April 26th, 2006

Are You Ready for the 2038 Bug?

By Alice Hill
RealTechNews

Ok, so this one doesn’t have the same note of panic when you consider it doesn’t happen for 32 years, but it doesn’t hurt to plan ahead. The culprit this time: the way Unix and some variations of Linux and Windows 2000 register time.

“Instead of relying on an in built calendar they count seconds from the notional date the system was conceived, at GMT 00:00:00, on Thursday, January 1st, 1970, and like a car’s odometer going round the clock, on bug day it will run out of digits and the counter will roll over and probably reset to January 1st 1901 or another equally invalid date. Source: Propellarhead

We Say: Hmm…sounds familiar. At least there won’t be the hysteria of elevators plunging to the ground at midnight and ATMs shooting cash out of their slots, and all the worst case scenarios we heard about in 1999. Anyone working on it?

Share and Enjoy:These icons link to social bookmarking sites where readers can share and discover new web pages.
  • del.icio.us
  • digg
  • Fark
  • NewsVine
  • Reddit
  • YahooMyWeb
You can leave a comment, or trackback from your own site. RSS 2.0

27 comments to "Are You Ready for the 2038 Bug?"

  1. Alan Fargusson says:

    Various people are working on this. It is really a 32-bit problem. The solution is to change the time to a 64-bit value. Here is a little C program that will show if your system has a problem:

    /*
    * Test the system time functions for the 2038 problem.
    */

    #include
    #include

    int main( int argc, char **argv )
    {
    time_t t;
    char *p;

    fprintf( stdout, “sizeof time_t = %d\n”, sizeof (time_t) );

    for ( t = 2147483641U; t

    April 26th, 2006 at 1:17 pm

  2. Alan Fargusson says:

    Looks like the C program got mangled. Oh-well.

    April 26th, 2006 at 1:19 pm

  3. robbage says:

    Fair enough, because we are all still going to be using Win 2000 and Fedora Core 5 in 38 years

    April 26th, 2006 at 2:25 pm

  4. Rob says:

    Heh … I’d be disappointed if we weren’t all on 64-bit procs by the end of the decade, nevermind by 2038.

    April 26th, 2006 at 2:30 pm

  5. Kevin says:

    By then we may be using Windows Vista. All our problems will be solved.

    April 26th, 2006 at 2:47 pm

  6. Tom says:

    What about the 2099 bug in Windows!? I just verified this by opening the calendar and the highest year it can be set to is 2099!!! I’m so close to panic!!!! Does anyone have a C program that will be able to test if the rest of my Windows PCs are affected???

    April 26th, 2006 at 2:51 pm

  7. Alice says:

    Start buying water and duct tape immediately!

    April 26th, 2006 at 4:35 pm

  8. ed3 says:

    How about the year 2007 bug?? Ok, not really a bug.

    In 2007, the days on which Daylight Saving Time starts and ends changes for us USians.

    And then to make things interesting, if a study concludes the change did not affect energy usage, congress has the option to revert DST start/end dates back to the old ones.

    …and all that said, it shouldn’t be a problem for computer systems. In *NIX systems it’s configuration file change. In *indows, it’s a registry entry.

    April 26th, 2006 at 5:45 pm

  9. Yan says:

    If people still run win 2000 in 2038, I’m going to transform myself into a feathered pig. :)

    April 26th, 2006 at 6:40 pm

  10. David Bird says:

    As a 23 year old mainframe COBOL programmer back in 1974 I remember someone saying that we should use 4 digit years in date fields. The response was “who in their right minds will be using our programs as written by us 26 years from now”. 26 years later a bank which shall remain nameles certainly was. The moral of the story is that 2038 is 32 years away but someone may still be using old software on the premise that “if it ain’t broke don’t fix it” when in fact it is broken but not in the way they think.

    April 26th, 2006 at 11:10 pm

  11. William H. Bowen says:

    I can see it now - fade to 2038, when the IT Managers of the day invade all the old folks homes looking for us old geezers that set up these systems in the far off past (like 2006). Just like the movie “Space Cowboys” (I’m holding out for the Clint Eastwood role).

    Guess I should store away copies of my old Solaris and Red Hat CDs for that time in the future when I’ll need them again :) )

    April 26th, 2006 at 11:31 pm

  12. C Hill says:

    Interesting comments - many highlight the reason why problems occur. By 2038, we may all be think “32 bits - hasnt everybody always been using 256/512/1024 bit processors for x years”. It wont be the processors or the operating systems that are the problem, it will be the data & the way it is stored that is the problem. In 32 years time there will be plenty of reasons to be processing data created today.

    April 27th, 2006 at 12:44 am

  13. John Corliss says:

    Who cares? There won’t be any life on Earth in 2038.

    April 27th, 2006 at 2:11 am

  14. John says:

    “This one will only affect computers based on the Unix operating system, which includes some versions of Linux, and Windows 2000″

    Who ever posted this needs to be more specific. Which variations of Unix, and Linux are included in this.

    April 27th, 2006 at 5:58 am

  15. Robert Claypool says:

    In Indiana, the government decided this year to go onto DST. The solution of course is to switch the computer from Indiana (East) to Eastern Time (US & Canada).

    April 27th, 2006 at 6:24 am

  16. Alex says:

    The 2038 bug is my sunset retirement job. Already got it all planned out. I’m thinking of stockpiling old compilers and books.

    April 27th, 2006 at 8:58 am

  17. Gord says:

    Please! Not another F&c*ing Y2K-like scenario. This is BS. And it’s so old… :(

    April 27th, 2006 at 9:53 am

  18. Lex says:

    Not to worry!! Intrepid time traveller John Titor has already saved the day!!! (Google or Wiki him… he he) He got an IBM 5100 from 1975, stopped off during 2000 to watch the Y2K thing (and chat on-line) and even now I’m sure he’s delivered the package safely home to 2036 where they are using it’s undocumented in-house features to de-bug legacy software! So let’s not worry about a thing, Hurray!!

    April 27th, 2006 at 10:16 am

  19. ValuedCustomer says:

    pfft, I’ll let my kids worry about it..

    April 27th, 2006 at 12:07 pm

  20. Jimmy says:

    Good idea alex, we can start the histeria at the beginning of the year and make all kinds of money upgrading systems to whatever will break the next year, muhahaha.

    May 2nd, 2006 at 6:04 am

  21. play slots online says:

    play slots for free for fun online …

    Here is Join come fold earn play free cleopatra slots online horses. …

    November 5th, 2006 at 10:16 pm

  22. poker games says:

    free online poker room …

    Once glimmer standoff management poker tournaments exotic? …

    November 6th, 2006 at 1:50 am

  23. free craps online says:

    play craps …

    Through We handle represent free craps online soft carpet. …

    November 7th, 2006 at 1:02 am

  24. world series of poker 2006 says:

    world series of poker circuit events …

    Ironically carpet cancellation player world series of poker bust! …

    November 7th, 2006 at 1:48 am

  25. play video poker says:

    play video poker …

    Thanks pushka meat play free video poker flush edge ring? …

    November 7th, 2006 at 3:09 am

  26. play video poker says:

    play video poker …

    Thanks pushka meat play free video poker flush edge ring? …

    November 7th, 2006 at 3:09 am

  27. Andy says:

    well of cause the 2038 ‘bug’ should have been fixed as part of the y2k bug squashing effort. For crying out loud guys it wasn’t hard to check both at the same time!
    With the y2k bug we got unlimited budgets to fix the problem once and for all.
    I went through ALL of our companies propriaty code and all of the Open source code we used and made sure that there was a 64 bit storage of dates.
    The real headach was in ‘optomised’ data stores where the increasing of timet caused fun and games by taking my data record size over a memory boundery (much of my work is with embedded systems and having records in the same page or block of memory can be a real performance boost).

    Indeed the only areas we have real concern is with file systems, optomised to keep tables inside a single 4k block. breaking this boundery can lead to twice as much IO requirement for every disk access - a huge hit in performance on IO dependant systems.

    March 11th, 2007 at 12:43 am

Leave a comment