Proposal: Change what is meant by the "PST" timezone

I need your opinion!

  http://www.jtcsv.com/cgibin/icu-bugs/formatting?id=3477

This bug describes some discrepancies in the ICU timezone data.

One issue that may be controversial is this one:  What is meant by "PST"?

Since JDK 1.1, Java code has been able to specify "PST" in order to 
retrieve what is really "America/Los_Angeles".  ICU4C and ICU4J have also 
supported this alias.

That is, "PST" (which stands for Pacific *Standard* Time) has been used as 
an alias for a timezone that includes both Pacific Standard Time (GMT-8) 
and Pacific *Daylight* Time (GMT-7).  The proper "short name" for this is 
"PST8PDT", not "PST".  However, in the past (both in ICU4C/J and in the 
JDK) "PST" has been used to mean "PST8PDT".

*** Part of the justification is probably:  When someone says "PST", they 
mean PST8PDT.  Unlike some other zones ("MST" == America/Phoenix) where 
you really have an area that observes xST all year long, that doesn't 
happen in the Pacific zone -- so anyone saying "PST" really means PST8PDT. 
***

This might seem like a fine distinction, but the bug proposes to change 
this situation.

Before:
  PST => America/Los_Angeles

After:
  PST     => GMT-8
  PST8PDT => America/Los_Angeles

This will, of course, break any code that assumes PST == 
America/Los_Angeles.

There are other options.  We could keep "PST" as an alias for "PST8PDT" 
since "that's what users really mean."

NOTE: While this email describes "PST", there are other proposed changes. 
"PST" is chosen because it may be more familiar to most readers.  The 
"PST" issue is illustrative of similar issues with other zones.

Please reply to this mailing list with comments.

[Alan S Liu/San Jose/IBM@IBMUS; [email protected];; IBM Globalization; 
5600 Cottle Road; San Jose, CA 95193;; (408) 256-3155]