On Thu, 20 Mar 1997 13:03:21 -0500 (EST) 
 der Mouse <mouse@rodents.montreal.qc.ca> wrote:

 > > At the very least, the "point-and-drool" crowd has money.  Money ->
 > > donations -> more resources for development.
 > 
 > True.  And if that's what you care about...why don't you just go work
 > for Sun, or SGI, or DEC, or any of those people who's selling a
 > "solution"?  Hell, even Microsoft; there's even more money (and
 > therefore resources for development) there.

...oh, c'mon.  Resources spent improving NetBSD are a good thing.
Obviously, since we give the software away with full source code,
money is not "what we care about".

However, money is required to purchase things like the the DAT changer
and tapes for the CVS server, and disks when we run out of space,
and new hardware that we want to support.

 > Yeah, again, I said something slightly different from what I meant.
 > Binary packages per se don't bother me; I just ignore the silly things.
 > What I really don't want is to see something like SGI's inst: it uses a
 > complicated and undocumented package format, is dogmatically inflexible

Hmmm ... are you sure it's not the package format described in the
SVR4 ABI?

 > about where it installs things, keeps records in an undocumented format
 > in an undocumented place...and it is the only supported way.

Oh, well, we won't suffer from that :-)

 > Again, this doesn't bother me per se.  What bothers me is the
 > possibility that the point-and-drool people doing real work will end up
 > being the market NetBSD aims for.  No thanks.  If I want that, I know
 > where to find it.

I don't forsee NetBSD's goals changing .. however, it's good to reevaluate
the approach used to achieve those goals.  Having spoken with several
"OS guys" about this at various times, I believe that providing a
packages subsystem would be a good way to help NetBSD in the OS research
market.  It's also a good way to help NetBSD in the ISP market.  It's
also a good way to help NetBSD in the "PC at home" market.

The appeal of such a subsystem is global.

 > Thinking back over the binary packages I've seen (yes, I have worked
 > with them - at work) and the source packages I've seen, and I compare
 > the high end of the quality spectrum for the binary packages with the
 > low end of the quality spectrum for the source packages...I'll take the
 > source package.
 > 
 > I compare the money...and I'll take the source package.
 > 
 > I compare the support...and I'll take the source package.
 > 
 > Every time.

Well, not everyone agrees...so you support both.  Problem solved :-)

Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                               Home: 408.866.1912
NAS: M/S 258-6                                          Work: 415.604.0935
Moffett Field, CA 94035                                Pager: 415.428.6939