[OpenIndiana-discuss] Rolling release considered harmful

[Judah Richardson]

On Mon, May 3, 2021 at 9:21 PM Reginald Beardsley via openindiana-discuss <
openindiana-discuss at openindiana.org> wrote:

>  A response to a couple of Judah's posts.
>
> I experiment with many operating systems including Plan 9 for which I got
> one of the first release floppies at Usenix '95 and chatted about it with
> Dennis for about 20-30 minutes.
>
> My interest in OI is not curiosity. I want a stable, industrial strength OS

Quite a few of these exist, in various forms and paradigms. The world - and
indeed organizations with needs far more complicated than that of any
individual user - runs on them. Many of them may not function internally
*exactly* the way you'd prefer or be sufficiently documented for you to
understand *exactly* what they are doing, but they have significant
developer attention and there are far more people able to help you with
edge case issues and/or customize things to your liking.

Today's world is far more interested in what systems can do (e.g. can I run
a service?) than in exactly how they do it (e.g I refuse to use any other
init system than <insert favorite here>).

and OI/Illumos has to date been my best option. I've used over a dozen
> non-Unix model OSes and close to 2 dozen Unix model OSes. Possibly more.
>
> I do *not* want to be running alpha software for production use, even in
> the context of a single user, home environment.

>From a practical perspective, I don't think OI has the option you want.
It's a rolling release distribution. I think OmniOS has been mentioned to
you previously but you'd be on your own with a DE there. However, I do
think you may have some success if you put at least much time in trying to
get a DE on OmniOS as you have trying to bend OI to your will. The reason I
say that is AFAIK OmniOS is a bit like the FreeBSD of Illumos in the sense
that it doesn't ship with a DE but running one on it does not necessarily
violate its paradigm. Anyway, I'm not sure how dead the OmniOS horse is on
your end, so there's that.

I'm quite happy to run alpha software for internet access, but not for my
> main system. For that I require the near bullet proof quality of
> traditional SunOS.
>
> Over the years I've watched many names disappear from the OI mailing list.
> I do not know if they still are using OI or not. Many of those are of my
> generation and frame of mind. I have gotten personal emails from some
> supporting my comments, but many appear to be gone forever. They are a
> significant loss.
>
> In short, if I am forced to run raw alphas I shall go elsewhere if I can
> find something with better QC and QA. After 30+ years using SunOS

OI's uname -a output notwithstanding, the release of the last OS to call
itself SunOS "on the box" was 26 years ago
<https://en.wikipedia.org/wiki/SunOS>.

I should like to bring OI to the state that it has a stable, tested branch.
> I have no objection to alphas, but I don't consider alphas as the only
> option acceptable.
>
> The hallmark of SunOS has always been robustness and reliability.

OI's rolling release is robust and reliable in the sense that stuff that
works in 1 BE is extremely likely to work in the next BE you boot into
after a pkg upgrade and packages don't just vanish from your installation
because they failed to build on the repo servers (I've had that happen with
other rolling release distros). To OI's credit, the only *loss* of
functionality I've experienced was Time Slider throwing an error code when
I tried to launch it. A quick pkg upgrade and reboot fixed that problem. In
today's security environment (monthly, at the very least) kernel patch
reboots are a necessary evil for every OS in existence anyway, so I
wouldn't call that a disadvantage.

If that is gone, there is nothing left of interest other than the corpse.

I wouldn't be so bleak :) I refer to my "religious artifact" comment a few
emails back. OSes and distributions are tools, not sacred objects. Tools
change, age, fall out of popularity, etc. So it has been since the dawn of
humanity. We all find ourselves having to align our expectations to the
tools (and the support for them) available.

Personally I find the Solaris(h) way of doing things fascinating, and
Solaris' excellent documentation, while not always applicable, is also
helping me understand functionality on other distros that aren't as well
documented. I do get frustrated. A lot. But I get up, walk away, work on
something else, circle back. However I do acknowledge that my
"explorational" use case is different from your desired one.

And I prefer to avoid those.
>
> Reg
> _______________________________________________
> openindiana-discuss mailing list
> openindiana-discuss at openindiana.org
> https://openindiana.org/mailman/listinfo/openindiana-discuss
>