Re: der Mouse [Re: Path to kmods]

[David Brownlee <abs%NetBSD.org@localhost>]

On Sun, 8 Feb 2009, lucio%proxima.alt.za@localhost wrote:

> Possibly, but it's called making a virtue of necessity.  I'm (also) a
> Plan 9 user and I see precisely the same symptoms on a much smaller
> scale on that side.  Fact is, the i386 rules the roost and there has
> been a much more pervasive shift towards accommodating (actually, I
> guess it's bedding in with) that reality.
>
> In short, investing in theoretically sound technology tends to be too
> expensive unless there is also a return on investment on the i386 end
> of things, which basically means that NO investment actually occurs in
> any other direction.
>
> I'm sure core will attempt to deny this bias, but it's there, _maybe_
> thinly camouflaged.

        How would this explain all of the added support for arm
        processors and related devices?

> Don't hold it against core, they have survival in mind.  Instead,
> those of us who might perceive the shift to a monoculture to be a
> short-term survival strategy must look at means by which diversity can
> be sustained, possibly at some (great?) cost.
>
> I guess I ought to go into details, but I must honestly confess that
> it is all rather blurry.  By all means help me define it all more
> clearly, that is an activity we can all contribute to if we care
> enough.

        On a volunteer project people work on areas which interest
        or benefit them, and code which is contrbuted back from
        companies by definition tends to be from projects which
        the company felt it could make money from.

        So... for NetBSD nowdays that tends to mean quite a lot of
        work on x86 and embedded ports (arm, mips, powerpc), plus
        the occasional developer who is scratching an m68k or vax
        itch.

        The needs of an embedded box are pretty different from
        those of a desktop, vs those of a xen DOM0 running on a
        large amd64 box, but they can all build from the same NetBSD
        kernel source, and in many cases share the same userland
        (or subset in the embedded case) and even some of the same
        pkgsrc packages.

        A general call to anyone still listening, if you have a
        specific use for NetBSd, then contribute towards that use,
        be it code, pkgsrc, documentation, updating wiki.netbsd.so,
        or just helping out people on the list. If someone else is
        contributing towards NetBSD in different areas, rather than
        just saying "Thats bad", look to see what they are doing
        and point out any specific issues which might affect your
        use of NetBSD, don't harass them about issues which will
        not.

        As an example - over the last couple of months I've been
        helping some people rework the atari install setup, including
        adding optional support to sysinst for gemdos filesystems
        and handling the case where the install sets are on a
        filesystem which truncates to 8.3. Its absolutely port
        specific and doesn't help non atari users at all. Should
        they all now be complaining that its taking NetBSD away
        from their focus?

--
                David/absolute       -- www.NetBSD.org: No hype required --