Subject: Re: PC: why January 4, 1980 Reply-to: GTBecker@RighTime.com
> Can someone explain to me why some PCs read January 4, 1980 > following the basic year-2000 test (setting clock to 11:59, > December 31, 1999 and turning computer off). Why not January 1, > 1900 in all cases?
The information you seek is available in the FAQ at
www.year2000.com/pub/year2000/y2kfaq.txt
and in several texts that are on
www.RighTime.com
Only a little searching will find it. I'll relate the summary again.
All AT-class PCs (all IBM-compatible PCs, 286 through Pentium, from the PC/AT of 1984 are at least AT-class) contain a hardware clock calendar component, the CMOS RTC (Real Time Clock). The original CMOS RTC device was a Motorola MC146818 which only maintained a two-digit year. Compatibility has dictated that every machine since behaves the same way. Although the century is stored by the BIOS in the CMOS portion of the device, it is not maintained by the hardware clock/calendar; consequently, in the CMOS RTC, year 1999 is followed by apparent year 1900.
When a machine is booted, the operating system (DOS, all Windows through Workgroups and Windows 95) initalizes its date - via the BIOS - from the CMOS RTC date. The operating system misinterprets any date in 1900 (which should be 2000) as January 4, 1980.
Some BIOSs apply windowing to the two-digit CMOS RTC year to infer the century, so machines equipped with such a BIOS will set the operating system date correctly, even though the CMOS RTC still says 1900. Some BIOSs will go a step further and correct the century in the CMOS RTC, but only at boot or when the time or date is set; they will not - in fact can't, reliably - correct the CMOS RTC century in real time. These BIOS behaviors are recent innovations, so any machine with a BIOS earlier than 1994 or so is unlikely to contain them, and many sold even today do not.
Some BIOSs will not accept nor report any date after 1999; some BIOSs will report 2096, rather than 2000. Both are by Award.
Rather than explain the details behind these behaviors, let it suffice to understand that the basic problem is caused by a design error in 1984, a two-digit-year hardware clock component selection. Some recent machines sufficiently conceal the flaw for most users, but it is present in _every_ machine, even one purchased today.
For more details read Year2000.Txt, RT2@PTTI.Txt and RighTime.Txt at www.RighTime.com.
Tom Becker
--... ...-- www.RighTime.com The RighTime Company Vox: (305)644-6500, Fax: (305)644-6515, BBS: (305)644-6185 Air System Technologies, Miami Vox: (305)644-8898