Re: Timezone information in HTTP

On Monday 11 June 2007 10:49, Julian Reschke wrote:
> But it *is* good to discuss it here. For instance, I'm not convinced
> it's the right approach; HTML5 is adding markup to allow user agents to
> do smarter things with time information.

  I'm willing to discuss this for as long as it takes to persuade everyone 
about its usefullness or someone persuade me that it is not required :-)

  I also want to talk about this as much as possible (but not to annoy you).

  I see that HTML 5 has markup for date, time and offsets. Perhaps the time 
offset could be embedded in the html tags. Something like:
<time offset="+0200">02:00:00</time> but this is still not what one should 
expect from HTML. Since HTML is a markup language, it may only tag the text 
and not transform it. The '02:00:00' will always be a string.

  Examples from HTML 5 draft:

<time datetime="2006-09-23">a saturday</time>
<time datetime="2006-09-24 05:00 -7">5am the next morning</time>
<time>08:00</time>

  Should it be possible for browsers to change 08:00 to their local time, 
supposing that TZ information is embedded in <time> ?

  Finally, this doesn't solve the problem of server side generated graphics, 
metadata, files etc (content that is not HTML) that include time information 
in them. Consider the MRTG and FAX examples that were mentioned in a previous 
email.

> Here's another one: I think currently the spec is not clear about
> whether it uses the RFC2616 or the RFC4234 ABNF format...

It uses the RC2616 ABNF format. In paragraph 1.3 it says:

   The HTTP header specification of this document is presented in the
   augmented Backus-Naur Form that is described in [RFC2616].

But your point is correct. I will remove RFC4234 from the informative 
references.

Received on Monday, 11 June 2007 15:08:19 UTC