A couple of question regarding the methods and tools used in creating KanjiVG

[Saar Korren]

I have a couple of questions regarding the methods and tools used in creating KanjiVG. These questions are relevant for making future contributions, so as to retain consistency.

The first question is regarding stroke order. Namely, how it was decided which stroke orders are correct, which are default, and which are variants. I'll give a more specific example: CHISE IDS describes 攀 as ⿱樊手. However, the stroke order for 樊 given in kanjivg/06a0a.svg is different from the order for the first 15 strokes of kanjivg/06500.svg. Why was one order chosen for one, and a different order chosen for the other? Were different sources used for each? And if so, which? What was the deciding criteria?

The second question is regarding the paths themselves. This is important for the possibility of expanding KanjiVG, and contributing new kanji, while retaining visual consistency. Which tools were used to draw the paths? Are the same paths copied across different kanji? Under what criteria are they scaled or aligned? Are the specific spacing requirements? Guiding grid? Angles used in certain curve types? Number and types of segments?

Having some "style guide" that would assist in drawing kanji that are visually consistent with the existing ones would be very helpful.

I would highly appreciate a response from Ulrich Apel on this, since it relates to the original creation process.

[Saar Korren]

Is there a way to ping someone here? It's kind of a blocker for my project. I need some sort of response.

[Ulrich Apel]

Hi, Saar,

an introduction to the KanjiVG project is:

Apel, Ulrich, and Quint, Julien (2004): „Building a Graphetic Dictionary for Japanese Kanji – Character Look Up Based on Brush Strokes or Stroke Groups, and the Display of Kanji as Path Data“. In: Zock, Michael, and Saint Dizier, Patrick: Proceeding of the Workshop on Enhancing and Using Electronic Dictionaries, Coling 2004, Geneva; pp. 36–39.

https://aclanthology.org/W04-2106.pdf

Stroke order is taught in Japanese schools following: 

Monbushô (文部省). 1958. Hitsujun shidō no tebiki (筆順指導の手引き). Hakubundō (博文堂), Tokyo.

It  seems to be online now, for example:

https://a4lg.com/downloads/library/a4lg/pdf/筆順指導の手びき.pdf

https://a4lg.com/downloads/library/a4lg/pdf/%E7%AD%86%E9%A0%86%E6%8C%87%E5%B0%8E%E3%81%AE%E6%89%8B%E3%81%B3%E3%81%8D.pdf

This is a variant of kaisho type characters with influences from print characters like Mincho.

A more comprehensive book on stroke order is:

Emori Kenji (江守 賢治) (ed.). 2003. Kai gyō sō—hitsujun jitai jiten (楷行草 筆順・字体字典). Sanseidō (三省堂), Tokyo.

It gives traditional kaisho stroke order and its variants.  For us Monbushô stroke order and derivations were most important. In many cases, however, we had to make our own educated guesses.  攀 is not a 常用漢字 and it is probably not in Emojis book either.  I am open for more consistency.  Probably, one should start in the middle, then left, then right.

I used Adobe Illustrator for drawing the characters.  Probably, I would use something else nowadays because of their current licensing politics.  Inkscape shifted layers sometimes.

Paths, elements etc. were copied across different characters according to position and shape.  I started with a database of elements and their position.  I could open stroke order graphics with characters containing the same elements via script from the database and copy-paste corresponding elements.  By doing so, one gets correct stroke order and the basic positions of the elements.  But the characters need corrections by hand.  There are certain esthetic rules how to position the strokes, so that they are filling up space or keep space open.  We also  followed different kyōkasho 教科書 fonts.  There are Japanese books on handwriting that explain some esthetic rules, but one can learn a lot also from looking at fonts.  

Actually, following kyōkasho too strictly doesn't result in the best esthetic results for handwriting.  One could think about adapting our current character forms to handwriting with a pen.

Best wishes

Ulrich 

-- 
-- 
You received this message because you are subscribed to the "KanjiVG" group.
For options and unsubscribing, visit this group at
http://groups.google.com/group/kanjivg
--- 
You received this message because you are subscribed to the Google Groups "KanjiVG" group.
To unsubscribe from this group and stop receiving emails from it, send an email to kanjivg+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kanjivg/035ee3ec-464b-430e-8c45-12b3373db5f6n%40googlegroups.com.

[Saar Korren]

" Paths, elements etc. were copied across different characters according to position and shape. "

" There are certain esthetic rules how to position the strokes, so that they are filling up space or keep space open. "

To try to understand this part more, I tried overlaying several characters with a common radical:
投排挑披

While the radical has the same number of points in each stroke in all of them, the position of the points seem completely unrelated. I tried scaling and moving them, but couldn't make the radicals "overlap". Were the points adjusted manually for each?

[Ulrich Apel]

Yes. the points and the strokes were adjusted by hand.  The 手偏 (⺘) needs less space generally than the right side.  But the more strokes are on the right side the slimmer ⺘becomes.  The first and third stroke may be at positions where they can fill out space, that the right side leaves free.  

Strokes of the same length should have the points around similar positions.  The stroke points can also be used for adjusting writing speed.  One should get natural writing speed within one stroke, if the time between two points is always the same.  In many cases, the writing becomes slower toward the end of a stroke; sometimes it doesn't. 

Best wishes 

Ulrich

To view this discussion on the web visit https://groups.google.com/d/msgid/kanjivg/27466561-270a-4bb9-ab82-c9b2fd98c011n%40googlegroups.com.

[Ben Bullock]

Somehow or another I don't seem to have received the message below, although I can't find any filters or spam settings which might have deleted the email accidentally so I didn't see it until now. I've added the important reference and also the PDF download links in the email to the kanjivg reference list at

https://kanjivg.tagaini.net/ref.html