W3C home > Mailing lists > Public > www-international@w3.org > July to September 2001

RE: Time

From: Carl W. Brown <cbrown@xnetinc.com>
Date: Sat, 11 Aug 2001 11:46:45 -0700
To: "John Cowan" <cowan@mercury.ccil.org>
Cc: <www-international@w3.org>
Message-ID: <FNEHIHOMIIDPDCIFEJEGMEDLCIAA.cbrown@xnetinc.com>
John,

>
> > If you want to get technical Kiritimati for example is GMT+14:00 not
> > GMT-10:00.  If you want to really get nasty I don't know of any
> time routine
> > that deals with historical dates.  For example last year Sidney
> changed the
> > DST dates because of the Olympics.  So last years dates should use a
> > different set of dates to calculate DST.
>
> The tz date routines, which are used in modern Unixoid systems, do in fact
> handle historical dates extremely well.  The current versions are
> pointed to by http://www.twinsun.com/tz/tz-link.htm , with much
> other information as well, including alternative C and Java
> implementations.
>

Are you referring to routines based on the TZ= variable?  The Windows
implementation is so bad that it only looks at the TZ offset and uses the US
DST formula everywhere.  You have to use the Windows proprietary API to do
any better and even that has limited historical implementation.  Sydney 2000
was implemented but if I remember correctly only for Win/Me.

I am familiar with the Olsen data which provides for historical routines.
But I have yet to see the full data implemented. Your page has lots of
links.  Does one point to an actual implementation?

Carl
Received on Saturday, 11 August 2001 14:46:51 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 19:16:57 GMT