筆順 · the component’s own sequence
Component consistency
When the same component is written inside different characters, how often its stroke-type sequence stays the sequence it has on its own.
These are frequencies in one KanjiVG snapshot. They are not a rule that a component must keep its sequence, or that a given position must change a stroke.
Each pair sets the component on its own beside one character that contains it. Cinnabar marks the strokes named in that pair’s caption. Click a plate to replay. With reduced motion the finished character stays still, and the marks stay.
- Cinnabar
On this page
What is counted
- Host
- A jōyō character in whose file the element occurs under the character group. The character’s own outer group is not a host of itself. 林 is one host of 木.
- Instance
- One component. A group with no
kvg:partis one instance. Groups that carrykvg:partand sharekvg:elementandkvg:numberare joined: nested copies of one part number are one part, and a new instance starts when the part number stops increasing. 林 has two instances of 木. 歌 has two instances of 丁. - Reference
- The
kvg:typesequence of the element’s own main file, when that file exists. The element need not be in the jōyō list. When the file does not exist, the reference is the most common sequence among the instances. On this page that second case is 𠂉, 䒑, and 龶. The table marks those three. - Match
- The instance’s strokes, in stroke-number order, carry the same kvg:type strings as the reference, letters included. A slash is part of the string.
- Split
- A stroke number falls strictly between the instance’s first stroke and its last, and that stroke is not part of the instance. The component is written on both sides of something else. Part numbers that follow one another, with no stroke between them, are not a split. A split can still match the reference sequence. The leave-and-return atlas counts the same part marks and asks what is written in the gap.
- What differs
- An instance that does not match is one of five. Shape: the same number of strokes, and the stroke characters are not a reordering of the reference. Order: the same stroke characters, letters ignored, in a different order. Fewer and more: the stroke count is not the reference count. Letter: the stroke characters and their order stay, and a trailing letter differs. These five add up to the instances that differ. Split is counted beside them, not instead of them.
The inventory
Every element in at least twenty characters
One row is one element. Hosts is characters. Instances is components. Match and the five difference columns are instances. Match and Differ add up to Instances. Shape, Order, Fewer, More, and Letter add up to Differ. Split is separate: an instance in Match or in Differ may also be split. The last row sums instances. It does not sum hosts, because one character hosts many elements. A mark under the element means the reference is the most common sequence, not a file of that element alone.
| Element | Hosts | Instances | Match | Differ | Split | Shape | Order | Fewer | More | Letter |
|---|
The same sequence
On its own
In another character
On its own
In another character
On its own
In another character
On its own
In another character
On its own
In another character
On its own
In another character
Where the shape changes
Every shape change
One row is one sequence written for one element. Instances counts each component. Characters lists each jōyō character once. Gap is whether a stroke of something else falls inside the component. Position is the kvg:position on the group, and how many instances carry it. No position means the file does not mark one.
| Element | Class | Sequence | Instances | Gap | Position | Characters |
|---|
On its own
In another character
On its own
In another character
On its own
In another character
On its own
In another character
Where the order changes
Every change of order
One row is one sequence written for one element. Instances counts each component. Characters lists each jōyō character once. Gap is whether a stroke of something else falls inside the component. Position is the kvg:position on the group, and how many instances carry it. No position means the file does not mark one.
| Element | Class | Sequence | Instances | Gap | Position | Characters |
|---|
On its own
In another character
On its own
In another character
Fewer strokes, and more
Every change of stroke count
One row is one sequence written for one element. Instances counts each component. Characters lists each jōyō character once. Gap is whether a stroke of something else falls inside the component. Position is the kvg:position on the group, and how many instances carry it. No position means the file does not mark one.
| Element | Class | Sequence | Instances | Gap | Position | Characters |
|---|
On its own
In another character
Split across other strokes
Every split
One row is one sequence written for one element. Instances counts each component. Characters lists each jōyō character once. Gap is whether a stroke of something else falls inside the component. Position is the kvg:position on the group, and how many instances carry it. No position means the file does not mark one.
| Element | Class | Sequence | Instances | Gap | Position | Characters |
|---|
On its own
In another character
On its own
In another character
Only the letter changes
Every letter-only change
One row is one sequence written for one element. Instances counts each component. Characters lists each jōyō character once. Gap is whether a stroke of something else falls inside the component. Position is the kvg:position on the group, and how many instances carry it. No position means the file does not mark one.
| Element | Class | Sequence | Instances | Gap | Position | Characters |
|---|
What this shows
70a0b7ae0c18ceb5cb358274b029cce0234a43bc of github.com/KanjiVG/kanjivg. The plate files are unmodified copies in site/assets/kanjivg/. The sequences are the kvg:type values in those files. Jōyō index: the list used on the Gallery.