I still think the rc.d stuff is needless complexity and just bowing down 
to the pressure of trendy fashion in the linux and sysV worlds. I 
personally find the monolithic rc script to be a much cleaner mechanism 
and will point out that as nice and pretty as some folks might think the 
present broken up netbsd rc.d is, just give it time to become 
disgustingly fragmented and start resembling ugly-as-hell linux rc's, 
which are a nightmare forest of environment vars and symlinks and all 
sorts of other ridiculous stuff. Theres no need for it to be like that. I 
cite the monolithic rc for NetBSD's rc's having remained so much more 
streamlined and simple for so many years, but now weve opened the 
floodgate for cruft and bit-rot in yet another avenue. that's my argument 
then and now. I'll continue using monolithic rc's, even if it is extra 
work for me to put them on newer systems, but it's just one more of the 
incentives to stop following newer netbsd releases..... the general 
mega-bloat and creeping linuxification/popularity contest stuff is the 
reason. 

as a side comment, i think the release cycle has been made artificially 
short in NetBSD. People are barely getting time ot fix bugs and make 
current releases stable when the ports have to contend with the new 
version's directives and must table the bug fixes and hardware support to 
work on much more bogus version-related stuff. On arch'es like i386 where 
there are a lot of developers, this might be fine, but other 
architectures are being increasingly marginalized, and it is only through 
heroic effort on the part of the core ports people that things still work 
and new hardware gets supported... 


 just my $.02
isildur




 On Wed, 27 Dec 2000, Jay Maynard wrote:

> On Wed, Dec 27, 2000 at 07:17:10PM -0500, Lord Isildur wrote:
> > yet another reason to ask 'why on earth did they do that?' I had a large
> > argument with several folks on port-alpha where this became a big issue
> > a while back. I never got an acceptable explanation as to why they went
> > with the (imho retarted) rc.d crap..
> 
> Since your message indicates that there's no reason you'd find acceptable,
> I'm probably wasting my time typing this...
> 
> For the rest of you, though, I consider it an eminently acceptable tradeoff
> to lengthen boot time - something that happens seldom - to make it much
> harder to screw up the system to the point that it won't boot.
> 
> The monolithic startup script should have gone the way of the dodo a long
> time ago. The only real objection to it has ever been in the BSD camp, a
> group known for virulent NIH syndrome.
>