[Microformats-dev] ISO Dates

[Tantek Ç elik]

On 7/19/05 10:36 AM, "Ryan King" <ryan at technorati.com> wrote:

> On Jul 19, 2005, at 8:35 AM, Tantek Çelik wrote:
>> On 7/19/05 8:01 AM, "Ryan King" <ryan at technorati.com> wrote:
>>> On Jul 19, 2005, at 5:26 AM, Brian Suda wrote:
>>> 
>>>> On the discuss list there has been so talk about ISO date formats.
>>>> Currently, X2V only supports Dates without the '-' seperators (which
>>>> is wrong and will be changed).
>>> 
>>> Good.
>>> 
>>>> X2V currently takes the date with a timezone and attempts to convert
>>>> it to a UTC date so that importing applications can adjust
>>>> accordingly. This is ALOT of work in XSLT, which i'm wondering if it
>>>> is needed at all?
>> 
>> Absolutely.  This is a requirement of iCalendar in particular.
> 
> Ah, I've misread the spec. Apparently, you can only specify TZID's
> not timezone offsets.

Exactly.

> Why is it, then that we publish hCalendar with offsets?

For a decent chance at human verifiability between the human readable
version and the machine readable version.

By allowing the publishing of the datetime with an offset, the actual
datetime value is likely to be in the "native" timezone of the person
authoring the event and thus will likely "match up" with the human readable
text.  Just hover over any typical hCalendar encoded event's start datetime
for example, and I bet you'll have a good chance at seeing the same
information (if punctuated a bit differently).

It is *much* more difficult to verify UTC encoded times (and I've spent a
lot of time with both).  Time zone math is difficult even for computers to
get right, never mind humans in their head.


> Why not allow TZID's?

The resulting structure that is required (see RFC2445) for defining TZIDs is
much more complex.

Thanks,

Tantek