[whatwg] <time> element feedback

Wed Sep 1 02:44:45 PDT 2010

Am 31.08.2010 22:21 schrieb Martin Janecke:
> Am 31.08.10 21:40, schrieb Aryeh Gregor:
>> On Tue, Aug 31, 2010 at 5:25 AM, Martin Janecke<whatwg.org at kaor.in>  
>> wrote:
>>> Besides,<time>2010</time>  in a British news article would allow 
>>> users e.g.
>>> in Japan to have these dates displayed as 平22年. That's clearly an 
>>> advantage
>>> over the number 2010 alone.
>>
>> I would say the opposite.  If they can read the English news article,
>> they'll necessarily know what "2010" means.  But they might not be
>> able to read Japanese.  Maybe they're borrowing a Japanese person's
>> computer, for example, or maybe the browser's idea of the user
>> language is otherwise wrong.
>>
>> Also, content that behaves differently based on the browser settings
>> of the viewer is confusing and can cause hard-to-debug problems.
>> Users will think that the author of that British article actually
>> wrote out a Japanese date, and be completely at a loss to explain why.
>>   Even if they can actually understand the date, the incongruity will
>> look like a bug.
>>
>> It could be outright misleading if there are two year display formats
>> that look the same but actually have different meaning.  A plain year
>> number in Arabic numerals like 2010 could refer to any number of
>> totally different year-numbering conventions, and the only way to tell
>> them apart currently is the page's context.  Having the browser change
>> the number to some convention that doesn't match its surroundings
>> makes it impossible to guess the convention.
>>
>> And finally, it just looks weird.  I would find it extremely strange
>> to have all dates on pages I'm reading replaced with Hebrew dates,
>> even though I understand those just fine.  I wouldn't want that at
>> all, and I find it hard to believe that many actual users do in real
>> life.
>>
>> Basically, any kind of attempt to have browsers localize dates that
>> are actually displayed in content is a terrible idea, and the spec
>> should remove all mention of any such thing.  I'm pretty sure I've
>> said all this before, though.
>>
> 
> I understand your point, the situation you describe would be unfavorable 
> indeed.
> 
> However, there's no need to make this unfavorable. The localized display 
> of times and dates can be realized via tooltips for example, as it is 
> often seen with abbreviations in texts. The localized date doesn't have 
> to be a replacement for the original date string but can be a helpful, 
> explaining addition.

This is a nice idea, but localisation should then be based on the 
language of the context of the time element, not based on the browser 
language:

<html lang="de">
   <body>
     <p>Die Party ist <time datetime="2010-09-01">heute</time>.</p>
   </body>
</html>

The tooltip might then display "1. September 2010" in German - 
independent from the computer or browser language. As Aryeh stated, 
displaying a Japanese or English date here on the computer in an 
internet café would be highly disturbing - even in a tooltip. 
Localisation should not mess around with the content unless explicitly 
triggered by the author.

-- 
Markus