Suggestions about separation and ISO code

Paul Eggert eggert at CS.UCLA.EDU
Wed Nov 29 19:38:01 UTC 2006

"Jonas Melian" <jonas.esp at> writes:

> It contains a two-letter code which is recommended as the general
> purpose code, _a three-letter code which has better mnenomic
> properties_

But that part of the TZ database is not intended for end-users.  It is
intended for internal processing.  So its mnemonic properties are not
that relevant.

I just now ran my host's time zone setup tool.  (I'm running Debian
stable, a fairly old GNU/Linux distribution.)  The tool gave me a map
of the world with cities and islands scattered over it, along with a
list of text time zone names like "America/La_Paz".  Nowhere did it
use two-letter codes.

If my experience is typical, changing from two-letter codes to
three-letter codes would not help typical users.  It would affect only
developers.  But most developers, I think, would prefer two-letter
codes, even if were no cost to switching th three-letter codes.
They're more familiar with the two-letter codes, due to their
widespread use in Internet domain names.

There's a nonzero cost to switching.  Existing software reads
and and expects these files to have a certain format.  To
switch, we need a compelling technical reason.  So far, I haven't seen

More information about the tz mailing list