Kanji with 臼; *HzFst.svg and *.svg should be swapped

[mifunetoshiro]

Basically, third and fourth strokes should be swapped, so that you draw the left three strokes first, then the two on the right, and the bottom one last.

臼 U+81FC
潟 U+6F5F
鼠 U+9F20
倪 U+502A
兒 U+5152
冩 U+51A9
啗 U+5557
毀 U+6BC0
寫 U+5BEB
滔 U+6ED4
瀉 U+7009
燬 U+71EC
猊 U+730A
睨 U+7768
稻 U+7A3B
竄 U+7AC4
舁 U+8201
舂 U+8202
舅 U+8205
舊 U+820A
諂 U+8AC2
貎 U+8C8E
蹈 U+8E48
閻 U+95BB
陷 U+9677
霓 U+9713
韜 U+97DC
餡 U+9921
鬩 U+9B29
鯢 U+9BE2
麑 U+9E91
鼬 U+9F2C
焰 U+7130

[ospalh]

The Japanese Wiktionary seems to list the current KanjiVG standard first:
https://ja.wiktionary.org/wiki/%E8%87%BC

[mifunetoshiro]

I see. Kakijun.jp uses the other variant, though.

Also, 097dc-HzFst.svg is missing the last stroke number (20).

[ospalh]

(Forget what i wrote a minute ago)

OK, found the file. There is a reason that that is not in the main kanji folder but in kanji-mismatch.

[ospalh]

… i mean, it would be nice if someone would fix those files in kanji-mismatch. (It's just that i don't feel motivated at the moment.)

I am also not saying that this change should not be done. Wiktionary isn’t really an authority on anything. It is just that there are different ways to draw this, and sometimes it can be hard to determine which variant should be the primary one. When i understand the comment under sources on the KanjiVG web site correctly, there isn’t any official stroke order for Japan.

[mifunetoshiro]

It says on kakijun.jp that there is no problem if the 3rd and 4th strokes are reversed (the current KanjiVG version), but their main one is from the kanji kentei manual book. The one preferred by Wikipedia is from a 1983 stroke order dictionary. Indeed both are correct, as there is no "official" order, but I'd personally prefer the order recommended by the kanji kentei people (and kakijun.jp). Not my call, though...

[ospalh]

My guess is that unless you do this yourself, it will stay the current way, even if only out of inertia.

The swap strokes script could be useful for this. (You would have to swap the strokes and the numbers with one call of the script each.)

eta
You need that script only in the cases where there is no -HzFst variant around. Anyway, you probably should fix the id attributes when you move the files around.

[mifunetoshiro]

Since both ways are acceptable and correct, I'm closing this. I think the only higher priority issue remaining is #49.