[ntp:questions] Re: Latest changes toTWikiGettingStarted

David L. Mills mills at udel.edu
Mon Oct 6 03:00:29 UTC 2003


Kind sir, please don't lecture me on my personal torque of the English
language, which I have been practicing in public probably since before
you were born. I can tell you haven't read my journal or symposium
articles and noticed in each a little Millspeak, which journal editors
know well and double-blind paper reviewers spot in a flash. I have had a
lot of good clean fun fun over the years turning a quaint or two in
otherwise dull technical prose, and I ain't gonna change notime soon.

Having said that, I don't agree at all with your assumption the target
machine has to be set up and presumably secured before a browser is
installed. Why assume the documentation is viewable on the target
machine at all? I've got some machines running NTP that will never
support a browser. Several machines on the wires here and in several
foreign countries are or have been installed and serviced via serial
console exactly as you say. For entertaining examples, see the SIGCOM
paper on the Fuzzball at www.eecis.udel.edu/~mills/papers.html.

To clarify the mantoad/webfrog thing, Harlan gets mad at me when I ask
that alternate formats not be provided in the documentation collection
itself, but that they be easily produced using supplied scripts. The
reason for this is both to slim the bloat and to make sure the original
reference pages are never obscured by translation unless the translator
fully understands that ambiguities and mistakes can occur. Multiple
translations would require each one be officially vetted (by me I would
assume) to insure meaning was not lost. That's hard to do when working
from a hypertext to flat documents.

As you say, you might be comfortable with that, but as you say the other
problem is the content itself. The content of the NTP reference pages is
due to several volunteers including me and, together with the NTP
project page, represents work spanning two decades. However, it is still
a work in progress and needs reorganization to appeal to modern tastes.
The twichy is no substitute for that, although it might be a good tool
to suggest reorganization style.

Having shot my load, I grow weary of the quest. You get the final word
and then I promise to get off the bus and let the other folks here read
something more interesting.


Brad Knowles wrote:
> At 8:14 PM +0000 2003/10/05, David L. Mills wrote:
> >  Can you cite examples where a sysadmin in Outer Mongolia wishing to
> >  install NTP in a Chinese ripoff of Unix for an IBM mainframe (!) does
> >  NOT have access to a browser of some kind?
>         When I install servers, I do so completely disconnected from the
> network, or at least to the greatest degree possible.  Ideally, it's
> me, the machine, a serial console, and a CD-ROM.  Even for machines
> that are in a co-lo facility thousands of miles away, I'll still do
> it the same way -- the only difference is that the serial console is
> a connection to a terminal server that I can connect to via ssh.
>         If a web browser ever gets installed on the machine, it will most
> likely be well after the machine has been set up and reasonably
> secured (host level firewalling, etc...), and even then it will be a
> text-mode browser like lynx or links.  Even that's pretty rare.
> >                                                              With this in
> >  mind, I tried to avoid anything fancy in the HTML format, so that those
> >  pages should light up fairly well even in a text-only browser. Most
> >  product manuals today require Acrobat, but I won't get into that.
>         So long as all the HTML documentation is also available in a pure
> text format, I wouldn't have any issues with that.
>         However, access is only one part of the problem.  The other part
> of the problem is the content itself.
> --
> Brad Knowles, <brad.knowles at skynet.be>
> "They that can give up essential liberty to obtain a little temporary
> safety deserve neither liberty nor safety."
>      -Benjamin Franklin, Historical Review of Pennsylvania.
> GCS/IT d+(-) s:+(++)>: a C++(+++)$ UMBSHI++++$ P+>++ L+ !E-(---) W+++(--) N+
> !w--- O- M++ V PS++(+++) PE- Y+(++) PGP>+++ t+(+++) 5++(+++) X++(+++) R+(+++)
> tv+(+++) b+(++++) DI+(++++) D+(++) G+(++++) e++>++++ h--- r---(+++)* z(+++)

More information about the questions mailing list