[threeten-dev] Possible addition of pattern letters in CLDR

[yoshito_umaoka at us.ibm.com]

Dan Chiba <dan.chiba at oracle.com> wrote on 12/13/2012 06:09:08 PM:

> For the discrepancies between 310 and CLDR, it would be ideal for us
> to be able to resolve them one way or the other. Personally I look 
> at them as issues because they often cause problems in similar cases.
> 
> I don't think there is an issue in "I" in CLDR. Printing "I" out of 
> an offset is an attempt to format a nonexistent field. It is the 
> same as attempting a time field when the value is date only or vise 
> versa or anything similar.
> 
> Another possible difference is punctuations. LDML says "... certain 
> ASCII punctuation characters may become variable in the future (for 
> example, ":" being interpreted as the time separator and '/' as a 
> date separator, and replaced by respective locale-sensitive 
> characters in display)." I think locale sensitive pattern letters 
> for punctuations are desired, or formatting to the local user's 
> expectation could be difficult to achieve. In particular, the 
> fractional seconds would be difficult, as predefined locale 
> sensitive patterns with fractional seconds are hard to find. 
> 
> Yoshito, would you let us know if you would like us to file a formal
> request for new pattern letters, please?
> 
> Regards,
> -Dan

I've already filed a ticket to CLDR trac system [
http://unicode.org/cldr/trac/ticket/5506].
I'm a CLDR TC member and I've already asked other CLDR TC folks to look 
into this.

I'm going to propose:

 - Add n/N/I to the LDML specification with the definition used by JDK
 - Add X for ISO type format, keeping "ZZZZZ" as an alias of XXX (long 
ISO)

We probably ended up to define another letter for what CLDR calls 
"locaized GMT format". There is another requirement for supporting 
shortest representation of localized GMT format.

For punctuation variable... In my honest opinion, I think there are many 
issues with introducing punctuation variables.


-Yoshito