Comfort for Doomsayers


[ HOURGLASS2 OUTPOST ] [ Follow Ups ] [ Post Followup ] [ FAQ ]

Posted by J.H. [JH] on October 28, 1999 at 07:33:11 {BrTq3yWfsY0MqZp34U2.HR5JhS9/hc}:

It's going to be a hard time next year for doomsayers, when the expected End of the World fails to materialize again. Actually, it's a hard time already. So long have they relied on Nostradamus' famous July 1999 prophecy, and now, after September, the last chance to play word-games to earn another month is gone. When the overhyped Y2K bug fails to materialize with anything but lots of annoying, but managable problems and lots of overtime work, what should a doomsayer do?

Fear not! Next out: the Y2.038 bug. Yes, I'm not kidding. The year 2038 will see a new date-related computer bug in old applications.

The problem stems from the C language definition of time_t. It is commonly defined as an integer giving the number of elapsed seconds since midnight, January 1, 1970. Most C and C++ runtime libraries define time_t as a long int. Since most system define long int as 32 bits, we have a range of 2,147,483,647 seconds from that date. This scope runs out on Monday, January 18, 2038. At that moment, all programs that were compiled on (very, very old-fashioned) 32-bit C/C++ compilers using these libraries will flip back to January 1, 1970.

I get scared just thinking of the insane hairstyles and fashion we'll see then...

Ah, I'm just happy that there will be work for an old computer programmer well into the next century. And the descendants (literal or spiritual) of our local prophet You Know will still have something to whine about!



Follow Ups:



Post a Followup

Name:
E-Mail:

Subject: *Comfort for Doomsayers

Comments:

Optional Link URL:
Link Title:
Optional Image URL:


[ Follow Ups ] [ Post Followup ] [ HOURGLASS2 OUTPOST ] [ FAQ ]