Re: [gpsd-dev] Clarifications needed for the time-service HOWTO

[Andy Walls]

On Wed, 2013-10-23 at 10:35 -0400, Eric S. Raymond wrote:
> Andy Walls <address@hidden>:
> > WRT adjusting fudges:
> > 
> > http://lists.ntp.org/pipermail/questions/2011-November/030951.html
> > 
> > I happen to use the "calibrate by hand" method mentioned in the post.
> > Open-loop is easier for me to follow.
> > 
> > I also have this cheat-sheet lying around for examining the data from
> > open-loop tests:
> > 
> >         http://www.eecis.udel.edu/~mills/ntp/html/monopt.html
> 
> I sense a lot of technical depth and value in what you've put together.

Thanks.

I'm a visual person.  I find plots indispensible for rapidly helping me
figure out my next move in improving/fixing a system.  For me, they make
the important and unusual stand out from the unimportant. Separating the
wheat from the chaff in the measurements becomes easy with plot.

E.g. plotting the offset's jitter as error bars on the plot of the
offset vs. time lets one see at a glance the intervals when a clock was
jittering badly and when it was not.  The same goes for the frequency
offset and wander (freq jitter).  

Alogrithmicly determining the good intervals from the bad intervals
isn't impossible; it's just a pain when the 'grey matter' computer can
do it so easily.


>  But
> Beat Bolli has done a better job of making the tuning process accessible in
> a way that fits the goals of this HOWTO.

Yes, he has made getting an answer to setup/configurations questions
much more accessible to the uninitiated.  That is sorely needed.

I've spent a lot of time getting NTP to synchronize systems on isolated
networks.  This forced me to read a lot of NTP documentation and Dr.
Mills' briefs and papers in the process.  Not what I really wanted to do
at the time.  :(


> Would you consider collaborating with me, and Beat if he's interested, in
> building and documenting some more advanced tuning tools?  This probably
> wants to be its own project, "ntptune" or something like that.

I would, but my spare time is extremely limited. (Heck, so is my for pay
time.)  I can almost guarantee I cannot produce end products in a timely
manner.  For example, linux ivtv and cx18 driver bug fixes and
maintenance have a very long turn-around time for me these days.

Regards,
Andy