Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Japan Bug Businesses Communications Programming Software The Internet Technology

Big Tech Warns of 'Japan's Millennium Bug' Ahead of Akihito's Abdication (theguardian.com) 211

MightyMartian shares a report from The Guardian: On April 30, 2019, Emperor Akihito of Japan is expected to abdicate the chrysanthemum throne. The decision was announced in December 2017 so as to ensure an orderly transition to Akihito's son, Naruhito, but the coronation could cause concerns in an unlikely place: the technology sector. The Japanese calendar counts up from the coronation of a new emperor, using not the name of the emperor, but the name of the era they herald. Akihito's coronation in January 1989 marked the beginning of the Heisei era, and the end of the Shwa era that preceded him; and Naruhito's coronation will itself mark another new era. But that brings problems. For one, Akihito has been on the throne for almost the entirety of the information age, meaning that many systems have never had to deal with a switchover in era. For another, the official name of Naruhito's era has yet to be announced, causing concern for diary publishers, calendar printers and international standards bodies. It's why some are calling it "Japan's Y2K problem." "The magnitude of this event on computing systems using the Japanese Calendar may be similar to the Y2K event with the Gregorian Calendar," said Microsoft's Shawn Steele. "For the Y2K event, there was world-wide recognition of the upcoming change, resulting in governments and software vendors beginning to work on solutions for that problem several years before January 1, 2000. Even with that preparation many organizations encountered problems due to the millennial transition. Fortunately, this is a rare event, however it means that most software has not been tested to ensure that it will behave with an additional era."

Unicode's Ken Whistler wrote in a message earlier this month: "The [Unicode Technical Committee] cannot afford to make any mistakes here, nor can it just *guess* and release the code point early. All of this is pointing directly to the necessity of issuing a Unicode 12.1 release sharply on the heels of Unicode 12.0, incorporating the addition of the new Japanese era name character, which all vendors will be under great pressure to immediately support in 2019 software releases."
This discussion has been archived. No new comments can be posted.

Big Tech Warns of 'Japan's Millennium Bug' Ahead of Akihito's Abdication

Comments Filter:
  • Oh damn! (Score:5, Funny)

    by RobinH ( 124750 ) on Thursday July 26, 2018 @08:09AM (#57012188) Homepage
    When Jesus comes back do we need to reset the year back to zero? Crap, as a programmer I hate our calendar system(s)!
    • Re:Oh damn! (Score:5, Funny)

      by ShanghaiBill ( 739463 ) on Thursday July 26, 2018 @08:14AM (#57012212)

      Crap, as a programmer I hate our calendar system(s)!

      Me too. It would be so much simpler if the earth rotated the sun in exactly 256 days, divided into exactly eight 32 day months.

      • Re:Oh damn! (Score:4, Funny)

        by Opportunist ( 166417 ) on Thursday July 26, 2018 @08:20AM (#57012252)

        Right, what computer illiterate idiot designed this system?

      • Re:Oh damn! (Score:5, Informative)

        by RobinH ( 124750 ) on Thursday July 26, 2018 @08:23AM (#57012270) Homepage
        Well, I've always thought the International Fixed Calendar [wikipedia.org] was a decent attempt at sanity, but if there's people in the world that can't adopt the metric system, there's no way in hell the calendar could change.
        • Well, I've always thought the International Fixed Calendar [wikipedia.org] was a decent attempt at sanity, but if there's people in the world that can't adopt the metric system, there's no way in hell the calendar could change.

          Maybe the best solution is to use a sane calendar system similar to that one as a base system (similar to Universal Coordinated Time [wikipedia.org]) and then just calculate offsets into whatever crazy calendar system some group prefers to use.

          Our current calendar system is pretty much bonkers anyway. We seem bizarrely attached to concepts like a 7 day week which is familiar but totally arbitrary. You could have a year with 73 weeks of 5 days each or a year with 5 months of 73 days and it would be equally valid and equal

          • by Diss Champ ( 934796 ) on Thursday July 26, 2018 @09:15AM (#57012538)

            Regarding a non-7 day week, this has been tried before and didn't work. See for example www.mdpi.com/2075-4698/3/4/399/pdf

            Sometimes a long-followed social construct has survived because it works well for us animals, and if that makes it harder for programming computers that is a reasonable cost.

            • Regarding a non-7 day week, this has been tried before and didn't work.

              Just because a small group tried something a long time ago doesn't have a lot of relevance to the discussion. Maybe it just wasn't the right system or the right time to do it. There is nothing magical about a 7 day week. It's fine but we could easily have a 6 or an 8 day week and there is no objective reason that couldn't work just as well. It's just something we continue to do because we've done it that way for a long time. The major benefit of it is that most of the world has standardized on it which

              • The Soviet Union is a rather large testing ground. They tried variations of 5 and 6 that worked poorly in an attempt to avoid using 7, which they viewed as a number that made it too easy for people to hold on to religion. While there are certainly an infinite number of variations one could try, its not like France (in the reference I cited) or the Soviets were too small a sample size to pretend that we have no evidence of relevance that 7 isn't just an arbitrary number and any other would work fine.

                • The Soviet Union is a rather large testing ground. They tried variations of 5 and 6 that worked poorly in an attempt to avoid using 7, which they viewed as a number that made it too easy for people to hold on to religion.

                  The Soviet Union tried a lot of things that didn't work very well. This is pretty far down the list among them. Plus like most standards it's the network effects of the existing standard that makes them hard to change once they are well established.

                  While there are certainly an infinite number of variations one could try, its not like France (in the reference I cited) or the Soviets were too small a sample size to pretend that we have no evidence of relevance that 7 isn't just an arbitrary number and any other would work fine.

                  Those are interesting case studies but one would have to dig a lot deeper into the reasons for those failures to determine if the length of the chosen week was the proximal cause of the failure. My guess is that it failed for other reasons similar to how the m

              • There is nothing magical about a 7 day week.

                Magical, no. But I reckon fatigue has something to do with it.

                I know some people do, but I wouldn't fancy working 9 days consecutively and getting one off. Not on an ongoing basis, anyway.

          • After the French Revolution, the Paris Commune created a revolutionary calendar that was certainly more internally coherent and logical, but it did not last long. Whether or not 7 days a week is arbitrary or not, the fact is that it has very deep roots. Much of our Western timekeeping dates back at least to the Sumerians, so it's probably one of the most long-lasting conventions ever made.

          • by djinn6 ( 1868030 )
            Unfortunately, UTC is also crazy in that it has irregular leap seconds. Then some idiot decided Unix time should be based on UTC rather than the other way around. So now you need to use TAI to bypass that problem, but software support is limited due to Unix time being "good enough" for most cases.
          • by RobinH ( 124750 )
            Arguably that's the only correct way to do it in a computer system, and any system that has to work in multiple time zones at once must actually be working this way (i.e. use UTC internally and convert to a display representation every time). However, UTC doesn't have any of the benefits of something like the International Fixed Calendar (where every nth of every month is always the same day of the week, there are always exactly 28 days in a month, and an economic quarter is always exactly 13 weeks long (a
          • What exactly is the point in having a calendar that's not synchronized to the solstices? Clocks measure time elapsed since an arbitrary epoch - calendars subdivide a year so that you can schedule seasonal activities. The two serve very different functions, and overlap only insofar as both measure the elapse of time.

            An accurate clock can't even measure the passage of days correctly, since every day is, on average, slightly longer than the one before it thanks to a combination of solar and lunar tidal drag,

            • I suppose I should say a *consistent* clock - one which holds the length of a second as constant over time. A sun dial accurately subdivides the day into sub-periods, but at the cost of constantly varying the length of those periods.

            • What exactly is the point in having a calendar that's not synchronized to the solstices?

              There is no objectively important reason that a calendar that tracks rotations of the Earth has to have any relationship whatsoever to orbits around the Sun. There is some utility in doing so but it's not as if we would be unaware of and unable to plan for the fact that the Earth takes approximately 365.25 days to orbit the Sun. Heck for quite a long time we used the Julian calendar [wikipedia.org] and it worked pretty well overall despite not quite so accurately reconciling the length of a year to the length of a day.

              • Except calendar aren't designed to track rotations of the Earth, they're designed to subdivide orbits so that people could effectively prepare for the changing of seasons well in advance, and served that purpose 10's of thousands of years before the Julian calendar was established. The inaccuracy of Julian calendar was itself due to insufficient correction, (they had 1 leap year every four) rather than not recognizing the fundamental problem. And considering that it took 100 years to drift (a bit less tha

                • Sorry - that bit near the top should have have been "took 100 years to drift a bit less than one day"

          • No, it's not. You just don't have enough astronomy. We invented measurements of time to know when to plant our crops and later to navigate the seas. It's entirely based on our observations of orbits and rotations. The math is all based on angles of circles and spheres. Our ancestors got really good at eyeballing the sky and figuring out where they were. It's probably in our DNA to some extent, since the ones that were bad at doing that tended not to make it home. If you want additional context, look at the
          • There is nothing arbitrary in our calendar system.
            Chrismas is 3 or 4 days after Jule, the winter solistice. (Because the Christians wanted to steal that germanic/nordic festival)
            A month is 28 days long, following the moon, hence its name.
            Our calendar is arranged around the 4 corner stones: the two equinoxes and the two solistices.
            I spare me to call you an idiot or a moron, blame your school system that you seem not to know the most simplest things in the universe.

        • Pretty bold calling a calendar system written by a British person the "International Fixed Calendar."

      • Re:Oh damn! (Score:5, Interesting)

        by AmiMoJo ( 196126 ) on Thursday July 26, 2018 @08:46AM (#57012400) Homepage Journal

        Interestingly in this case it probably won't be the numerical side of things that causes problems. Most IT systems use the same Gregorian calendar we do, and have a conversion function to handle translation to/from imperial eras.

        The problem is that, like leap seconds, there is no way to predict when eras will change so you have to update all your software every time there is a new one. Particularly for systems that handle personal data it's still common for people to enter their birthday using the imperial era system, so when the new era starts all those systems need to be able to handle it.

        It was bad enough the last time it happened, but this time IT is much more pervasive and user facing. All sorts of industries are affected, e.g. airlines need to be able to handle children born in the new era from day one.

      • by necro81 ( 917438 )

        It would be so much simpler if the earth rotated the sun in exactly 256 days, divided into exactly eight 32 day months.

        Ya know, with enough planets and moons of the right masses and orbits, you probably could generate a system that was locked into that kind of orbital resonance [wikipedia.org]. Maintaining stability in such a system, where small perturbations accumulate, would probably be difficult to guarantee, though.

        • by AmiMoJo ( 196126 )

          I wonder what the Outsiders charged them for those planetary drives?

          Whatever it was it would be worth it to make years 256 days long.

      • by PPH ( 736903 )
        The only correct solution: Stardate.
      • Comment removed based on user account deletion
    • Re:Oh damn! (Score:4, Insightful)

      by xxxJonBoyxxx ( 565205 ) on Thursday July 26, 2018 @09:11AM (#57012512)
      >> When Jesus comes back do we need to reset the year back to zero?

      Serious answer: No, because the current calendar is based on the BIRTH of Christ.

      https://en.wikipedia.org/wiki/Anno_Domini

      If Jesus comes back "in glory" it's unlikely that he'll be back as an infant (unless your a Stewie fan), so your timekeeping based on the birth of a highly religious infant should remain intact. Whether or not He will need clocks in the eternal world to come remains up for debate.

      https://en.wikipedia.org/wiki/Second_Coming
      • I think you missed the memo... the new dating scheme is BCE and CE https://en.wikipedia.org/wiki/... [wikipedia.org]

        BC and AD have gone out the window...

        • BC and AD have gone out the window...

          Not quite... here's another Wikipedia link: https://en.wikipedia.org/wiki/... [wikipedia.org]

          BC and AD are the traditional ways of designating eras. BCE and CE are common in some scholarly texts and in certain topic areas. Either convention may be appropriate for use in Wikipedia articles.

        • I think you missed the memo... the new dating scheme is BCE and CE https://en.wikipedia.org/wiki/... [wikipedia.org]

          BC and AD have gone out the window...

          It isn't a "new dating scheme", the year numbers are exactly the same. The only change is the labels, because, as you may be surprised to find out, non-Christians would rather not to refer to the current date as being in "the year of our lord".

        • I guess you got that reversed.
          No one - except some idiots- is using BCE and CE anymore.

      • Whether or not He will need clocks in the eternal world to come remains up for debate.

        How do you know it's eternal unless you're watching a clock? Maybe it's only a month, and then He'll get tired and want to do something else. (ADD Jesus.)

        OTOH, do they have night in Heaven? What if I'm an astronomer and want to see the stars? Oh, right; either I can just zoom over to see them, or I'll have forgotten all about them and be basking in the Glory of the One True Star. So I guess He's like a black hole with all of the other souls spinning and dancing around him.

        (Yeah, I'm a non-believer a

    • That is a classic out by 1 programming error. There is no year zero, the first year is year one.

    • "When Jesus comes back do we need to reset the year back to zero? "

      I suspect that if/when Jesus comes back (Why would he/she/it come back to Earth? There HAVE to be more interesting places populated by less obtuse/obnoxious entities to spend time with), Calender reset (Why would we do that in any case?) will be somewhere around item 43612 on our list of problems.

  • n/t
  • by TFlan91 ( 2615727 ) on Thursday July 26, 2018 @08:10AM (#57012196)

    "the official name of Naruhito's era has yet to be announced"

    Then... just announce it?

    • by Opportunist ( 166417 ) on Thursday July 26, 2018 @08:21AM (#57012258)

      I'm pretty sure it will be known in IT circles as the "placeholder era", and decades from now we'll find something along these lines in a lot of code...

      • by djinn6 ( 1868030 )
        Nah, just call it Heisei 31 until they figure out what the name of the new era is called. Or better yet, just use 2019. There's not a lot of reasons to give it a name other than improving the imperial family's prestige.

        Speaking of which, why are they still around? Isn't it about time to abolish this nonsense?
        • It is called Heisei 31 until May 29, then it's Unknown Era 1 May 30 (Or whatever the changeover date is)

    • Re:So... (Score:5, Informative)

      by SinGunner ( 911891 ) on Thursday July 26, 2018 @11:59AM (#57013706)
      Japanese "eras" have only been tied directly to the emperor since the Meiji Restoration (mid-1800's). The current era is only the 4th since this change. The Emperor takes a new name when they ascend the throne, so there has never been a way to know the name of the era in advance. Given the intense superstition prevalent in Japanese society, it seems incredibly unlikely a name would ever be disseminated in advance.
    • "the official name of Naruhito's era has yet to be announced"

      Then... just announce it?

      Why is this a problem? In most computing systems, is the era encoded as a string or as an enumerated value? If it's enumerated, the corresponding human-readable string can be changed later.

  • Fake news! (Score:4, Interesting)

    by Jesus H Rolle ( 4603733 ) on Thursday July 26, 2018 @08:23AM (#57012268)
    Japanese regnal years are not used for any significant calculations. Behind the scenes it's YYYY, with regnal years used only for display. This is an aesthetic issue only, and hardly unforeseen.
    • Re:Fake news! (Score:5, Insightful)

      by nojayuk ( 567177 ) on Thursday July 26, 2018 @08:54AM (#57012438)

      Heisei calendaring appears on official documents, train tickets etc. (quick check, a Japan Rail reserved-seat ticket I was issued back on the 12th of May this year has the date of issue as 30.-5.12, that is Heisei year 30, month 5, day 12). Nearly all common date expressions in Japan use Christian Era numbering though.

      • Would it kill anyone were their 2020 ticket to say 32-5-12? There's no ambiguity, even though it's technically wrong. Hey, it's Showa 93 right now! But seriously, a train machine printing an unambiguous but technically wrong date on a ticket is hardly a problem.
        • But seriously, a train machine printing an unambiguous but technically wrong date on a ticket is hardly a problem.

          Oh it's a huge problem!
          Lucky for you, I happen to be in the train ticket-machine business.
          Unfortunately the old machines cannot be updated, but I'll give you a..... mmm....... 4% discount on new ones.
          Unprecedented!
          Are YOU Going to be the one shameful business which is printing years in an antiquated era? If so, shame on your family's next four generations.

        • You are unfamiliar with trains in Japan then. It's an issue when they leave 20 [nytimes.com] or 25 seconds [soranews24.com] early. Lateness is inexcusable. If there is a change to the tickets that causes any confusion, it's going to be a madhouse.

          In most other countries, no big deal.

      • Just checked some packaged food, and the "use by"/"best before"/"packed on" dates are in Heisei Era on all of them. Some have both Heisie and CE dates, but none have CE only.

    • Not just display. Input is an issue also. If people are inputting new regnal years and the system interprets them as old regnal years that is going to lead to some substantial headaches.
  • And I'm a throwback apparently for using BC and AD ...
  • by Zontar_Thing_From_Ve ( 949321 ) on Thursday July 26, 2018 @08:48AM (#57012416)
    Officially years in Japan start with 1 and the coronation of the new emperor. Now it seems that the year in Japan is Heisei 30, which means it's the 30th year since the Heisei emperor (Akihito) took the throne. This would be something like in the US calling 2016 as Obama 8 or 2018 by the term Trump 2. The real problem here isn't that computers are going to shut down in Japan when the current emperor abdicates next year as planned. This isn't really a Y2K problem. The Guardian usually does good work but to say this is similar to Y2K is just not correct. The article even admits that some older computers have actually never updated the year from the Showa era (when Hiirohito was emperor) so they think this year is Showa 93. Those computers will have a problem in 2025 as their calendars were never designed to hold 3 digit years, which would make 2025 be Showa 100. The real problem with the abdication is that the next era for the upcoming emperor has not yet been named. OK, so why is that a problem? Well, Japan has a history of creating a brand new character for the era when it change and Unicode has a major release scheduled for right before the abdication is scheduled to happen. The brand new character is the problem because the next release of Unicode won't support it because nobody knows what it will be yet. They have the ability to guess, but nobody wants to guess because they could be wrong. So all this hubbub is that next year's major Unicode release will require a patch shortly after release with the patch including the new character for the new era. Do keep in mind that Akihito could die of natural causes before the abdication and this problem will happen immediately upon his death. And this problem will happen every time a new emperor takes over. I'm not convinced that this is really a major problem. Computers could easily just show Heisei 31 and so on until the Unicode fix is in place. I guess it's just fun on a slow news day to blow things out of proportion.
    • by omnichad ( 1198475 ) on Thursday July 26, 2018 @09:44AM (#57012714) Homepage

      If Unicode was at least sane enough, there would be a range of consecutive reserved codepoints. However, 0x337A is already assigned to something else (as is 0x337F if they started going in the other direction). Having a character doesn't matter as much as having a reserved codepoint. You can test without a real character.

  • by Joe_Dragon ( 2206452 ) on Thursday July 26, 2018 @08:51AM (#57012432)

    will they also fix the 2038 bug as well?
    or hold off so they have jobs in 2037 fixing it.

  • "Akihito has been on the throne for almost the entirety of the information age"

    Umm, no. We were well into the information age by 1989. The use of computers in vital infrastructure (e.g. banking) dates back to the late 50's. I checked out the author of the article, and as expected, he appears to be under 30; this would explain his myopic view of history.

  • by K. S. Van Horn ( 1355653 ) on Thursday July 26, 2018 @09:45AM (#57012718) Homepage

    "...Unicode, the international standards organisation which most famously controls the introduction of new emojis to the world."

    This is a new level of cluelessness. "Most famously"? Like, internationalization and localization were just afterthoughts; it's the emojis that they really focus on.

    How is this guy a technology reporter for a major newspaper?

  • by rxmd ( 205533 ) on Thursday July 26, 2018 @09:59AM (#57012778) Homepage

    Akihito's coronation in January 1989 marked the beginning of the Heisei era, and the end of the Shwa era that preceded him

    Actually it's not the Shwa era, but the Showa era, with a bar on top of the o. The character in question (U+014D) is used in transliterating Japanese in Latin script to indicate pronunciation. It has been part of Unicode since 1991.

    It's interesting to see in the summary a discussion of Unicode 12.1 vs. 12.0, when Slashdot itself doesn't support the Unicode 1.0 characters necessary to write the summary :)

"Never face facts; if you do, you'll never get up in the morning." -- Marlo Thomas

Working...