Re: Timezone discrepancies

From: Thomas Lockhart <lockhart(at)alumni(dot)caltech(dot)edu>
To: Peter Eisentraut <peter_e(at)gmx(dot)net>
Cc: Ed Loehr <eloehr(at)austin(dot)rr(dot)com>, pghackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Timezone discrepancies
Date: 2000-09-21 15:20:59
Message-ID: 39CA275B.E93685FD@alumni.caltech.edu
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

> Another comment: I just peeked and, to be blunt, the time zone database
> looks like a mess.

Great! (Well, not great, but you know...). Let's clean it up. But we'll
need to get representation from the various regions covered to avoid
dropping useful fields.

> For example, there's CET (central european time), but
> not CEST (central european summer time). Instead there's CETDST, which
> I've never heard used.

I recall working on "MET DST" (note space) at the request of someone in
"MET". The "DST" qualifier works for every "standard timezone". Have you
seen this usage? "CETDST" is probably there for historical reasons, to
handle this case before we could manage the standalone "DST" qualifier.

> Then there's MEST, MET, METDST, MEWT, MEZ, all of
> which are supposed to be "Middle Europe" variations, none of which I've
> ever heard of. (MEZ and MESZ are the German translations of CET and CEST,
> but as listed they claim to be English terms.)

Ah, that may be. So I was talking with a German or Austrian or ??
earlier...

MET shows up in my zic timezone database. We should of course retain
entries for all corresponding entries in those databases, across the
various platforms we support.

> Also I've never heard of
> "Dansk Normal Tid" (DNT) or "Swedish Summer Time" (SST), both of these
> places use Central European Time.

The use of CET in those countries might be modern developments (since
1986??) or perhaps "DNT" and "SST" are much older. Most of the character
string representations for timezones came from Postgres' pre-history at
Berkeley, and I just carried them forward.

I've found that Sun seems to have more accurate timezone support than
other systems, at least for pre-1947 details. And afaik they do not use
zic so we should look at both of those to get a more complete story.

> There are several other obscure candidates where I don't have direct
> geographic knowledge, such as "Moluccas Time" (MT, I though that would be
> Mountain Time), or "Seychelles Time" (SET).

In the US, all time zones have three characters. So "Mountain Time" is
"MST" and "MDT" for "Mountain Standard Time" and "Mountain Daylight
savings Time".

I had thought that there was a "Seychelles time", but perhaps someone
who has been there can speak up?

- Thomas

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Jeroen van Vianen 2000-09-21 15:41:10 Bug in RI
Previous Message Gunnar R|nning 2000-09-21 14:48:36 Re: Timezone discrepancies