argh, weve had this thread before, to its exhaustion! 
Many people need ACLs. Theyre not inherently bad. The discussion was (and 
i hope will soon return to, if this thread continues) on the details of 
doing them right. Some folks (myself included) argued strongly against 
adding them to ffs specifically, or if it was, then against doing it in 
a non backwards compatible manner- there was a good bit of talk of doing
them in such a way to meet the demand that an acl-enabled ffs could be 
mounted and used, read-write, on a system with no acl support, without 
damaging or corrupting the acl information, and also a non-acl ffs be 
used on a system with an acl aware kernel, which does not cause the addition
of acl features to the filesystem but is still useable by that acl aware
system, just sans acls. Some methods of doing this were discussed. Yes, there
was a lot of flamewar going back and forth but luckily that ended after a
few days. please whoever is trying to resurrect that flamewar, go read
the old one. Many folk (again myself included) wanted to be very sure such
a thing would always remain strictly an option ( for example, nothing in
the distribution would require or depend on it except the obvious cases like
acl-related commands) and also the issue of what granularity to have was
discussed. per-file is pretty involved, but there is something conecptually
cleaner about doing things per-file, since a file is a file is a file in
UNIX, but many also argued that directories would be the better 
granulaity. I dont think a conesensus was reached on most of these 
issues, and the thread just faded out but lets not have it again unless 
we pick up where we left off. Many of us dont see ACLs as very important. 
I for one dont really need them. I can see what situations people do want
them though. As long as it is done the right way, it will be a good addition
to the system. 

Isildur

On Fri, 24 Aug 2001, Rick Kelly wrote:

> Luke Mewburn said:
> 
> >As discussed last time ACLs came up, just because *you* don't have a
> >use for ACLs does not mean that ACLs are useless for other NetBSD
> >users.
> 
> >If a suitable ACL implementation was made available for NetBSD (e.g,
> >what's in TrustedBSD), I see no problem with it being added as an
> >option to the system, just like quotas, endian-independent ffs file
> >systems, and other such optional features.
> 
> As long as it is truly optional. It's useful if needed, but otherwise
> adds uneeded overhead to the file system. 
> 
> -- 
> Rick Kelly  rmk@rmkhome.com  www.rmkhome.com
>