[ntp:questions] Re: kiss-of-death packet - terminology

Brian Garrett mgy1912 at cox.net
Wed Aug 6 19:23:22 UTC 2003

"Vernon Schryver" <vjs at calcite.rhyolite.com> wrote in message
news:bgrj0p$i0s$1 at calcite.rhyolite.com...
> >I myself play in the Professional Arrogance League, so I am in no
> >position to lecture.  But we are gathered here to fuel the NTP Jihad,
> >and its goal is to deliver as much benefit as possible to the human
> >race through the magic of NTP.  Anything which gets in the way of that
> >goal (including one's emotional foibles) must be set aside.
> >
> >I think we can safely say that nobody will object to that reasoning.
> >(Because it would be far too embarrassing to do so.)
> I think that position is itself embarrassing.  At this late stage in
> its life, it would be embarrassing for the protocol NTP to admit caring
> how many more flock to its banner, if protocols had feelings.
> >In practice, though, one of the most difficult barriers to getting
> >software adopted is not its technical excellence but its ease-of-use.
> >"If you reduce the IQ needed to run it by 5 points, you quadruple the
> >potential market."  And since the most rapidly growing part of the
> >Internet is non-English-speaking, minimizing language difficulties is
> >becoming an important issue.
> Even if you accept the goal behind that statement, removing the colorful
> language that has always been a hallmark of NTP would not be useful.
> NTP cannot be a protocol for dummies because the underlying notion of
> synchronizing a rabble of tickers into a hierarchy is too subtle.
> It's not that people can't understand it, but that they are sure they
> understand all they need to know but in fact don't have a clue.  Even
> protocols that try to do simpler things like timed and SNTP are too
> complicated for dummies.  (Ask me for war stories about dummies doing
> their impressive best to wreck timed installations based on how they
> "knew" the protocol and code worked.)  Dealing with dummies who think
> they are too smart to need the easy effort of consulting the web to
> decode "kiss of death" or "kiss o' death" is a frustrating waste of
> time.  The best way to serve dummies is to make the documentation more
> opaque to force them to hire smart people, or more accurately, people
> who will make the effort required to understand.  The dummies are
> usually dumb not because they lack brains but because they stop when
> as soon as they think they understand, often before reading anything.
> They generally cannot be convinced they missed anything, but odd turns
> of phrase like "kiss o' death" might help.
> Suffering fools should be left to the professionals in that field,
> not by dumbing down the universe to be comfortable for fools.
> All that asside, I see little or no evidence of real confusion, but
> plenty of envidence of people with a need to stick their oars into
> the stew pot, and never mind the danger of spilling everyone's supper
> into the fire.  Too many people figure that if they can't make their
> marks by inventing a protocol, they can do it by what they are pleased
> to call contributions.  If I didn't already know that from watching
> this and other circuses, I'd have been taught it quite forcefully
> during the last year or two with my anti-spam thing, the DCC.
> Vernon Schryver    vjs at rhyolite.com

What he said.

In view of the many non-native speakers of English on this newsgroup who
speak English better than most Americans and have clearly had no difficulty
coping with Millsspeak up till now, I have to regard any debate about
"kiss-o'death" and other such colloquialisms in the NTP literature (apart
from RFC's) as specious.  Apart from the ubiquitous "my computer is
displaying the wrong time zone" complaints, most of the problems I see
people asking about in this newsgroup are due to problems with their
specific environment that the documentation does not necessarily address.
(I'm talking about the more unusual problems, not the frequent stuff that
could be solved by RTFM).

Let's face it, folks--if someone is intelligent enough to be hacking for a
living and fluent enough to be able to understand _any_ of the NTP
documentation (or even to understand what a KoD packet _is_, for gods' sake)
they're smart enough to cope with any dialect issues at the University of

FWIW, IMHO, YMMV, substantial penalty for early withdrawal.


More information about the questions mailing list