Re: 仮想ボディと外枠の関係性

To place text you have to use the font origin so the body discussion cannot omit it. The fact it usually is the Latin baseline is coincidence.

—Nat
________________________________
From: 木田泰夫 <kida@mac.com>
Sent: Saturday, July 27, 2024 11:16:54 PM
To: Nat McCully <nmccully@adobe.com>
Cc: Toshi Kobayashi <binn@k.email.ne.jp>; Taro Yamamoto <tyamamot@adobe.com>; public-i18n-japanese@w3.org <public-i18n-japanese@w3.org>
Subject: Re: 仮想ボディと外枠の関係性


EXTERNAL: Use caution when clicking on links or opening attachments.


Thank you Nat.

For CJK typography there is the opposite problem: the body as you explain is the basis for all layout and text positioning and there is no need to understand where the Latin baseline is.

That is exactly what I want to express in JLReq-d, and in that, I believe we are in complete agreement.

So far, the approach has been to explain Japanese layout based on the body model.

For mixed layouts, we need to explain how characters with the body model and those with the Latin baseline model should be aligned, and how text systems should handle the translation between them.

Of course, we need to be realistic. If fully supporting the body model requires text systems to have two different modes, the idea would not fly.

By the way, I think it might be a good idea to always use the term "Latin baseline" to emphasize that the model is not universal.

- kida

2024/07/28 14:37、Nat McCully <nmccully@adobe.com>のメール:


Yes. However for Latin typography (and by extension digital typography and fonts for a screen) there is no need to worry about the body of the font or the top or center or bottom of that body. They only need to place the Latin baseline in the canvas and draw there.

For CJK typography there is the opposite problem: the body as you explain is the basis for all layout and text positioning and there is no need to understand where the Latin baseline is. However in a digital context the only locator for text drawing is that very baseline which is not thought about or understood. This is why digital type and fonts are incompatible with CJK typography unless you know about both. So far only Adobe has tried to support both so the user can either think in terms of baseline or in terms of the body, and the engine does the translation. I want all engines to know it is their job to do this translation.

—Nat
________________________________
From: Kobayashi Toshi <binn@k.email.ne.jp>
Sent: Saturday, July 27, 2024 7:51:07 PM
To: Taro Yamamoto <tyamamot@adobe.com>; 木田泰夫 <kida@mac.com>; Nat McCully <nmccully@adobe.com>
Cc: JLReq TF 日本語 <public-i18n-japanese@w3.org>
Subject: RE: 仮想ボディと外枠の関係性

EXTERNAL: Use caution when clicking on links or opening attachments.


山本 様
みなさま

 小林 敏 です．

  Taro Yamamoto さんwrote

> 1.Upper side of the body
> 2.Mean height of the body (i.e. centerline).
> 3.Bottom side of the body

横組で仮想ボディを考えると，原点は，1は左上，2は上下中心，3は左下ということですか．880：120の位置はないということですか．

いずれにしても，どんな場合でも，原点が決まり，そこから文字サイズの応じて字面の大きさと位置も決まり，そして，仮想ボディも文字サイズに応じて原点からの距離で決まる．つまり，仮想ボディは，原点がどこにあろうが一義的に決まる．仮想ボディとその中にある字面の関係は決まってくる．あとは，その仮想ボディの配置を説明すればいいことでしょう．（これは活版モデルといえる，原点からの距離は写植モデルといえる．そして，この2つは，計算はややこしいかもしれないが換算できるでしょう，ということです．）

そうしたことを基本に，文字の配置方法は，字詰め方向も，行送り方向も，すべて仮想ボディを前提にして位置関係を説明することは可能だと思います．JIS X 4051も，いってみれば活版モデルです．

そして，jlreq-dでは，活版モデルで説明したらどうかというのが，この間の私と木田さんのやりとりだと，私は認識しています．

いずれにしても，仮想ボディで位置関係が決まれば，あとは原点からの移動量を計算すればいいことで，それは原点の位置により決まってきます．異サイズの混植や行頭・行末では，確かに原点の位置により計算はややこしくなるが，それはユーザにとっては，とりあえず必要はないことです（私は興味があるが）．

言い換えれば，仮想ボディで位置関係を説明するということは，原点の位置を回避できるということです．

そうであったとしても，880と120という数値は，仮想ボディモデルであっても必要になる場合がある．なぜなら，仮想ボディの880：120に位置に合わせたいという場合もあるからです．