On Sun, Jan 02, 2000 at 04:58:57PM -0700, Warner Losh wrote:
> Karl, I was my hands of this conversation. You aren't listening.
>
> We have custom hardware. We're a control and measurement system. The
> <10ns is needed for that control and measurement part. The sync we
> get of the system clock, like I said before, is on the order of a few
> hundred ns on pentium hardware and a few us on 486 hardware.
>
> That's why we spend more $$ on the hardware. If you don't like it,
> forget the chill pill and just play in traffic.
>
> I'm done with this conversation.
>
> Warner
That's fine Warner, but isn't germane to the point.
Poul claimed that his favorite device was *superior* in *ordinary* use.
Then he admitted that he needed custom hardware (as you do as well)
to get that kind of accuracy.
Recommending something that is twice as expensive for people trying to
keep network time when the INHERENT stability of the system clock is at
least a couple orders of magnitude worse than the proclaimed PPS stability
is just flat STUPID.
Maintaining and defending that position when it is pointed out in bold
letters that the uncertainty on the system clock and interrupt latency
is FAR greater than the claimed 10ns resolution has only three possible
explainations:
1. The poster has an interest in the device being shilled for.
2. The poster has an IQ below that of tap water's temperature in
Celcius.
3. The poster has never in their life had to specify an uncertainty
with a measurement, which in and of itself disqualifies them
for the discussion at hand.
As for whoever the person is who force-removed me from the list, trust
me on this - I won't forget that act, and until you're identified and
permanently removed from both the list and the entire project you'll
have no contributions from me to your little treehouse project in any
way, shape or form.
--
--
Karl Denninger ([EMAIL PROTECTED]) Web: http://childrens-justice.org
Isn't it time we started putting KIDS first? See the above URL for
a plan to do exactly that!
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
