[ntp:questions] Re: A utility called "S"?
Richard B. Gilbert
rgilbert88 at comcast.net
Sat Jan 3 00:34:24 UTC 2004
Ah, nostalgia! The model 26 card punch!!! I still have a card saw and
a registration gauge somewhere! I misspent my youth in the Princeton
University Computer Center. . . . Just one more compile. . . . Now I'm
misspending my old age. :-)
The easy fixes will be along shortly. The difficult will take a little
longer, if I can do them at all.
David L. Mills wrote:
>Peace! Those scripts were written twenty years ago when I was a young
>man. Now in the modern time I repent dearly my misspent youth. The
>scripts are easy to fix and so I will the adventure to you. Should you
>so choose, please post your adventures so those evil old scripts can be
>consigned to the god of IBM 704 punched card. And I punched a few of
>"Richard B. Gilbert" wrote:
>>I've been using gawk to extract the records for each server and Gnuplot
>>to plot offset, delay and dispersion for each day.. I stumbled across
>>the awk scripts while trying to find a different awk script entirely
>>with the intention of recycling pieces of it do some simple statistical
>>studies that your scripts already do
>>The existing summary.sh script does not appear to be entirely Y2K
>>compliant; if I'm reading it correctly, it's still looking for
>>peerstats and clockstats in the 20th century. It also appears to expect
>>the stats files in the same directory where it and the awk scripts
>>live. I think I can fix that and, perhaps, also modify it to use
>>Gnuplot. I'll be happy to share if I can get it working.
>>David L. Mills wrote:
>>>Brace for a blast from the past. The S package was distributed some
>>>years ago by AT&T (no kidding). It was quite a classy package and
>>>produced wonderful publication documents on Sunses and allegedly
>>>Windows. We used it here for awhile and it might even yet be available,
>>>but real engineers now use Matlab. Every few months or so I make another
>>>small step to overhaul the scripts for Matlab, but that ain't done yet.
>>>Volunteers are always cherished.
>>>I built the statistics files especially friendly to common Unix tools.
>>>The scripts are designed to bun once a week or so and produce summary
>>>reports for usage statistics, general monitoring and fault detection.
>>>There is also a little dormant project designed to watch the statistics
>>>in real time and call a beeper if something goes wrong. The mind
>>>"Richard B. Gilbert" wrote:
>>>>In .../ntp-4.2.0/scripts/stats/summary.sh we find:
>>>># Process clockstat files with S and generate PostScript plots
>>>>While berivity may be the soul of wit, "S" is a little too brief to be
>>>>useful unless you already know where to get it and how to use it. I'm
>>>>imagining trying to search the web for "S"....
>>>>Can anyone spell this out for a poor newbie?
More information about the questions