TZ database content

[Paul Eggert]

> Date: Fri, 09 Feb 2001 11:32:02 -0600
> From: Syed Sajjath <Syed.Sajjath at wcom.com>

> It would be very helpful for me if there is a separate file with

> Country Name,
> Zone  short name (like ET),
> Zone Long Name(Eastern Time),
> Zone Summer short Name(EDT),
> Zone Summer Long Name (Eastern DayLight Time),
> Zone standard short Name(EST),
> Zone standard Long Name(Eastern Standard Time)
> GMT Offset (Standard)
> Currently used indicator (this will indicate if the timezone is just for
> historical purposes)

This could be automatically generated from the existing database,
except for the following problems:

* The existing database does not have long zone names.

* The existing database assumes English.

One possible way to attack this problem would be the following:

1.  Convert the existing database to use unique abbreviations for each
    distinct zone name.  For example, we would use a different
    abbreviation for Australian Eastern Standard Time than for US
    Eastern Standard time (both currently "EST"), and a different
    abbreviation for Indian Standard Time than for Israel Standard time
    (both currentl "IST").

2.  Have a new English-language table that contains

     - the short name
     - the long name

    The names must all be unique, and cannot be the same as an ISO
    3166 country code.

3.  Have a translation table per locale, which translates short names,
    long names, and ISO 3166 country codes to the language of that
    locale.  This could be maintained with the standard gettext
    mechanism. 

The translation table would allow Israeli and Indian users to both see
their familiar "IST", if they set their locale properly.

Step (3) would require modifying localtime etc. to know to look up
zone names in locales, and to provide a new interface to get the long
names.  So it's not a trivial project.

If all you care about is English, then steps (1) and (2) would
suffice, but it might cause problems for folks whose abbreviations
would change.