Subject: Re: RZ23 problem, now with bonus RZ24!
To: NetBSD/vax Mailing List <port-vax@netbsd.org>
From: Brian Chase <bdc@world.std.com>
List: port-vax
Date: 03/22/2000 14:21:41
On Wed, 22 Mar 2000, Johnny Billquist wrote:
> On Wed, 22 Mar 2000, NetBSD Bob wrote:

> > As long as there was a switch there or query there to force a newfs,
> > if desired.  Sometimes, I really do want to blow away anything previous,
> > as cleanly as I can, after botched installs, bad crashes, system burps,
> > admin stupidities, changing flavors of OS, etc.  IF newfs did not, then
> > that would leave a llfmt as the only recourse.
> 
> Actually no. Since you are allowed to edit the disk label...
> 
> But anyway, I also think that it should be able to override. Further I
> think that newfs should ask instead of having a switch.
> 
> But that's just me.

We're talking Unix here though, so I would prefer the default behavior be
bloody and brutal.  There should be no interactive checking unless you
request it or unless it was established when the utility was first
introduced.  And even in those cases, the utilities should have
non-interactive override options.

If you take a machete and start chopping at your arm, then you shouldn't
complain when your arm falls off.  You should learn that "Oh it's bad when
I do that.  I won't do that ever again."

And then just for backwards compatibility, the default behavior shouldn't
change.  Adding explicit options to do the checks, or even adding a Unix
WIMP_MODE environment variable to check against would be okay. :-)

But even for newbie sysadmins, I believe that lessons in proper paranoia
are best taught by allowing them to hack off their own limbs a couple of
times.

-brian.
--- Brian Chase | bdc@world.std.com | http://world.std.com/~bdc/ -----
              And at work today, I screwed up by sending out an 
               apology for something I didn't even do!  -- K.