December 2000 - tz - lists.iana.org

Re: NTP and POSIX Time in conflict?
by Alan Perry Dec. 9, 2000

Dec. 9, 2000

Did this discussion thread go anywhere? I had a "poorly timed" disk crash and lost a bunch of e-mail. What operating systems use local time as their basic clock these days? Just wondering. What was the issue with "NTP and POSIX Time in conflict?" alan >Some operating systems use local time as their basic clock. Novice >programmers often tell me that this is a good thing. I always find it >amusing to compare their arguments to the anti-leap-second arguments: > > Clocks are local time Clocks are non-leap-second counters > -------------------------------------------------------------------- > Computer clocks are set by Computer clocks are set by > humans, who use local time. NTP, which counts non-leap seconds. > > I don't know the time zone. I don't know the UTC-TAI difference. > > Who cares about occasional Who cares about occasional > errors in time subtraction? errors in time subtraction? > > Governments change time zones. IERS changes the leap-second table. > Keeping up to date is painful. Keeping up to date is painful. > > I just signed a contract that I just signed a contract that > specifies a future local time. specifies a future time in UTC. > How do I represent that time? How do I represent that time? > > An isolated system can't learn An isolated system can't learn > about time-zone changes. about new leap seconds. > >The unfortunate reality is that an isolated system can't maintain an >accurate local-time clock. How do we react to this? Do we screw up the >semantics of localtime(), saying that it's just fine for localtime() to >ignore changes in time zones and in the leap-second table? Or do we >maintain our standards for the semantics of localtime(), and acknowledge >that isolated systems can't support it properly? > >---Dan > > >

2 1

NTP and POSIX Time in conflict?
by Joe Gwinn Dec. 7, 2000

Dec. 7, 2000

First, let me introduce myself. I am the Chair of the Realtime part of the POSIX standards committee, the Vice Chair of all POSIX, and only slightly time-obsessed. I have been trying to straighten POSIX time issues out, but hasten to add that the larger POSIX community does not really care that much about accurate time, so it may take some persuading to make any changes. The current POSIX standard (IEEE Std 1003.1-1996) is very confusing and perhaps confused in its definition of time and time-related functions. Their intent appears to have been to dodge the whole issue of leap seconds (because isolated systems cannot know of leap seconds), to ensure that such things as file modification dates were causal (to the resolution of the clock), and to be simple to define and program. The fundamental clock of POSIX is "Seconds Since the Epoch", essentially a count (time_t) of SI Seconds elapsed since the Epoch (00:00:00 UTC 1 January 1970 AD), to which leap seconds are not applied. I call this clock "POSIX Time", and observe that it parallels TAI, albeit at a constant offset. "Broken-down time" in POSIX resembles UTC, but as leap seconds are not applied,is not in fact UTC. The fundamental problem with POSIX here is that the functions specified for conversion between POSIX Time and broken-down time fail if leap seconds are involved, in particular there are time values in one form that cannot be expressed in the other form. True UTC, as defined in ITU TR 460-4, does not share this problem, so posix is broken here. The above is background. Now some questions. 1. I have heard it claimed that the broken-ness of posix time prevented NTP from handling leap seconds correctly. Would you care to comment on this? Specifically, what specific brokenness causes what problem? 2. Arthur David Olson's library "tz" defines the POSIX Epoch as "1970-01-01 00:00:10 TAI", which is to say that POSIX Time differs from TAI by ten seconds. However, when I compute the difference using the USNO table of leapseconds, I get 8.000082 seconds, or very close to eight seconds, this being the number of leap seconds in UTC as of the POSIX Epoch. Why the difference? Ten seconds is the value when leap seconds were first applied, just before 1 January 1972, but is too large for 1 Jan 1970, and so it's tempting to simply define POSIX Time as offset from TAI by exactly eight whole seconds. Being off by 82 microseconds seems far less of a sin than being off by two seconds. Arthur David Olson recalls in an email that the intent was that if time_t is zero, then tm_min and tm_sec will also be zero, and the non-leapsecond and leapsecond scales will coincide at the Epoch, but diverge thereafter. One consequebce of this choice is the added constant two-second offset. Do you know of any added rationales? The issue here is which offset should POSIX choose, and why. Joe

4 4

RE: DST in west Amazonas and Fernando de Noronha
by Jesper Nørgaard Dec. 7, 2000

Dec. 7, 2000

Unfortunately, I couldn't load the pages http://pcdsh01.on.br/Fusbr.htm and Fusbrhv.htm today. I don't know if they have been permanently removed or renamed. Anyhow we should not make any changes to Fernando do Noronha just based on these pages, so if no other sources can be found that Fernando do Noronha was not on the same time as Pernambuco, it probably wasn't. Indeed it is also likely that I read them wrong if Paul Eggert doesn't read them that way. >The above-listed web page mentions Atalaia do Norte, Boca do Maoco, >Benjamin Constant, Eirunepé, Envira, and Ipixuna as being in west >Amazonas. I don't know offhand whether these are cities or districts. >I can't access the Brazilian Census Bureau pages at ><http://www.ibge.gov.br/ibge/default.php> (my Netscape 4.76 Solaris >browser rejects it with javascript errors) I had no problems loading these pages with Internet Explorer 5, but don't understand much portuguese. If "municípios" corresponds to the same thing as in spanish, these are municipality inhabitant counters, not strict city inhabitant counters. Indeed I am not completely sure if some names might only be a municipality and not also a city, but at least all the brazilian cities I know are mentioned so there is probably not much difference. I could confirm that population counters on the two mentioned pages were equivalent, and also that Eirunepé seems to be the most populous of the West Amazonas cities, so it seems to be the best choice for a new "America/Eirunepé" West Amazonas time zone. Thanks for the link to the brazilian population counters, they came in very useful for me. Regards, Jesper Nørgaard Email: jnorgard(a)Prodigy.Net.mx Programador/Analista CIMMYT - Centro Internacional de Mejoramiento de Maíz y Trigo Dirección: CIMMYT, Lisboa 27, 06600 Mexico, D.F. Tel.: +52 (5) 58-04-20-04 ext. 1374 Fax: +52 (5) 58-04-75-58 Tel. Casa: 53-10-05-95 ó 53-10-97-78 CIMMYT home page: http://cimmyt.cgiar.org Check out my Time Zone Converter: http://timezone50.homepage.com ---------- From: Paul Eggert[SMTP:eggert@twinsun.com] Sent: Lunes 4 de Diciembre de 2000 17.41 To: jnorgard(a)Prodigy.Net.mx Cc: tz(a)elsie.nci.nih.gov; worldclock(a)timeanddate.com Subject: Re: DST in west Amazonas and Fernando de Noronha > Date: Fri, 03 Nov 2000 19:17:14 -0600 > From: <jnorgard(a)Prodigy.Net.mx> > http://pcdsh01.on.br/Fusbr.htm > http://pcdsh01.on.br/Fusbrhv.htm Thanks for those pointers; I'll put them somewhere in my next proposed patches. > Since the TZ database for Brazil has been elaborated per state, we > have an obligation to document the exceptions from that. Yes, I think we should add a new entry, since as far as I can see, west Amazonas observed DST in 1993/1994, whereas Acre did not; so it's not the case that the two regions' clocks have been the same since 1970. > Therefore I suggest to put a second entry in the TZ database for > Amazonas as "America/Atalaia_do_Norte" or something like that. The above-listed web page mentions Atalaia do Norte, Boca do Maoco, Benjamin Constant, Eirunepé, Envira, and Ipixuna as being in west Amazonas. I don't know offhand whether these are cities or districts. I can't access the Brazilian Census Bureau pages at <http://www.ibge.gov.br/ibge/default.php> (my Netscape 4.76 Solaris browser rejects it with javascript errors), but <http://www.dcc.ufrj.br/~cristina/dadosam.html> gives the following cities in Amazonas, in decreasing order of population: MANAUS 1157357 PARINTINS 71574 MANACAPURU 65577 ITACOATIARA 64937 TEFE 62616 COARI 53327 MANICORE 37704 MAUES 36628 TABATINGA 32009 LABREA 27517 SAO GABRIEL DA CACHOEIRA 26992 IRANDUBA 26612 EIRUNEPE 25776 ... BENJAMIN CONSTANT 23633 ... ENVIRA 18192 ... IPIXUNA 11120 ... ATALAIA DO NORTE 8108 ... So it appears that America/Eirunepe would be a better choice than Atalaia do Norte. However, I don't see a population entry for Boca do Maoco, and I'm not sure what that is (perhaps a district?). > According to these official pages, Fernando do Noronha has been on > DST for some while back (e.g. before October 15 where the status > probably changed). I don't read the pages that way. http://pcdsh01.on.br/Fusbr.htm and http://pcdsh01.on.br/Fusbrhv.htm both list Fernando de Noronha as being 2 hours behind UTC. It is a little tricky, since the pages also list the offset from Brasilia, and this offset does differ between winter and summer.

1 0

Re: zoneinfo/right/
by Markus Kuhn Dec. 6, 2000

Dec. 6, 2000

Andreas Schwab wrote on 2000-12-06 22:15 UTC: > Chmouel Boudjnah <chmouel(a)mandrakesoft.com> writes: > |> Could someone point me who use /usr/share/zoneinfo/right/ ? and if > |> it's needed ? Nobody needs this. The "right/" time zones are an experimental mechanism to play around with an alternative non-POSIX time_t definition, in which the seconds-since-the-epoch quantity increases even during a leap second. This is in violation of ISO/IEC 9945-1:1996 section 2.2.2.113 and is clearly not recommended for general use, otherwise your time_t values will differ from the rest of the world by the TAI-UTC difference (currently 30 seconds). It has the advantage that time_t becomes a true second counter that does not behave funnily during leap seconds and difftime() becomes equivalent to subtraction. Apart from adventurous leap second geeks, nobody should ever use the "right/ " time zones, as they might cause your NFS file timestamps and quite a couple of other applications that depend on the POSIX definition of seconds-since-the-epoch (which does not count leap seconds) to be wrong by half a minute. They can safely be omitted from a Linux distribution if disk space is precious. > ftp://elsie.nci.nih.gov/pub/tzarchive.gz Markus -- Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK Email: mkuhn at acm.org, WWW: <http://www.cl.cam.ac.uk/~mgk25/>

1 0

various time changes
by Rives McDow Dec. 5, 2000

Dec. 5, 2000

Dear Paul, Here are some changes in time zones that have come to my attention recently. Tonga is observing DST as of November 4, 2000, and will stop on January 27, 2001. Dominican Republic changed it's mind and presidential decree on Tuesday, November 28, 2000, with a new decree. On Sunday, December 3 at 1:00 AM the Dominican Republic will be reverting to 8 hours from the International Date Line, and will not be using DST in the foreseeable future. The reason they decided to use DST was to be in synch with Puerto Rico, who was also going to implement DST. When Puerto Rico didn't implement DST, the president decided to revert. Jordan will most probably implement DST next March (2001), on the last Thursday. However, with them, it is not possible to predict this with any certainty. If they do implement it, it will run to the last Thursday of September, 2001. Rives McDow

2 1

Fiji and Tonga has stopped using DST
by Steffen Thorsen Dec. 5, 2000

Dec. 5, 2000

Risto Nykänen <ristonykanen(a)luukku.com> has reported to me that both Tonga and Fiji has stopped using DST. I have tried to research this a little bit further, and found an official statement about Fiji: from http://www.fiji.gov.fj/core/press/2000_09_26_4.html > The interim government has decided not to introduce "daylight saving" > this year. > > A spokesman for the Ministry of Labour said the decision follows > strong opposition from some commercial enterprises and the Ministry of > Education who were concerned about the disruptions caused to school > children, especially during the examination periods. > This means that as from Sunday 4th, November, there will be no change > as normal standard time will keep on operating. I have not found anything for Tonga, but it is actually no surprise if they have stopped using DST, as the "first into year 2000" hype has ended. Back in March I found a notice on the website http://www.tongaonline.com that Tonga changed back to standard time one month early, on March 19 instead of the original reported date April 16. Unfortunately, the article is no longer available on the site, and I did not make a copy of the text, and I have forgotten to report it here. (Original URL was: http://www.tongaonline.com/news/march162000.htm ) Regards, Steffen -- Steffen Thorsen <steffen(a)thorsen.priv.no> <steffent(a)pvv.ntnu.no> http://www.thorsen.priv.no

2 1

TZ and DST changes for East Timor and Dominican Republic
by Steffen Thorsen Dec. 5, 2000

Dec. 5, 2000

* East Timor seems to have changed time zone from GMT+8 to GMT+9, around the end of September. * The Dominican Republic is now using the same time zone as Eastern US with DST during summer (EST/EDT). detail information: East Timor: Charles C. Schlegel" <chucks(a)bkk3.loxinfo.co.th> reported this: > I have just returned from East Timor, and have discovered that the UN > Transitional Authority there has decided to change the time zone of > the country. I think it took effect at the end of September. E. > Timor time is now 2 hours ahead of Bangkok, not one hour as before. > > I don't know just who you could contact to get formal confirmation of > this. The political/administrative structure in E. Timor is still > very much in a state of evolution. The legal governing body at > present is = the UN Transitional Authority for East Timor (UNTAET). Dominican Republic: Enrique Morales <emorales(a)pnud.org.do> reported to me that the Dominican Republic has changed the time zone to Eastern Standard Time as of Sunday 29 at 2 am. Some sort of "official" statement is available in Spanish here: http://www.listin.com.do/antes/261000/republica/princi.html Use of http://babelfish.altavista.com to "translate" to English: > Power Executive modified last night decree previous which the next > Sunday 29 arranged to retard in one hour the clocks in all the country > to 12:00 of the midnight, that now will be the same day but to 2:00 of > the dawn, to avoid probable disadvantages in the services of > telecommunications. > > Decree 1011, emitted last night, arranges that > the delay is from the 2:00 of the dawn of the same Sunday 29, instead > of the 12:00 of the midnight of that same day, as it established the > previous one, the 820-00, of the 25 of last September. The new decree > arranges in addition that the Sunday first of April of next year 2001, > to 2:00 of the dawn, all the clocks are advanced in one hour. [continues ...] I have some other tz updates to report as well, I hope to have them ready tomorrow. Best regards, Steffen -- Steffen Thorsen <steffen(a)thorsen.priv.no> <steffent(a)pvv.ntnu.no> http://www.thorsen.priv.no

2 1

RE: Brazil again
by Jesper Nørgaard Dec. 4, 2000

Dec. 4, 2000

The recent changes to abandon DST in the states Pernambuco and Roraima as of October 15, and Alagoas, Sergipe, Paraíba, Rio Grande do Norte, Ceará, Piauí and Maranhão as of October 21, will surely make necessary to invent new time zone items in the TZ database. But I suggest two more changes that are not directly generated from these events. I checked out the pages mentioned by Steffen Thorsen http://pcdsh01.on.br/DecHV.html and found a couple of pages that should be of interest Brazil on Standard Time: http://pcdsh01.on.br/Fusbr.htm Brazil on Daylight Time: http://pcdsh01.on.br/Fusbrhv.htm Unfortunately these two pages have not been updated as per the last changes from 15.th. and 21.st October 2000, but they reflect well how the situation was supposed to be before those dates. NEW TIMEZONE IN AMAZONAS Since the TZ database for Brazil has been elaborated per state, we have an obligation to document the exceptions from that. In the case of Amazonas we only have one entry "America/Manaus" on GMT-4.0 without DST, but according to the pages there is a region of the Amazonas State with GMT-5.0 without DST (like Acre or the "America/Porto_Acre" entry). The city information I have about this follows here: GMT-4.0 without DST @ Brazil - Amazonas East, Manaus,-3.10,-60.00 Borba,-4.39,-59.59 Coarí,-4.11,-63.14 Fonte Boa,-2.53,-66.02 Humaitá,-7.52,-63.02 Lábrea,-7.27,-64.78 Moura,-4.12,-63.42 Bôca do Acre,-8.75,-67.38 Jutaí,-5.18,-68.90 Floriano Peixoto,-9.07,-67.42 GMT-5.0 without DST @ Brazil - Amazonas West, Atalaia do Norte,-4.33,-70.20 Benjamin Constant,-4.37,-70.03 Eirunepé,-6.67,-69.87 Envira,-7.30,-70.22 Ipixuna,-7.57,-72.60 The format is {@ Country/State} City, Latitude, Longitude Therefore I suggest to put a second entry in the TZ database for Amazonas as "America/Atalaia_do_Norte" or something like that. FERNANDO DO NORONHA ON DST According to these official pages, Fernando do Noronha has been on DST for some while back (e.g. before October 15 where the status probably changed). This is as a contrast to other islands like Trindade and Martin Vaz, which have not been on DST according to the pages. This change to the status of Fernando do Noronha could well be linked with the change of status of the islands from being "self ruling" to be under the rule of state Pernambuco. I don't have information when these events should have happened, but the issue is worthwhile investigating. The ironic thing is that at this point when I'm reporting that Fernando do Noronha has been on DST, it most probably is no longer on DST. This is because it now is under Pernambuco State rule, and since Pernambuco abandoned DST on October 15, probably the island did the same. By the way the TZ database is mentioning Fernando do Noronha as "Fernando de Noronha". I believe it should be spelled with "do" instead of "de". Regards, Jesper Nørgaard Email: jnorgard(a)Prodigy.Net.mx Programador/Analista CIMMYT - Centro Internacional de Mejoramiento de Maíz y Trigo Dirección: CIMMYT, Lisboa 27, 06600 Mexico, D.F. Tel.: +52 (5) 58-04-20-04 ext. 1374 Fax: +52 (5) 58-04-75-58 Tel. Casa: 53-10-05-95 ó 53-10-97-78 CIMMYT home page: http://cimmyt.cgiar.org Check out my Time Zone Converter: http://timezone50.homepage.com ---------- From: Steffen Thorsen[SMTP:straen@thorsen.priv.no] Sent: Martes 31 de Octubre de 2000 17.09 To: Time Zone List Subject: Brazil again I forgot one entry in my last mail: Manoel Andrade <mandrade(a)hotlink.com.br> reported to me that some states/territories in Brazil have stopped using DST, just about a week after the DST period started. Now looking at the official page http://pcdsh01.on.br/DecHV.html : It seems that Pernambuco and Roraima stopped using DST on 15 October at midnight. Then, just one week later, Sergipe, Alagoas, Paraiba, Rio Grande do Norte, Ceará, Piauí and Maranhão stopped using DST on 22 October at midnight. Regards, Steffen -- Steffen Thorsen <steffen(a)thorsen.priv.no> <steffent(a)pvv.ntnu.no> http://www.thorsen.priv.no

3 2

Re: NTP and POSIX Time in conflict?
by Bradley White Dec. 4, 2000

Dec. 4, 2000

> 1. I have heard it claimed that the broken-ness of posix time > prevented NTP from handling leap seconds correctly. Would you care > to comment on this? Specifically, what specific brokenness causes > what problem? POSIX legislates a particular mapping between time_t's and the time of day (e.g., 536457599 shall correspond to "Wed Dec 31 23:59:59 UTC 1986"). So, POSIX time must be adjusted over leap seconds. The premier NTP implementation goes to convoluted lengths to perform that adjustment. The "correct" way to handle leap seconds is a matter of opinion, but `ado-right' stalwarts would contend that using a continuous time base results in a better system, and treating time_t as opaque gets you there. See time2posix(3) in the ado source for more details. Bradley

1 0

RE: Asia/Dacca Spelling
by Olson, Arthur David (NCI) Dec. 4, 2000

Dec. 4, 2000

> The following was reported by Steve Minkin regarding the current time zone > Asia/Dacca. Please note that Steve is not on this distribution list. Should > the spelling for this time zone be changed? Should an alias be defined? > "Please note that about 10 years ago the spelling of the capital of > bangladesh changed from Dacca to Dhaka. No one uses Dacca any more." A search at www.un.org yields 30 hits on Dhaka and one on Dacca. The page http://www.xrefer.com/entry/501998, attributed to "The Macmillan Encyclopedia 2001, © Market House Books Ltd 2000," contains: "Dhaka (name until 1982: Dacca)." So...my sense is that the thing to change "Zone Asia/Dacca" to "Zone Asia/Dhaka" and add a "Link Asia/Dhaka Asia/Dacca" to the "backward" file. --ado

1 0