇒. In 刀 the left-falling stroke starts about level with the hook and is the shorter of the two. In 力 it starts above the hook and is the long stroke. Same types. The length and the starting height of s2 separate them.
What this difference does
犬 and 太 share ㇐, ㇒, ㇏, ㇔. The dot is s4 in both. In 犬 it sits at the upper right, above the horizontal. In 太 it sits under the crossing. 大 is the first three of those types and has no dot. It does not share the four-type sequence. Among jōyō characters it shares ㇐, ㇒, ㇏ with 丈, and that pair is in the count below. The dot’s place, or its absence, is the difference on the plates.
The last stroke is the same class in all three. What changes is how far it rises. The full type string is not identical in every file.
What this difference does
The third stroke is ㇟ in all three. In 己 it starts on the middle bar and does not rise, so the top stays open. In 已 it rises past that bar and stops short of the top. In 巳 it rises to the top. 已 and 巳 use s1 ㇕; 己 uses s1 ㇕c. 已 and 巳 are not in the jōyō list, so they are not in the count below. The extent of s3 is what separates the three graphs.
The stroke classes are the same set. The index of one of them is not, so the type strings are not the same sequence.
What this difference does
The sequences are not the same string. 右 opens ㇒ then ㇐. 石 opens ㇐ then ㇒. From s3 both write 口 as ㇑, ㇕b, ㇐b. The first two classes trade indexes. The same exchange, tied to which stroke is the short one, is the subject of Same parts, different order for 右 and 左. Here the two files are two characters.
What this difference does
These three do not share one type string. 王 is ㇐, ㇑a, ㇐, ㇐. 玉 writes that sequence and then a dot, s5 ㇔, on the lower right between the middle bar and the bottom bar. 主 writes the dot first, s1 ㇔, above the top bar, and then those four types. The four non-dot types match. What changes is whether there is a dot, and which index it occupies.
Counted over the jōyō list
| kvg:type sequence |
|---|
No sequence in this filter.
What this shows
70a0b7ae0c18ceb5cb358274b029cce0234a43bc of github.com/KanjiVG/kanjivg. Plate files are in site/assets/kanjivg/. Jōyō index: the list used on the Gallery.