[Unicode]   Common Locale Data Repository : Bug Tracking Home | Site Map | Search

CLDR Ticket #11375(new)

Opened 3 months ago

Last modified 4 days ago

Start date for the Meiji era is wrong for the Japanese Calendar

Reported by: jefgen Owned by: anybody
Component: other-supplemental Data Locale:
Phase: dsub Review:
Weeks: Data Xpath:


I initially filed a bug on ICU for this, but with the pending change to use the data from CLDR, the suggestion from Yoshito was to file a bug on CLDR itself. :)

ICU Ticket Link: https://unicode-org.atlassian.net/browse/ICU-20024

This is regarding the start date listed in the "supplementalData.xml" file for the Meiji era.


Copy/pasted from the ICU ticket:

We have reports that the start date for the Meiji era/period (in the Japanese Calendar) in ICU is incorrect.

The data in ICU currently lists the start date for the era (in the Gregorian calendar) as starting on 1868-9-8.

From the file: icu4c/source/i18n/japancal.cpp

//  Gregorian date of each emperor's ascension
//  Years are AD, months are 1-based.
static const struct { 
    int16_t year;
    int8_t  month;
    int8_t  day;
} kEraInfo[] =  {
    //  Year  Month Day
    {   645,    6, 19 },   // Taika   0
    {  1868,    9,  8 },   // Meiji    232
    {  1912,    7, 30 },   // Taisho   233
    {  1926,   12, 25 },   // Showa    234
    {  1989,    1,  8 }    // Heisei    235

However, it seems that this is actually the date in the Japanese Lunar calendar. When converted to the Gregorian calendar, the date should actually be 1868-10-23.

In other words:
1868-10-23 (Gregorian) == 1868-9-8 (Japanese Lunar)

So it seems like it should be this instead:

    {  1868,    10,  23 },   // Meiji    232

So for the CLDR data, it seems that the line:

    <era type="232" start="1868-9-8"/>

perhaps should be this instead :

    <era type="232" start="1868-10-23"/>


Change History

comment:1 Changed 3 months ago by jefgen

  • Cc kristi added

comment:2 Changed 3 months ago by jefgen

Note: The comment from Yoshito in the ICU-TC call was that the dates prior to 1872 might all be somewhat suspect due to the confusion between the Lunar vs Solar calendars.

comment:3 Changed 3 months ago by jefgen

  • Keywords jaEra added

comment:4 Changed 7 weeks ago by mark

What we should look for is the simplest specification for developers to have code to interpret the values. I believe that is proleptic Gregorian, so that all of the date fields for all eras in all calendars (and elsewhere in CLDR) can be interpreted with the same code.

That would involve converting all dates that are not preleptic Gregorian (suspect that is basically Japanese eras) to be proleptic Gregorian.

comment:5 Changed 4 days ago by mark

  • Milestone changed from UNSCH to to-assess

Add a comment

Modify Ticket

as new

E-mail address and user name can be saved in the Preferences.

Note: See TracTickets for help on using tickets.