Yes, BSD is a bit more fractured, but at this point I don’t think we can
consider FreeBSD the same operating system as NetBSD, DragonFly, or
OpenBSD.  There is a lot of cross fertilization, true, but the core kernels
are actually quite different under the hood.  I know that kernel software
is boring for most folks, but I still consider it the heart of an OS.

Most of the illumos distros have what is essentially still the same
kernel.  Maybe after time they will drift apart to more closely resemble
the situation with the BSDs, but I think our community is too small to
sustain that.  There are a few commercial contributors, but apart from that
there only a few hobbyists.  (And, if I can speak frankly, only a handful
of the hobbyists have shown ability or motivation to work in the kernel.)

No, what we have is more like Linux distributions, with some kernel
differences (Ubuntu is not the same kernel as RedHat after all, although
they are 99+% the same), but mostly differentiated in what they have in
userland.  Most especially in packaging, but also in administrative tools,
etc.  (Systemd may have been driving the differences further apart, as
Linux distributions look less and less like traditional UNIX.)

Some years ago I had an idea to try to create a “core” distribution, sort
of a minimalist distro for other distribution builders to enhance or derive
from, but also to act as a kind of “minimal OS” for hosting & testing the
illumos-gate software.  My goal was POSIX compliance, with legacy SunOS
compatibility as distinct 2nd case.  (For example, I changed the default
uname output to reflect that SunOS is no longer meaningful, and frankly in
my tree I had changes which made the compile-time environment more like
POSIX, and even Linux, than SunOS.)

Lots of this was controversial, and I sort of abandoned the effort due to
lack of interest, and a realization that I was probably tilting at
windmills.  The fact that I didn’t have any commercial support behind me,
and the primary commercial illumos player (Joyent) had expressed adamant
opposition to the effort led to me to realize that my efforts were almost
surely fruitless.  I think this is unfortunate, as illumos has relegated
itself to compatibility with a system that few still use or care about (and
the current number is dwindling), and our big “compatibility” play is now
the LX zones in SmartOS.

I have some ideas for ways to improve illumos and help it remain vital and
relevant.  Unfortunately, there seems to be no support for most of those
ideas from the largest commercial contributors, who have their own agendas.

Ultimately, these separate agendas almost guarantee that illumos itself
will have to fork into different systems; from what I can see the only
other option is for everyone to just accept the leadership of the primary
commercial contributor, and be happy with what they provide.  I don’t think
that is what anyone (including said commercial contributor) really wants
though.

I’d be willing to work with others on some of these ideas, resuming some of
the work I began years ago, but I’ve long since realized that it is both
more than I can do alone, and not work I can afford to do for free — since
I still have things like a mortgage to pay for.


On Fri, Sep 15, 2017 at 8:36 AM Alan Coopersmith <
[email protected]> wrote:

> On 09/15/17 02:18 AM, Alexander Pyhalov wrote:
> > FreeBSD doesn't have this problem. They have different teams, interested
> in
> > different areas, but still share single code base.
> 
> Except for the folks who have different ideas and work on NetBSD, OpenBSD,
> DragonFly, etc. instead.
> 
> https://en.wikipedia.org/wiki/List_of_BSD_operating_systems is a lot
> longer
> than the list of illumos distros.
> 
> -alan-

------------------------------------------
illumos-discuss
Archives: 
https://illumos.topicbox.com/groups/discuss/discussions/T83f198c8597cf8e3-Mccdcb17006aac12eb5d45329
Powered by Topicbox: https://topicbox.com

Reply via email to