Re: Competition (was: Ports & Packages ...)

[Jimmy Johansson <jimmy%Update.UU.SE@localhost>]

My response got kind of long winded and spread all over the place, time
to go to bed I guess...

On Wed, Aug 24, 2011 at 09:25:44PM +0200, Magnus Eriksson wrote:
> 
> This got a bit longer than I thought it would, sorry.  It's been
> brewing for a while, and not directed at anyone in particular.
> 
> 
> On Wed, 24 Aug 2011, Jimmy Johansson wrote:
> 
> >Nobody ever told me it was a competition.
> 
> >What is the prize and isn't Apple competing too?
> 
> The prize is having an at least half-decent OS that runs on commonly
> available hardware. NetBSD is still in the race, for most values of
> "runs" and "commonly", but honestly by now I'm looking for
> alternatives for whenever the next hardware upgrade is due; I'm
> under no illusion that I could just grab whatever I find online at a
> good price and expect it to run.  There's been far too much talk
> along the lines of "so and so graphics cards don't work, we don't
> have kernel support", and too much problems with e.g. wifi, for me
> to assume anything - other than that it will most likely continue to
> run fine for server duties on the platforms that core developers
> have access to.

First, maybe I was a bit flippant in my response there, I was in a good
mode though and couldn't resist.

Second, I think you can. You just have to take a bit of care. I think the
same holds if you for example would like to run Linux. With care I mean
get a few candidates and then google for a few minutes. I think I would
do the same when building a system running Windows. Some hardware works
better than other.

I actually built this machine a few weeks ago. It runs NetBSD 5.1 if I
disable ACPI and NetBSD current I could install without changing
anything, but it was unstable until I set "USB controller mode" to
FullSpeed and after that it is rock solid.

As a side note, Ubuntu linux doesn't even boot on this system, I just
get a purple screen. I tried Ubuntu when trying to see why the system
was unstable.

Most of the hardware that NetBSD doesn't run on is probably hardware I
would stay away from anyway imho.

I guess I might be lucky picking hardware though. :-)

[snip, virtualisation]

> And in the thread that was just posted on -users, on whether NetBSD
> runs on platform X, the answer was "it's just a question of time and
> effort". Developer time and effort.  So why does this community seem
> to think that it's somehow a bad thing trying to increase the pool
> of users and potential developers?  Every time the issue comes up,
> people just say "ugh, don't like the hype" and go back to coding on
> their Macs.

Isn't this one of the problems in most communities where the members do
things in their free time? I mean, I believe that most of the code added
to the Linux kernel now a days is added by paid programmers during work
time. Ubuntu uses the work done by the Debian project and then add
advocacy, some own code and documentation done by paid staff and also
stuff from the community for sure. It is kind of hard for a project like
NetBSD to compete with that, most of the people that write code for fun
after work on their own time wants to do what they find fun, and I can't
question that, their time so they do what they want. And this isn't
likely to change, unless somebody joins who just think that advocacy it
the best thing ever.

> Just yesterday, I mentioned the fact that I was running NetBSD in a
> chat, _on a system that itself runs NetBSD_, and was met with what I
> must assume have been a blank stare on the other end; and "...but
> why?".  That's one guy who won't be trying out our favorite OS, no
> matter its technical merits.
> 
> Choosing your main OS isn't for most people a matter of
> dispassionately counting a platform's merits and drawbacks using
> some checklist, it is often a matter of minimizing mental effort by
> just going with what people around you are using, and being nudged
> by the occasional "hey, this stuff is really cool".
> 
> So lets talk to people about the stuff that is cool once in a while.
> 
> (and why not send some suggestions to -advocacy, it needs some
> traffic other than spam)

What I find cool is usually stuff that makes peoples eyes glance over
though. Started in school "You mean that you like math?" and it carries
on to this day. I'm no fun at parties at all! :-)

> >On a more serious note, as far as I know the NetBSD developers are more
> >focused on delivering a good, well designed and solid operating system
> >rather than competing with linux.
> 
> Just like every other developer community out there..  So why aren't
> you and I running AROS, Haiku or Plan9?  Anything in particular that
> is wrong with their architecture?

I actually kind of run Plan9... Just for fun though. See cool above.

> I suspect not; at least for me the answer to the "but why" question
> above is as simple as that NetBSD was available on the platform I
> was using at the time, met the general idea of what I wanted ("free,
> unix"), and I knew about it.  Clean design only mattered once I had
> multiple options that all seemed roughly similar; in my case I
> switched to, and back from, Debian.
> 
> 
> >I have used all of NetBSD, FreeBSD, and OpenBSD.
> >...
> >All of those projects have merit and I don't see any of them going
> >away soon.
> 
> They'll all be around for a long time.  The issue for me isn't so
> much whether NetBSD will magically disappear or stop working in a
> year or two, but the general experience (what hardware hoops do I
> have to jump through to make it run?) and long-term sustainability
> (what happens when the currently active generation of developers
> lose interest?).
> 
> I don't want to come off completely negative, there are several
> positive signs - GSoC projects means new developers get involved,
> the emips port looks really cool (and unique!), MP work on Xen, etc,
> etc.  But we need some advocacy, and the attitudes towards that are
> IMHO weird.

I acutally read the FreeBSD thread the OP had in the original post and
they had some good points in that thread. Sure, the NetBSD project needs
some advocacy, but I also think that that it needs a way that makes it a
bit easier to get involved in the NetBSD project in some way.

As it stands today it feels like NetBSD is geared more toward the
experienced Unix user. Installing NetBSD for the first time can be kind
of daunting and documentation can be hard to get at. The amount of
advocacy doesn't matter if a possible user decides to try NetBSD and
then gives up because he can't install it.

I think that what I'm trying to say is that if the project wants to get
a bit more main stream to "compete", then I think you have to both work
at advocacy and on delivering what the potential user wants. You also
have to define a direction, a goal. For me, as an outsider, I see NetBSD
as a solid server OS, an OS for embedded systems and also an OS for
enthusiasts and programmers who want to work on something in their free
time with mostly clean design and a good code base. As far as I know it
has never been geared towards the casual user.

For example Ubuntu is though. They want to deliver an OS for the masses.
I'm not sure that is a direction the NetBSD project wants to take and
I'm not so sure it is a direction that they have to take either to stay
relevant.

> sig:
> 
> >If you don't shoot the bearers of bad news, people will keep
> >bringing it to you.
> 
> I already have my asbestos suit on.  Does the bulletproof vest go
> over or under it?

Maybe I like bad news? :-)

Regards,

Jimmy
-- 
If you don't shoot the bearers of bad news, people will keep bringing it to you.