Re: JAreq?

皆さん、

今のドラフトです。日本人には当たり前のことを延々と書いてあります。3章として具体的なrequirementを書くつもりです。AG
WGからnoteで出せるかどうかはまだ分かりません。

村田 真

Title: Accessibility requirements for Japanese Web Content

Editor: Murata Makoto

Abstract


This document identifies accessibility requirements arising from the
characteristics of the Japanese language and writing system. It provides a
foundation for improving Web accessibility specifications and related
technologies for Japanese users who experience reading difficulties.

1. Introduction
1.1 Purpose

The purpose of this document is to provide a common reference for the
development, maintenance, and review of Web accessibility specifications
and related technologies for Japanese. It does so by identifying
accessibility requirements that arise from the characteristics of the
Japanese language and writing system, with particular attention to the
needs of users who experience reading difficulties.

The requirements identified in this document may also be useful to
developers of authoring tools, user agents, reading systems, digital
publication technologies, and other implementations that support Japanese
content.

1.2 Relationship to Japanese Script Resources

This document assumes familiarity with the terminology and concepts defined
in Japanese Script Resources. Rather than repeating general descriptions of
the Japanese writing system, this document highlights only those
characteristics that have significant accessibility implications and
derives accessibility requirements from them.

2. Accessibility-Relevant Characteristics of Japanese

The following characteristics of Japanese have significant implications for
accessibility.

2.1 Bunsetsu

Bunsetsu is a linguistic unit consisting of a content word together with
any accompanying function words. It plays a central role in Japanese
reading and is often a more useful unit than the word for accessibility
purposes.

2.2 Kanji and Kana

Modern Japanese uses a combination of kanji, hiragana, and katakana. These
scripts have different linguistic functions and present different
accessibility challenges for different users.

1) Kanji

In compulsory education, children learn 2136 kanji characters defined by
the Japanese government. Beyond compulsory education, there is no official
list of additional characters to learn. Nevertheless, newspapers, books,
personal names, place names, and specialized documents regularly use
several thousand additional kanji outside the jōyō set. Most of these
characters occur infrequently, and many Japanese readers cannot recognize
or read all of them without assistance.

Thus, the readability of Japanese text depends on whether readers are
familiar with the kanji being used. Reading aids are therefore important
for many users, including children, people with dyslexia, language
learners, and readers encountering uncommon characters.

2) Kana

Some readers experience difficulty reading kana, including both hiragana
and katakana. The nature and severity of these difficulties vary among
individuals. For these readers, text written primarily in kana can present
additional reading challenges.

2.3 Unspaced Text

Japanese is normally written without spaces between bunsetsu. Instead,
readers often rely on the alternation of kanji and kana to recognize
bunsetsu boundaries. When text contains a high proportion of kana, these
visual cues become weaker, making segmentation more difficult.

The absence of explicit bunsetsu boundaries can present a major obstacle
for some readers, especially those who experience reading difficulties.
Providing automatic phrase segmentation or allowing users to insert spaces
between bunsetsu can significantly improve readability for these readers.

Example: The sentence 太鼓とバチを持ってきて ("Bring the drum and the drumsticks.")
becomes たいことばちをもってきて when written entirely in hiragana. A Japanese reader
with dyslexia reported interpreting this as たい｜ことば ｜ち (red snapper | word |
blood)をもってきて, rather than たいこ｜と｜ばち(drum|and|stick)をもってきて, and therefore
could not understand the instruction. This example illustrates how the
absence of explicit bunsetsu boundaries can interfere with text
segmentation and comprehension.

2.4 Vertical and Horizontal Writing

Japanese is commonly written in both vertical and horizontal directions.
Books, newspapers, and literary works are often published vertically, while
web pages, technical documents, and many digital publications are commonly
presented horizontally.

While many readers can read both writing directions without difficulty,
some readers find vertical text substantially more difficult to read than
horizontal text. Reports from Japanese accessibility organizations and
readers with dyslexia indicate that vertical writing may present a
significant accessibility barrier.  A much smaller number of readers report
the opposite preference.

2.5 Ruby Annotations

Ruby annotations provide pronunciation or other explanatory information
associated with base text. In Japanese, they are most commonly used to
indicate the pronunciation of kanji.

The need for ruby annotations varies considerably among individuals. Some
readers require ruby for nearly all kanji, while others need it only for
unfamiliar or difficult characters. The amount of ruby required varies
considerably among individuals.

Ruby annotations themselves can also create accessibility barriers. Some
readers have difficulty distinguishing ruby text from its base text, while
others, including some people with low vision, find conventional ruby text
too small to read comfortably. Allowing users to adjust the size and color
of ruby annotations, or to choose other alternative presentations, can
substantially improve readability.

Unlike ordinary text, ruby annotations form a parallel annotation
associated with the base text rather than a simple sequence of characters.
This means that Japanese text cannot always be regarded as a simple
sequence of characters.

2.6 User-Perceived Characters

In Japanese, user-perceived characters are well represented by Unicode
extended grapheme clusters. Consequently, this document uses extended
grapheme clusters as the basis for character-level accessibility
requirements.

Note: Emoji require special consideration and are not discussed in this
document yet.



2026年7月17日(金) 17:48 Yasuo Kida <kida@mac.com>:

> 別文章を作るのは賛成です。作って存在しているということ自体が意味を持つと思います。
>
>
> ただし、完結した文書としてjlreq-dにおいても重要な点をカバーしておくべきかと思います。jlreq-dだけ見る、というケースも考える必要があります。全ての人に知って欲しいことはjlreq-dにあり、そしてアクセシビリティに目を向けた人のための専用文書がある、という二段構成はどうでしょう。
>
> 木田
>
> 2026/07/17 15:52、MURATA <eb2mmrt@gmail.com>のメール:
>
>
>> >Accessibility Guidelines
>> WGと話していると、基本的な前提（たとえば単語や文節）について認識の違いが大きいことに気づかされます。もちろん縦書きやルビなんて名前しか知らない。
>> >
>>
>>
>>
>> また，執筆編にとっても，アクセシビリティは重要な要素ではあるが，内容の1項目にしかならない．また，この執筆編は，ある意味で日本語をある程度は常識として持っていることが前提です．
>>
>>
>> ということを考えれば，jlreq-dでもアクセシビリティは重要だということは触れるとして，別に日本語を知らないということを前提にして，別にした方がよいかもしれない．
>>
>
> やはり、別文書のほうがよさそうですね。
>
> RichardのJapanese Script Resources <https://www.w3.org/TR/jpan-lreq/>
>  を私は知りませんでした。組版に深入りせずに、日本語テキストとは何かを説明しているので、広範囲の人に読んでもらえる文書だと思います。
> これを参照しつつJAreqを書くと、私がさっきまで考えていたより、はるかに楽になります。
>
> --
>  --
> 東日本国際大学客員教授
> 村田 真
>
>
>

-- 
 --
東日本国際大学客員教授
村田 真