Notes

Get involved

You can take part in this work.

First, get a suika.fam.cx account (if you have not) and join the spec-ja group from your user page.

Then,

Translate a paragraph that is not translated yet

Double-click the paragraph and input the translated text into the second textbox.

Fix a typographical error in the translated text

Fix a paragraph that has been translated incorrectly

Double-click the paragraph in which the error is found and fix the error.

Alternatively, you can report the problem to the _bugs page.

Discuss the choice of a Japanese term

Discuss any other topic

Write on the _notes page.

Introduction to the interactive translation interface

For the translating of HTML5 specifications, a new interactive paragraph-level editing interface has been developed. Since the HTML5 specification adopts a new specification development model — publishing a large Editor's Draft with sections in different levels of maturity, in daily basis — it is unrealistic to follow the traditional model for the development of a Japanese translation of a technical specification — obtaining a copy of a relatively stable Working Draft or a final Recommendation, translating the whole document at once, and then publishing the finalized Japanese translation.

The new interactive translation editing interface (and its backend management tools) enables a different style of the translation work:

• Use of an interactive Web interface. A translated version of a paragraph can be added or modified by just choosing the paragraph to add or modify and then editing the text in a Web browser. You no longer have to prepare your environment for the translatioon work; you just need an interactive standard-compliant Web browser.
• Paragraph-level editing. The minimum unit of text is paragraph (in this context the contents of a heading, list item, or table cell is also referred to as “paragraph” unless it contains another paragraph). The Web interface and the management tools handle text in per-paragraph basis. Since the original spec text is frequently updated and paragraphs are frequently modified or reorganized, handling translated texts in this level seems better choice than other possibilities (e.g. section-level).
• Automatic synchronization with the original spec text. Translation of an old draft version of the specification is rather harmful, since readers of the translation don't check whether the text is up-to-date or not in general and therefore they would not be even aware that what they read is out-of-date. This system is so set up that any change to the original specification text will be reflected to the translated version within a day. If a paragraph that was already translated is then updated in the original text, the paragraph is taken over by the newer English version. It would take a step backward, but is better than leaving out-of-date translation as is.

How to use Web interface

By double-clicking a paragraph that can be translated, or that has been translated, a editing form is shown as below:

Original text:

( Pattern)

Translated text:

Tags:

Save and close Close (

Load: 200 OK

; see also license)

The “Original text” field

The original English text in the paragraph of the original document. It may contain HTML tags and references as in the original document. A run of white space characters is considered equivalent to a space (U+0020) character. Leading and trailing white space characters are ignored.

The “Translated text” field

The translated text for the paragraph. It may contain HTML tags and references.

In addition, you can use special notation for RFC 2119 keywords as:

• しなければ[[MUST:なりません]]。
• しては[[MUST NOT:なりません]]。
• [[REQUIRED:必須]]です。
• [[SHOULD:するべきです]]。
• [[SHOULD NOT:するべきではありません]]。
• して[[MAY:構いません]]。

The “Pattern” flag

If this flag is not set, then the “Original text” is compared with paragraphs in the original document by exact match (except for the aforementioned white space normalization). You should use exact match (leave the flag unchecked) as far as possible.

If this flag is set, the “Original text” is compared with paragraphs in the original document by pattern matching rule, where a * character in the “Original text” matches one or more characters. If there are more than one patterns matching the same paragraph, then only the longest pattern is considered as matching. When this flag is set, variables $d, where d is the index of the * (starting with 1) in the “Translated text” is replaced by the substring matching the corresponding * in the pattern.

The “Tags” field

The tags associated to this paragraph. One tag can be specified in a line. Tags are not used in general; it can be specified for the convinience of the paragraph search. Common tags are:

cell

td/th contents.

dt

dt contents.

headings

hn/caption contents.

item

li contents other than steps.

non-normative

Non-normative contents.

step

Steps.

Any other tags may also be specified. Tags are case-ensitive.

The “Save and close” button

Save changes and close the form.

The “Close” button

Close the form without saving any change.

Text next to the buttons

Progress of transactions is shown here.

Translation Guidelines

Use desu‐masu style.

Markup RFC 2119 keywords, otherwise what is requirement and what is not would become unclear, since there is no upper‐ and lowercase letters in the Japanese script. Note that keywords in non-normative parts (e.g. examples) of the spec are not RFC 2119 keywords, so don't markup them.

More to come...