W3C home > Mailing lists > Public > www-international@w3.org > April to June 2014

Re: [css-text] I18N-ISSUE-335: health warnings for when language not known

From: Mark Davis ☕️ <mark@macchiato.com>
Date: Wed, 23 Apr 2014 11:43:04 +0200
Message-ID: <CAJ2xs_HiJy90VM5Ux6ioT=o9EQ6vF_qQVib0V_zNGJsvz75q0A@mail.gmail.com>
To: Koji Ishii <kojiishi@gluesoft.co.jp>
Cc: "Phillips, Addison" <addison@lab126.com>, "CSS WWW Style (www-style@w3.org)" <www-style@w3.org>, www International <www-international@w3.org>
> should have a health warning that language-sensitive actions are not
possible and should probably not be guessed at

The term "guessed at" is unnecessarily pejorative. In many cases, the only
recourse—and best user experience—is to use language detection. Moreover,
there are clearly times when even explicit language settings are clearly
wrong: <span lang="en">朝日新聞デジタル</span>. That particular example may be
unlikely, but the language tags on a whole document are often incorrect for
all or part of the document.

The most that should be said is something like:

Where the language cannot be determined with a high degree of confidence,
language-sensitive actions should be avoided.


Mark <https://google.com/+MarkDavis>

 *— Il meglio è l’inimico del bene —*


On 23 April 2014 10:25, Koji Ishii <kojiishi@gluesoft.co.jp> wrote:

> > A number of requirements or suggestions in the document require knowing
> the language in which the document is intended to be interpreted or of the
> span of text affected by the styling. Since it is possible to have a
> document or span of text for which the language is unknown, the document
> should have a health warning that language-sensitive actions are not
> possible and should probably not be guessed at. There is a hint of this in
> the statements in Section 2.1 Example 2, but it isn't clear whether or not
> it generalizes beyond that example (it should) and it doesn't exactly leap
> out at the reader.
>
> I’m not sure if it’s a good idea to say “language-sensitive actions…should
> probably not be guessed at.” So the suggestion is, even if code points
> used, fonts used, or the current system language/region indicates that the
> content is Arabic, UA should not apply any Arabic specific rules?
>
> Can you elaborate the motivation to prohibit that?
>
> /koji
>
>
>
Received on Wednesday, 23 April 2014 09:43:54 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 23 April 2014 09:43:55 UTC