Bruno wrote:

> I never put that perspective at stake, just said that *now* is a great 
> time to show the NetBSD value to everybody else. If we can show that 
> NetBSD is better than FreeBSD right now in terms of performance the 
> NetBSD community can achieve the interest of new users or even 
> developers. And more people mean more scrutiny, even though the 
> community as is has allways done a terrific job - in a 
> quality/quantity ratio it's probably the best OS community around the 
> globe :)


I agree that it may be the time to do some mktg for netbsd. however, I 
don't think it should be oriented against freebsd. After all, how many 
people are using any *bsd? up so far, one of the great features of 
netbsd was the "no hype" choice. It would be easy to fall in the trap 
"they are weak now, let's get in", and then get nowhere.

If netbsd should talk and defend itself, it should do so independently 
of what is happening around. it should do so with the conviction that 
the message will be heard, and that the consequences will be in its 
favour. This is far from obvious. people at sun and other companies have 
been fighting for years, with a lot of money behind, to get where?

If you wanna defend the system, do it without relying on temporary 
failures of others, because if you don't, you might regret it tomorrow. 
Also, all this stuff about freebsd being unstable/unxxx is to be 
considered with care. fb 5 is still usable and much more stable than 
other systems, eventhough it is less stable from a bsd viewpoint (after 
all, the most widely used OS is the most unstable one:). So one should 
take the time to understand the meaning of somhing like "the 5 is not a 
production system". And I don't see a lot of *bsd in production systems, 
do you?
And  just because MD have quitted to create his own bsd isn't enough. 
after all, why didn't he not join netbsd? once upon a time TDR have 
quitted netbsd and created a successfull child. This should be 
considered as a great thing: since their creation, BSD systems have been 
proven to be usable to create new systems (Other systems just collapse 
once their code or their teams grow beyond some limit).

My meaning is that the mktg operation shouldn't be oriented "against" 
freebsd. that would only frustrate a lot of people. Instead, make it 
freebsd friendly (after all, the freebsd situation is an experience tha 
shows bsd systems can integrate other flavours while still being runnable).

The most valuable feature of netbsd is the "no hype" strategy: Do it 
right whatever they can say ("bien faire et laisser braire" dunno how 
this goes in english, but it means something like : do it right and let 
the donky shout). If we go into the hype world, let's only go once we 
know we will survive it. Linux, with all its forces and people is still 
having a difficult battle, how would we stand?

So yes, write, talk, and chat about netbsd. say it's great and 
wonderful. but say it the scientific way. no benchmarketing, no war, no 
hype...

with regard to the portability thing, I think we should use it more than 
we do. Apart from saying "mine is larger than yours", most people don't 
care if their OS runs on a lot of hardware, but:
- running on a lot of hardware implies some level of quality of the 
system source code
- the fact that most of the code is hw independent means that netbsd is 
ready for next generation hardware.
- running on a lot of hw means that the algorithms of the system are the 
right ones and aren't just simple hacks done for a given platform. This 
is important when comparing to freebsd and linux. both these systems 
started out as x86 kernels, patched long and wide to support other 
hardware (of course, both aren't just a series of patches, they have 
also been redesigned, but redesign is a risky thing, as we see in the 
case of freebsd...).

So let us say "It's the most portable system", "the champion of 
portability", ....

other things of interest are: security, flexibility (it is easily usable 
as a server, gateway, firewall, dev machine, desktop, ...), std 
compliant (both posix & gnusix), ease of use (well that lacks some 
docs), robustness/reliability/perf/... (needs real numbers but some 
tests we've done in my prev job leave me optimistic here), ...... and 
it's of course open source and open minded.

cheers,
mouss