Re: Web Notifications I18N Review: [I18N-ISSUE-161, I18N-ISSUE-162]

On Jun 28, 2012, at 1:42 , Anne van Kesteren wrote:

> On Thu, Jun 28, 2012 at 8:13 AM, Norbert Lindenberg
> <w3@norbertlindenberg.com> wrote:
>> Or can there be senders that just have no information at all about the user's
>> preferred languages?
> 
> No, the application "knows" the language (assuming it is localized to
> begin with). That's always the case on the web, localization is a
> server affair. Having said that, the notification can still be in
> another language. E.g. when I book a ticket via Flying Blue rather
> than KLM the email I get will be in French (god knows why) which
> presumably would be partially quoted in the body of a notification.

OK - the spec doesn't actually say that senders must be web applications, but you get to pick the target you're designing for.

Your example shows that even multilingual web applications may rely on infrastructure (owned by Air France, I'd guess) that doesn't know about or respect the user's language preference, but that's a problem within the app that the web notification API doesn't have to solve.

>> Language detection unfortunately is quite unreliable for short strings, so there's
>> no equivalent "auto".
> 
> I tend to think we should wait and see if people run into actual
> problems here. It is quite trivial to add language in the future too.
> Is it normal for whenever an application exposes just a string there's
> always a language and direction field associated with it? Doug seems
> to be saying direction is already missing sometimes. Language is even
> more obscure (and often wrong on the web, too).

It is not normal yet for applications to properly identify the language of the text they handle, but there's a clear need for it. Just imagine a screen reader trying to read that French email to you pronouncing it as if all words were Dutch.
http://www.w3.org/International/questions/qa-lang-why

Norbert

Received on Thursday, 28 June 2012 16:18:09 UTC