Day Periods (was Re: Additional French Time Patterns ?)

[Edwin Hoogerbeets]

Speaking of the day period support, the UTS #35 specification does not
make it clear whether the start/end times are in the local time zone or
UTC time, and whether or not they use the local time system/calendar or
the Gregorian time system/calendar. From the looks of it, they are in
local time and in the Gregorian time system. Now that I think of it, it
would be pretty silly to specify it in UTC, but what I am asking for is
that the spec be amended to make this explicit and clear, instead of an
implicit assumption.

When I say "local time system/calendar" vs. "Gregorian time system",
what I am talking about is Ethiopic times which are also 12-hour, but at
different times of day than our regular Gregorian time system. The
Ethiopic day starts at sunrise, which is right about 6:00am Gregorian
time. Being near the equator, this doesn't vary much in Ethiopia. So,
12:00 morning Ethiopic = 06:00 Gregorian, and 6:00noon Ethiopic = 12:00
Gregorian, 12:00evening Ethiopic = 18:00 Gregorian, etc.

For consistency and ease of implementation, I think the day periods in
CLDR should always be specified in Gregorian time as the single
normalized form, and it should be up to the implementation of the date
formatter to translate 06:00 Gregorian to "ጠዋት 12:00 " when formatting
an Ethiopic date/time with the Ethiopic time system/calendar.

Now having said that, the Gregorian calendar is sometimes used in
Ethiopia as well, mostly when dealing with foreign companies or people
so as not to confuse them, and the regular Gregorian AM/PM meridiems are
generally used with Gregorian time formats. That is, the day periods are
dependent on the calendaring system in use, just like the date formats
already in CLDR. I think we should add a "calendar" attribute to the
dayPeriodRuleSet or dayPeriodRule elements, with the default value when
not specified being "gregorian". Or are day periods perhaps a separate
concept from meridiems?

Can we clarify the spec on these points?

(And before you ask, yes, I have already submitted ticket #8473 to
support am-ET in the day periods!)

Thanks,

Edwin

On 05/05/2015 05:52 AM, Mark Davis ☕️ wrote:
>
> As far as the am/pm goes, we are rolling out new dayperiod support. It
> is targetted mostly at languages that use different periods than AM/PM
> or 24 hour, but there is support for longer or shorter forms.
> Cf. http://cldr.unicode.org/translation/date-time-names#TOC-Day-Periods-AM-and-PM-
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://unicode.org/pipermail/cldr-users/attachments/20150505/be3e5284/attachment.html>