You'll say ``Argh!'' now, i know. After a long and heated discussion in the FreeBSD core team, we came to the conclusion that changing the timezone name for Central Europe from MET to CET is a very bad move. The reasoning: the name MET is in widespread use here on almost all Unix systems around. (With AIX being the most noteworthy and most funny exception; they call it ``Norway-France-Time'', NFT. :) However technically correct your choice for CET might be, it breaks tradition, and is in general something people here will consider a ``US-centric'' decision. MET might sound like a half-translated term in German, but people can easily memorize it, and identify their own timezone by this name. Other languages like Danish even use this term as well as their local abbreviation (which is no surprise given that Danish is closer to English than German is). You'll note that we don't mind you US inhabitants naming the New York timezone ``EST'' either, even though the term ``Western Atlantic Timezone'' is far more correct from our point of view, since you are west of us, not east as the name suggests. Needless to say, i'd leave the decision for this name at the people that have to live with the name. But i expect you doing likewise: respect our wish to keep the name `MET' for the Central/Middle European Time. Changing it will gain nothing, but cause support hassles for our userbase. Btw., there's at least one alternate (but not less bogus) translation for this abbreviation: Mediterranean Time. I'm also sending this request on behalf of Poul-Henning Kamp, another FreeBSD core team member. Cc to him for this. -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)
J Wunsch wrote:
You'll say ``Argh!'' now, i know.
After a long and heated discussion in the FreeBSD core team, we came to the conclusion that changing the timezone name for Central Europe from MET to CET is a very bad move.
Hi, This change resulted from an original mail of mine which asked why Unix used MET when the rest of the European, English-speaking world uses CET. Central European Time is the generally accepted English-language term for the time zone of France, Germany etc. I (as an English-speaking, - in fact, English. - European, informatician) applaud the use of CET and of CEST instead of CETDST. Any move back would reinstate a situation which is completely confusing to anyone who is used to the common name for this time zone. How long until people say that SkyTV can't talk about CET because it's not standard?! Cheers Pete -- Peter H.C. Hullah Technical Services mailto:Peter.Hullah@eurocontrol.fr EUROCONTROL Experimental Centre Phone: +33 1 69 88 75 49 BP 15, Rue des Bordes, Fax: +33 1 60 85 15 04 91222 BRETIGNY SUR ORGE CEDEX France
As Peter Hullah wrote:
This change resulted from an original mail of mine which asked why Unix used MET when the rest of the European, English-speaking world uses CET. Central European Time is the generally accepted English-language term for the time zone of France, Germany etc. I (as an
Well, but only very few English native-speakers actually live in MET. So why don't you value the opinion of the people living here more than the opinion of `foreigners'?
English-speaking, - in fact, English. - European, informatician) applaud the use of CET and of CEST instead of CETDST. Any move back would reinstate a situation which is completely confusing to anyone who is used to the common name for this time zone.
My strongest point is still that you're simply not mighty enough to change the world. Your timezone package is great, but you by far don't constitute a majority in Unix installations -- and it's Unix what we are talking about here. So see my first mail, while it's arguable whether or not your are technically or politically or linguistically more correct, you break the tradition people got used to. Breaking tradition is almost always a Very Bad Thing, in particular if it's for nothing more than cosmetical reasons. Your naming might have been very acceptable 5 or 10 years ago, when the Unix world in Middle (Central, if you prefer) Europe still evolved, but right now, with a userbase of about a million or perhaps more: sorry, you're too late. Btw., the German does also have a word for `central', but doesn't call the timezone ``Zentraleuropäische Zeit'' either. Neither do the Danes. Perhaps that's why ``MET'' is more popular here. (MET DST looks a bit strange at the first glance, but the term `daylight savings time' is a so much better description than the term `summer time' so people tend to accept this name well.) Being faced with the opportunity to use a locally changed name in FreeBSD, we rather decided to approach you to revert your opinion. If we changed our zone name in the shipped systems, you wouldn't gain anything by it (except your personal pride, perhaps), but we'd have more hassles in maintaining your contributed software in our repository. Finally: the opinions expressed here are not only my own. Even without a straw poll, i know that other people have also been surprised by this change -- and that's not even with a _shipped_ system, but just only with the development version by now. -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)
From: J Wunsch <j@uriah.heep.sax.de> Date: Tue, 26 Nov 1996 21:13:04 +0100 (MET) So why don't you value the opinion of the people living here more than the opinion of `foreigners'? I value the opinion of all contributors. In this case, the problem is that some people use `CET' and some use `MET', and the question is which abbreviation to prefer in the tz database. My strongest point is still that you're simply not mighty enough to change the world. The real-world tradition prefers `CET' -- at least, that is what I found when I looked into the issue, and I haven't seen much contrary evidence though I would welcome it. The question was whether to keep the tz database's idiosyncratic `MET/MET DST' abbreviations, or to switch to the real-world abbreviations. I'm sorry to hear that the change is upsetting some users. But the tz database continues to support TZ='MET' for people who prefer the old style. And one is always free to edit the zoneinfo files to output `DNT', `EMI', `EMT', `FWT', `HFH', `HOE', `ITA', `MET', `MEWT', `MEZ', `NFT', `NOR', `SET', `SST', `SWT', or `WUT', all of which I have seen listed as abbreviations for winter time for some locales in that time zone. I'm afraid that questions like these can never be completely resolved to everybody's satisfaction. Perhaps now's the time to start thinking about a system for localizing the time zone abbreviations. It would be nice if this could be done to the tables without breaking existing instances of `zic'. Then e.g. German-speakers could easily see `MEZ'/`MESZ', which are the preferred German abbreviations.
As Paul Eggert wrote:
I'm afraid that questions like these can never be completely resolved to everybody's satisfaction. Perhaps now's the time to start thinking about a system for localizing the time zone abbreviations. It would be nice if this could be done to the tables without breaking existing instances of `zic'. Then e.g. German-speakers could easily see `MEZ'/`MESZ', which are the preferred German abbreviations.
Good idea, but probably fairly complex, too. -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)
From: J Wunsch <j@uriah.heep.sax.de> Date: Tue, 26 Nov 1996 10:04:20 +0100 (MET) However technically correct your choice for CET might be, it breaks tradition, and is in general something people here will consider a ``US-centric'' decision. Actually, if anything, `MET' is the US-centric name, since it was chosen by an American, whereas `CET' was championed by a European. It wasn't an easy decision to change `MET' to `CET'. However, when I researched the subject I found that `CET' was by far the most popular name in real-world tradition; the `MET's that I found were usually derived from the tz database itself, and were thus suspect. I'll send you some of my results under separate cover (I already sent it to the tz mailing list). If you're having backward compatibility problems, you can set TZ='MET' and you should continue to see `MET'.
In message <199611261916.LAA15292@shade.twinsun.com>, Paul Eggert writes:
From: J Wunsch <j@uriah.heep.sax.de> Date: Tue, 26 Nov 1996 10:04:20 +0100 (MET)
However technically correct your choice for CET might be, it breaks tradition, and is in general something people here will consider a ``US-centric'' decision.
Actually, if anything, `MET' is the US-centric name, since it was chosen by an American, whereas `CET' was championed by a European.
Well I happen to disagree here. At least for Denmark MET is actually the abbreviation recommended by the ministry of the interior. -- Poul-Henning Kamp | phk@FreeBSD.ORG FreeBSD Core-team. http://www.freebsd.org/~phk | phk@login.dknet.dk Private mailbox. whois: [PHK] | phk@ref.tfs.com TRW Financial Systems, Inc. Future will arrive by its own means, progress not so.
In message <5562.849036754@critter.tfs.com>, Poul-Henning Kamp wrote:
At least for Denmark MET is actually the abbreviation recommended by the ministry of the interior.
Be careful here: Is MET the official abbreviation associated with the Danish or the English language? I expect it is only Danish! Abbreviations depend on the language! In Germany, there is a law that defines MEZ as the German abbreviation, nevertheless the official translation into English is CET (as used by the EU buerocrats when they mess around with the summer time rules every few years). For the Olson data base, it was a design decision that only abbreviations based on the English names of the time zone will be used (as found for example in international airline association manuals). Multiple language versions of the database is a possible future project for which nobody has found the time and energy to realize it so far. If you want to replace the English abbreviation CET by the Danish abbreviation MET, then I could argue as well that we should replace it instead by the German abbreviation MEZ as there are much more German speaking Unix users than Danish speaking users. You see to where this argument would lead? Markus -- Markus G. Kuhn, Computer Science grad student, Purdue University, Indiana, USA -- email: kuhn@cs.purdue.edu
Date: Tue, 26 Nov 1996 20:32:34 +0100 From: Poul-Henning Kamp <phk@critter.tfs.com> for Denmark MET is actually the abbreviation recommended by the ministry of the interior. That's odd, since the Danish Standards Association recommends `CET' (reference: Posix 1003.1-1996 [ISO/IEC 9945-1: 1996], page 614, line 105). I just looked in Dejanews for recent Usenet articles mentioning `Denmark CET MET' and talking about the issue, and I found only the following relevant articles, which seem to prefer `CET' to `MET': Henrik Buch Mortensen <hbm@emi.dtu.dk> Re: MET timezone: premature end of DST <URL:news:u3d8ynaylw.fsf@amalthea.emi.dtu.dk> (1996-10-13) This article presents a new zoneinfo file written from scratch by the Danish author, which uses CET and not MET. Thomas Petersen <petersen@post1.tele.dk> Linux Danish HOWTO (part 1/1) <URL:news:Danish-HOWTO.1-7533.847339934@cc.gatech.edu> (1996-11-06) This article talks about setting the time zone for Denmark, mentions CET before MET, and then gives instructions using the old zoneinfo MET file. Do you have a reference for the `MET' recommendation? Is this `MET' an abbreviation for Danish or for English?
In message <199611270100.RAA16930@shade.twinsun.com>, Paul Eggert writes:
Date: Tue, 26 Nov 1996 20:32:34 +0100 From: Poul-Henning Kamp <phk@critter.tfs.com>
for Denmark MET is actually the abbreviation recommended by the ministry of the interior.
That's odd, since the Danish Standards Association recommends `CET' (reference: Posix 1003.1-1996 [ISO/IEC 9945-1: 1996], page 614, line 105).
Ok, they may have changed it recently then. My reference is from 1991, I guess this could be another bright move from the eurocrats in EU :-(
Do you have a reference for the `MET' recommendation? Is this `MET' an abbreviation for Danish or for English?
Danish, it refers to "Mellemeuropaeisk tid" and states the "MET" as the abreviation. Interesting footnote here: They go to great length to explain how to write the 02:00 - 03:00 interwall in the fall transistion: During the first interval use the suffix 'A': 02:14A During the second interval use the suffix 'B': 02:07B Thus the timescale will be: ... 01:58 01:59 02:00A 02:01A ... 02:58A 02:59A 02:00B 02:01B ... 02:58B 02:59B 03:00 03:01 Rather computer-unfriendly really... -- Poul-Henning Kamp | phk@FreeBSD.ORG FreeBSD Core-team. http://www.freebsd.org/~phk | phk@login.dknet.dk Private mailbox. whois: [PHK] | phk@ref.tfs.com TRW Financial Systems, Inc. Future will arrive by its own means, progress not so.
As Paul Eggert wrote:
Actually, if anything, `MET' is the US-centric name, since it was chosen by an American, whereas `CET' was championed by a European.
:)
It wasn't an easy decision to change `MET' to `CET'. However, when I researched the subject I found that `CET' was by far the most popular name in real-world tradition; the `MET's that I found were usually derived from the tz database itself, and were thus suspect. I'll send you some of my results under separate cover (I already sent it to the tz mailing list).
I've already been forwarded this discussion before, and the reason is that i simply don't agree with it, thus i wrote to this list. Again, the Unix world has a long tradition to use the name `MET', and by now, most vendors (AIX being the only exception i know) use it here. This is nothing you (or we, for that matter) have the power to change. Creating even more confusion by breaking what has been usual for what i've called `cosmetic reasons' is something that will only cause us support nightmares, but gain absolutely nothing. You're probably right that `MET' is just a lazy translation of the German (and Danish and ...) timezone name into English, but that's apparently what people can memorize. If at all, you should at least call for some discussion and probably even vote in a group like de.comp.os.unix (and other country's counterpart groups, i don't know whether there's a fr.comp.os.unix, too) first. That's where the people can be found that will finally be most affected by your opinion and decision (and that's where the people are who will finally bug _me_ with their questions, since they know i'm a FreeBSD core member, and thus expect to get answers from me). -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)
In message <199611260904.KAA10440@uriah.heep.sax.de>, J Wunsch wrote:
You'll say ``Argh!'' now, i know.
After a long and heated discussion in the , we came to the conclusion that changing the timezone name for Central Europe from MET to CET is a very bad move.
Argh! :-) - According to my Langenscheidt Englisch-Deutsch Woerterbuch (a very popular German-English dictionary) MEZ is translated into English as CET and MESZ is translated into CEST. - Several other German-English dictionaries that I checked agree. - English geography textbooks call the area in which Germany and its neighbors are located "Central Europe" and almost never "Middle Europe". - The head of the PTB time department (German government timekeeping lab) confirmed to me that CET/CEST is the common usage. - CET clearly dominates MET by a factor >10 in the media. For example just have a look at European satellite TV schedules on videotext. Just do a Web search: When you see MET, then this is almost exclusively an output of the Unix "date" command, i.e. someone didn't care about the abbreviation at all when this was written. - I hardly ever have seen MET and METDST used anywhere outside the Unix world and we can very well assume that MET is today only known because it somehow found its way into the Olson database some time ago. DST is an abbreviation exclusively used in the U.S., most other countries say "summer time". Please forward this list of arguments to the other members of the FreeBSD core team to allow them to benefit from the research that others and I have done on this subject before the tz databse was changed and to allow interested people to base future "long and heated discussions" on some more real world evidence and not just on what Unix geeks are used to know about the real world.
The reasoning: the name MET is in widespread use here on almost all Unix systems around.
Just because the tz database is in widespread use on almost all Unix systems around. As you can read in any statistics textbook: Before you draw conclusions form correlations, you must eliminate common causes.
(With AIX being the most noteworthy and most funny exception; they call it ``Norway-France-Time'', NFT. :)
Oops ... well, what else would you ecpect from AIX ... ;-)
technically correct your choice for CET might be, it breaks tradition, and is in general something people here will consider a ``US-centric'' decision.
The main arguments for the MET->CET came from people in France and Germany. Calling the discussion for the MET->CET change "US-centric" certainly lacks any basis. Markus (German) -- Markus G. Kuhn, Computer Science grad student, Purdue University, Indiana, USA -- email: kuhn@cs.purdue.edu
As Markus G. Kuhn wrote:
- I hardly ever have seen MET and METDST used anywhere outside the Unix
But we _are_ in the Unix world. Our ``target market'' is not outside, but inside this world.
world and we can very well assume that MET is today only known because it somehow found its way into the Olson database some time ago. DST is an abbreviation exclusively used in the U.S., most other countries say "summer time".
There's much more Unices around than those using the Olsen database. Surprisingly (for you :), almost all of them agree on MET, provided they cover Europe at all. (There are probably more people disagreeing on the DST naming though, but remember that the old TZ variable scheme often only provided for three-letter names for the DST zone, too.) Even if the Olsen database were really the only source: _you are too late_. Correcting a mistake should have been done earlier, not after years of its use. The American Indians don't really live in India, but changing this name after years of usage wouldn't make sense, either.
Please forward this list of arguments to the other members of the FreeBSD core team...
No need. We are presenting you our opinions, and those of other people who we know have the same opinion.
others and I have done on this subject before the tz databse was changed...
Sorry, i can't remember you starting a discussion in de.comp.os.unix. That's where those affected by your decision can be found. Why didn't you ask there what people would prefer?
technically correct your choice for CET might be, it breaks tradition, and is in general something people here will consider a ``US-centric'' decision.
The main arguments for the MET->CET came from people in France and Germany.
So it must be interesting for you to know that the main arguments against it do also come from people in these countries now. :-) (I know of other examples where you've been gratuitously changing timezone names. The resulting ``MOST'' was entirely unintelligible for me, be assured. :) -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)
In message <199611262146.WAA14614@uriah.heep.sax.de>, J Wunsch wrote:
As Markus G. Kuhn wrote:
- I hardly ever have seen MET and METDST used anywhere outside the Unix
But we _are_ in the Unix world. Our ``target market'' is not outside, but inside this world.
The time zone string CET or MET is never parsed by any software, because in all applications (e-mail, news, etc.) the time format standards require that a numeric local time - UTC offset is also provided (e.g., required by RFC822 except for US names) and the numeric value is then used. Therefore, MET->CET is no compatibility problem. Consequently: Only humans have to recognize the time zone abbreviation and it definitely does not make sense here to use one that does not follow common practice (Langenscheid, PTB, SkyTV, CNN, etc.).
Sorry, i can't remember you starting a discussion in de.comp.os.unix. That's where those affected by your decision can be found. Why didn't you ask there what people would prefer?
May be you simply read the wrong groups. A large MET->CET discussion took place in de.comp.standards and CET was after a brief exchange of arguments quite quickly the common consens. Give me around 48 hours, then I have some time to start a small educational flame war in de.comp.os.unix. I think, I am quite good at this ... ;-) Let's see in two weeks again what the USENET mob thinks then.
The American Indians don't really live in India, but changing this name after years of usage wouldn't make sense, either.
Here you really have very litte idea of what you are talking about: They are called "Native Americans" today. Political correctness ideology requires the Americans to change terminology every ten years and believe me, they are quite good at this! Markus -- Markus G. Kuhn, Computer Science grad student, Purdue University, Indiana, USA -- email: kuhn@cs.purdue.edu
participants (5)
-
J Wunsch -
kuhn＠cs.purdue.edu -
Paul Eggert -
Peter Hullah -
Poul-Henning Kamp