|
|
Posted by Moe Trin on 06/26/07 01:59
On Sun, 24 Jun 2007, in the Usenet newsgroup comp.os.linux, in article
<pan.2007.06.24.19.14.41.843749@zianet.com>, ray wrote:
>If you'd like to elaborate, I be interested in why you have the
>requirement that time be monotonically increasing.
Talk to the friendly government banking oversight agencies. It _may_
be part of Sarbanes-Oxley Act of 2002 (Public Law 107-204), but the
idea is the trustworthiness of the data. It's that old "what did
they know, and when did they know it" problem. (I was hoping that
Professor Mills included something about the reasoning in the RFCs,
but a quick scan doesn't turn up an obvious reference.) The main
point is that there are some situations where _changing_ the clock
setting is a Class 1 felony. To keep the system in time, you are
allowed to _drift_ the clock, but as in the world, time never goes
backwards, nor stand still.
The "source" to the frequency that the O/S is using to keep track of
time is usually a US$3.50 (in unit quantities - cheaper in volume)
quartz oscillator. The specs _USUALLY_ are something like +/- 0.01
percent accuracy (+/- 100 ppm). Translated, that's plus/minus 8.64
seconds a day. The original IBM PC had a crystal on the motherboard,
along with a variable capacitor, located in a metal shield (on the AT,
this was next to the second 16 bit ISA slot on the left), and you
could "set" the 1.8432 MHz oscillator frequency to minimize the error.
By the time the PS/2 came out, this had been replaced by the sealed
can type of oscillator. Today, the oscillator may be contained in the
BIOS chip. Because everything is sealed, there is no adjustment, and
the O/S corrects for timing errors by varying the division ration of
that 1.8432 MHz signal - that in turn generates the INT (or IRQ) 0
interrupt in an x86 class system.
>In the real world, things happen,
And when thing happen, you document the crap out of it, and then figure
out how to prevent that crap in the future. while you also hope that
the Feds (perhaps the SEC) will buy your story.
>and no software is perfect or bug-free.
Cue the joke about an update to correct an error invariably ADDs more
bugs.
Old guy
Navigation:
[Reply to this message]
|