A likely explanation for the iOS Do Not Disturb bug

Jacqui Cheng for Ars Technica:

Apple has declined to publicly comment on the issue and won’t explicitly confirm the reason for the bug. But numerous developers experienced in the complicated world of calendaring have stepped up with their observations and the overall consensus has converged on the ISO week date.



  • Domicinator

    I so miss the DND functionality. Honestly, I can’t wait for the 8th. It’s really nice to have my phone automatically not bug me at night.

  • Techpm

    Dates are hard, yesterday Google was also caught with another little date parsing error – http://www.seroundtable.com/google-search-by-date-international-16165.html

    • http://twitter.com/billyrazzle Billy Razzle

      WHAT DO YOU MEAN GOOGLE HAD AN ERROR!?!?!!? DON’T YOU KNOW GOOGLE IS PERFECT!?!?!?!?!?!?!? THEIR MAPS ARE NEVER WRONG!!!!

  • JimH

    The article went into very good detail. And it was nice to see it point out a simple fact about complex software. While efforts are made to try to think of corner cases in advance for both the code and the test cases, test suites always have holes in them due to the sheer volume of corner cases that can occur. No one can possibly get 100% coverage on something as big as an OS, let alone many smaller programs. But once something like this is found it is usually added to the test suite improving coverage over time. It was also good to see the article point out that pushing out a code fix within 1 week of a bug surfacing is pretty much impossible. Any code change has to undergo additional testing to ensure that it doesn’t break anything else in the build. And software build are dynamic living things in any software that is getting ongoing fixes and updates. As for Apples ‘cool’ response, I was actually impressed that they responded as quickly as they did and let us know when the problem would disappear.

  • pchmiel

    Just remember that for instance in my corner of the world (we start weeks on Monday) this error appeared exactly one day later than in US for instance.

    So ISO date theory may not be completely true. But interesting nonetheless.

  • http://dirk.haage.info Dirk Haage

    Actually there is a major error in it: 2013 started in 2012, 31.12.2013 == 2012-w01-1

    The original article at TUAW was corrected yesterday.