[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Timezones and Olson Tables
IANA
should be kept informed by sending email to iana@iana.org. Publishing a RFC under the
supervision of the IESG Area Directors is appropriate in this case, and IANA is
willing to participate in this process when necessary.
Thank
you.
Josh
Elliott, Administrator
IANA
As you recall
from the minutes of the Atlanta meeting and dialogs on the list, we have to
come up with a way to manage a common table of timezones or at least agree
on what we all will call them, etc in CAP. One of the things that was
suggested was to use the Olson Timezone tables. I talked to Keith
Moore about this and about registration and administration with IANA.
Here is what he replied and suggested. They are excellent points
and a good definition of the problem and issues.
Keith wrote:
- - - - - -
The normal way to set up an
IANA registry is to publish the rules
for maintaining that registry in an
RFC. These rules need to specify
the requirements for adding or
changing entries in the registry
(what pieces of data are needed and how
they are to be vetted).
The calsch working group should get consensus on
these rules and then
they should be sent to the IESG with a request to
publish as a
BCP RFC. The same RFC could also specify which
portions of the
current Olson database should be initially included in
the IANA registry,
that way, it wouldn't be necessary to separately
register each timezone
from the Olson database.
The rules should
be developed with someone from IANA in the loop;
contact iana@iana.org to
establish contacts.
p.s. although I am
very supportive of the idea to have IANA
maintain the Olson timezone
registry, I do have two concerns:
1. we need to make sure that
everyone, including Olson, understands
and agrees on issues like
copyright ownership, distribution rules
(presumably free to
everyone), and change control
2. I think calsch should attempt to
establish globally-scoped
timezone names, probably using
different (or alternate) names
to those used in the Olson
database. If I understand correctly
Olson prefers to
associate timezones with the names of continents
and nearby
cities; but since most timezones are set by political
jurisdictions I suspect that we will get more precision (and the
timezones will be more "obvious") if we name the timezones after
the political regions for which those timezones are established.
So I would recommend that primary timezone names be:
country code
for the top level of timezone name, followed by
subdivisions
(for those few countries for which it is necessary),
followed
very occasionally by ad hoc divisions when these are
necessary.
E.g. /NO (for time in Norway) and /US/Eastern or
/US/Indiana
(for time in Eastern US (EST/EDT) or the portion of
Indiana
that always uses CST. Note that the requirements
for calsch
timezone names (which are mostly used to indicate
events in the
future) may be different than those for the
traditional uses of
the Olson database (which oddly enough are
mostly used to
indicate events in the past - e.g. file
modification times).
Of course not all timezones are set by
countries, so we would
still need a way to register other kinds
of timezones.
Ideally countries could eventually get direct change
control
over their timezone records at IANA.
- - - -
Ok list, comments, suggestions.
Doug, I know you have been working on this a lot. Your
thoughts?