W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > July to September 2014

Re: Usefulness of language annotations

From: Christophe Strobbe <strobbe@hdm-stuttgart.de>
Date: Mon, 11 Aug 2014 20:37:46 +0200
Message-ID: <53E90D7A.7020603@hdm-stuttgart.de>
To: w3c-wai-gl@w3.org

On 11/08/2014 19:10, Jens O. Meiert wrote:
> (...)
> Hence I want to fish for arguments here: How useful are language
> annotations via @lang?
> In particular:
> 1) Do user agents, including assistive technology, use this
> information in a way that is *actually* relevant and meaningful to the
> user?

Yes. For example, there are combinations of screen readers and speech
synthesisers that can automatically switch to a different "voice" or
Braille table when they encounter an explicit language switch inside a

> 2) Isn’t, or shouldn’t, language determination primarily be made a
> user agent, and not a developer responsibility?

Do you mean through language guessing in the user agent? I have rather
bad experiences with language guessing in word processors - i.e. in
documents where the languages changes between or inside paragraphs.
(Example in Word 2007: a paragraph in English contains a few words in
French; I explicitly set the language of those French words to French;
later I add some English words at the end of that paragraph; Word starts
its language guessing algorithm again sets the language of the *entire*
paragraph to English - sigh. And this is for "big" languages. And when I
use abbreviations like DVD inside a Chinese text, Word 2010 insists on
marking "DVD" as English, no matter what I do. I avoid MS Word whenever
I can but its language guessing is one only of the reasons.)

I don't expect browsers to perform better than word processors in this

Braille tables differ between languages for things such as punctuation
and "contracted Braille" (although the use of contracted Braille appears
to be diminishing, just like knowledge of Braille generally).

> 3) Does it matter at all?
It matters to users of synthetic speech and Braille. It therefore also
matters when you generate digital talking books (DAISY) out of existing

Best regards,


> I’m presenting this here (and perhaps on one or two more lists) to
> fish for arguments I may be missing. I’ve set my mind on this a while
> back [1]—and would answer 1) not usefully implemented, 2) tool
> responsibility, 3) probably not—but wonder if there is something that
> strongly disputes my view.
> Otherwise we can maybe clear up a mythical argument and requirement
> that developers need to mark up languages.
> Cheers,
>  Jens.
> [1] https://plus.google.com/+JensOMeiert/posts/PfasbPRMuX8
> --
> Jens O. Meiert
> http://meiert.com/en/
> ⚐ http://coderesponsibly.org/

Christophe Strobbe
Akademischer Mitarbeiter
Adaptive User Interfaces Research Group
Hochschule der Medien
Nobelstraße 10
70569 Stuttgart
Tel. +49 711 8923 2749

"La vie est courte, hélas! et je n'ai pas encore lu tous mes livres!" (d'après Mallarmé).
Received on Monday, 11 August 2014 18:38:10 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:33:02 UTC