Re: Call for comments: HTML5.x time zone proposal [I18N-ACTION-327]

On 8/8/2014 1:04 PM, Arle Lommel wrote:
> The specification has the following:
>
>     The |tz| attribute has no meaning when applied to |month|,
>     |yearless date|, |week|, or |duration| values.
>
>
> I’m wondering why yearless date is treated this way, especially as 
> Addison’s last comment states:
>
>     …I personally tend to think of week and month values as "floating
>     times". However, there are probably cases where they are not
>     intended to be floating and where time zone would apply…
>
>
> I think yearless date is even /more/ likely to be non-floating since 
> recurring events are often tied to a calendar date. For example:
>
>     Payment for the big bowl of sauerkraut is due annually by the
>     close of business in Albuquerque on <time datetime="07-15"
>     tz=“America/Phoenix">July 15</time>  or on the next working day if
>     that date falls on a weekend or holiday.
>
>
> So I would argue that tz should have a meaning in this case as well.
>
> For duration, I think tz would have no meaning (since tz could not 
> impact a duration), but the other items can all be bound to a specific 
> location.

Is that statement correct? Take:

"You have one week from the date of submission to complete the application".

This expresses a duration (one week), but the number of hours to the 
deadline will vary if the week crosses a daylight savings event. (I've 
used prose for the example, but you could substitute a numeric 
expression, e.g. "7 days").

To correctly calculate a deadline based on this information, you'd need 
both the current date and the time-zone to which the calculation 
applies, to know whether a week has 167, 168 or 169 hours.

A./
>
> Best regards,
>
> Arle
>
>

Received on Saturday, 9 August 2014 00:25:55 UTC