Subject: Re: take 2; which way should we go for /etc/rc...
To: None <>
From: Greg A. Woods <>
List: tech-userlevel
Date: 12/03/1999 01:47:56
[ On Thursday, December 2, 1999 at 23:37:00 (-0500), der Mouse wrote: ]
> Subject: Re: take 2; which way should we go for /etc/rc...
> I've always thought of NetBSD as a relatively niche OS, one that fits
> well the people who (like me) are into treating OSes as toys to be
> played with, into OS hacking for its own sake.  Loosely speaking, a
> research system.  Perhaps it wants to turn into something more
> "mainstream", more "a bigger user base (regardless of who) is Good".
> And if that's the way core wants NetBSD to go, if that's their metric,
> well, they probably can do it.

I've always thought of NetBSD as a really cool and highly portable
kernel that just happens to come along with a relatively orderly,
complete, clean, and usable (and equally portable) user-land.

However I've never once even considered calling NetBSD a good production
system, at least in its native form.  Sure it can be twisted into
reasonable form without too much trouble and it does make a solid base,
and it certainly makes more sense to use in some production scenarios
than most any alternative commercial platform, at least in my eyes.

I don't think though that a good production system needs to be at odds
with being a good research system at the same time (except for some of
the more obvious quandries that I think involve "security" more than
"robustness" and "maintainability" and "configurability").

I see this slow and painful process of cleaning up the system startup
and shutdown procedures as one of the many steps NetBSD has indeed made
towards becoming a more production-oriented system.  There are of course
many other facets being worked on and great progress has already been
made on many fronts (rc.shutdown itself being no small achievement)!

I do think NetBSD has to become even *more* amenable to being used for
long terms in very serious production environments.  A user-supported
system like NetBSD cannot survive if it is only a suitable as an OS
research system as otherwise it will inevitably become easier to do the
same research on the main-stream systems and indeed we are already
seeing increasing amounts of OS research (formal and informal) happening
on other systems.

I suppose this line of discussion should naturally wander quickly off to
netbsd-advocacy (which I thankfully do *not* subscribe to!  ;-).

							Greg A. Woods

+1 416 218-0098      VE3TCP      <>      <robohack!woods>
Planix, Inc. <>; Secrets of the Weird <>