Appologies for bringing this up again, but I was talking about this
_briefly_ at dinner tonight (thanks Erik), and forgot to cover the
following which I figured _might_ be of wider interest.

The main reason I like start/stop scripts has nothing to do with the
issues of boot/shutdown, but rather day to day maintenance and more
importantly fault resolution via phone etc.  Being able to tell an
operator (assumed to be clueless) to run

/etc/init.d/ntp stop
/etc/init.d/ntp start

s/ntp/sendmail/ or any other daemon that normally runs all the time but
occasionally needs to be restarted to deal with a fault.  Is far less
stressful for everyone than explaining how to run ps(1), grep(1) etc.

I actually put my start/stop scripts in /etc/rc_d (to ensure they
don't clobber vendor stuff) and they generally provide a restart and
check target as well so that a cron job can run

/etc/rc_d/ntp check

regularly, and the above mentioned operator could have been told to run
/etc/rc_d/ntp restart

and of course once you have the start/stop/check/restart scripts for
the above, it makes a lot of sense to leverage them for booting.  I
generally do nothing for shutdown (on BSD anyway) - the SIGTERM from
init is normally all that is needed, but its handy to have the hooks
there anyway.

Ok, back to work.

--sjg