FW: Comments and a question (VTIMEZONE from zoneinfo)
Olson, Arthur David (NIH/NCI) [E]
olsona at dc37a.nci.nih.gov
Thu Sep 15 21:39:39 UTC 2011
I'm forwarding this message from Phil Budne, who is not on the time zone mailing list.
Those of you who are on the list, please direct replies appropriately.
--ado
-----Original Message-----
From: Phil Budne [mailto:phil at ultimate.com]
Sent: Thursday, September 15, 2011 5:33
To: tz at lecserver.nci.nih.gov
Subject: Comments and a question (VTIMEZONE from zoneinfo)
Some meta-comments, some comments, and a question:
The page at http://www.twinsun.com/tz/tz-link.htm
says: "Please send corrections to this web page to the time zone
mailing list."
Reading the directory at
http://www.twinsun.com/tz/
shows the html file hasn't changed since 2008-06-30.
Until I checked the twinsun.com WHOIS entry and saw it's owned by
(current tz editor) Paul Eggert, I was wondering about whether to
trust the page at all!
At some point I lost the twinsun URL, and couldn't find a link in
any obvious places, like Wikipedia, or ftp://elsie.nci.nih.gov/pub/
So my first comment is, the tz-link page needs more meta information
(eg; date of last change & author). If the tz-link page is being
maintained, there should be visible links to it in obvious places, and
if it isn't, the page should have a comment reflecting this. It would
be useful to have a .html file at ftp://elsie.nci.nih.gov/pub/ with
current information, a link to current information, or at least a link
to dated, but possibly useful information!
Since this is my first public comment on the TZ list, let me express
my long time admiration (I remember 4BSD when the rules were compiled
into libc, and systems like TOPS-20, where they were in the "kernel")
to all, past and present who have endeavored to make sense of this
herculian problem (the proportions and magnitude of which are so
truely perverse, that they could only have been made by mankind which,
unlike nature, has no collective ability (or perhaps more succinctly
or single authority powerful enough) to conform to Occam's Razor.
As a side comment, there's nothing more Talmudic than
commentary on commentary, so the appreciation of the tzdata
files at
http://blog.jonudell.net/2009/10/23/a-literary-appreciation-of-the-olsonzoneinfotz-database/
as talmudic is especially apropos!
Finally, I know it's unlikely to change, but I find the fact that the
tzdata tar files extract to the current directory (as opposed to
creating a new directory) to be, to be kind "quaint". And I'll
confess, I have a project that's so dated (a 1990's "finger" program,
from back when exposing personal information on the Internet wasn't
considered a problem) whose tar file is still available on my FTP
server that has the same "charming" behavior of unloading it's
contents into an unsuspecing extractor's current working directory.
On to non-meta commentary:
The link for "vzic":
http://www.dachaplin.dsl.pipex.com/vzic/
is dead.
There is a vzic project at:
http://vzic.sourceforge.net/ and http://vzic.sourceforge.net/
Registered in 2006, and unchanged since 2010.
http://code.google.com/p/tzurl/
says "This project is a fork of the vzic project."
and appears to be under active development/maintenance.
Related pages/projects are:
http://tzurl.org/
which supplies "the latest Olson DB timezone definitions in iCalendar format"
(generated using "tzurl")
And "ical4j"
http://code.google.com/p/ical4j/
"A suite of Java libraries for managing iCalendar and vCard data"
and
http://wiki.modularity.net.au/ical4j/index.php?title=Main_Page
Which at
http://wiki.modularity.net.au/ical4j/index.php?title=Timezones
reads:
iCal4j provides it's own TimeZone implementation as opposed to
using the default Java timezone implementation. This is to
ensure that all timezones in iCal4j can be accurately
represented using a VTIMEZONE component in calendar data.
The iCal4j timezone is essentially an implementation of
java.util.TimeZone that is backed by a
net.fortuna.ical4j.component.VTimeZone instance. Whilst a
timezone instance may be created explicitly, typical usage is
to obtain a TimeZoneRegistry instance and retrieve the
timezone instance from the registry.
And FINALLY, a question:
Has anyone heard of code that reads binary tz data, and outputs
VTIMEZONE info? Extracting rules for a specific year is sufficient.
I can SWEAR I once saw an "unzic" program, but maybe I'm having a new
bout of a particular form of senior moment I've experienced before,
where I remember things that have never happened.
We've been using vzic, but it isn't a great solution since RedHat
Linux doesn't supply an RPM for the timezone source files that vzic
needs to run, so we're stuck having to track version of the tzdata RPM
on the specific version of RedHat that the product installs on.
I got involved because I felt the seismic activity when the problem
landed on the desk of an engineer without sufficient experience to
grasp of the enormity of the hairball that had just hit him.
Thanks
Phil
More information about the tz
mailing list