[OpenIndiana-discuss] I figured this deserved a separate thread

[Christopher Chan]

On Thursday, November 04, 2010 10:11 PM, Jonathan Adams wrote:
> not that my $0.02 is worth much, but Illumos and OpenIndiana, although
> connected are separate entities.
>
> If you compare this to Linux, Illumos is the kernel and OpenIndiana is
> a distribution.
>
>
> If we were planning on having a stable, a development and a bleeding
> edge then that would be suitable in both Illumos _and_ Openindiana,
> but not one or the other.
>
> For the Kernel (Illumos) I would suggest that the current build might
> be considered development, the most recent previous development build
> be considered stable and the "bleeding edge" being an almost untested
> git clone and build.
>
> For the Distribution (OpenIndiana) I would suggest trying to follow
> the example of Ubuntu in having a regular build and maybe a LTS every
> 18 months to 2 years (in a similar way to Solaris10 update N)

Why do we have to follow the model used by Ubuntu or Fedora?

The way things work, Illumos being not just the kernel but also certain 
core user land stuff like shells and libraries point to what I feel 
would be a better arrangement. Perhaps the way things are might be a 
good thing with KDE being out of Openindiana at the moment. Desktop side 
of things are mostly composed of two things: shared libraries and 
applications. 'Long term support' can be done for these shared libraries 
without having to have the applications (which are just 'consumers') 
also under 'Long Term Support'. If there is an efficient and good 
approach to these shared libraries that do not qualify as system 
libraries, we can share the Q&A load and oversight while maintaining 
quality. It will mean that a bit more communication/coordination is 
necessary in certain areas but things on the edge (consumer software - 
software stacks/applications) can be upgraded faster without having to 
wait for a big release where everything has to be checked out.

Ubuntu is currently one big mess and their LTS release policies are 
laughable. Yeah, I have a stable version of Pidgin but it did not do 
Yahoo anymore on Hardy well before Lucid got released.

>
> These should in general be considered stable, and try not to upgrade
> the kernel, or major versions of apps unless security holes exist.

Apps are 'consumer' entities and should be upgradable any time. Kernels 
and core user land stuff are another story. I think it would be 
interesting to stuff non core stuff into their own zfs filesystem and 
make it possible to upgrade the boot environment at will without 
touching say the KDE env or the GNOME env or whatever. Much like how 
some stuff are now being stuffed in /opt which is necessarily affected 
by changes to the BE.


>
> I would suggest trying where possible to separate usability apps from
> core apps, with core apps admin being zealous of protecting the core
> apps repository (yes I know that is annoying to most dev's who want
> their files to be core) but it guarantees a core "server" style
> installation without forcing a gui or photo managers etc.

I see we see eye to eye on this.


>
> On top of the core apps and usability apps repositories, I would
> suggest another cutting edge repository with the latest versions of
> all apps (Mozilla/Adobe this means you) and worry less about trying to
> remain stable and known.

Yes, something different from the current way things are done with 
Ubuntu. Maybe pkg can have a registrar of approved addon repos that it 
can search and those repos can be managed apart from OpenIndiana.


>
> But this is not really anything to do with me.
>
> Jon
>
> On 4 November 2010 12:14, Nathan Evans<ndickinson.evans at gmail.com>  wrote:
>> Who cares who can sue who,
>> Well, the problem is here how do we market OpenIndiana and all this open
>> source software effectively, how do we grow the community in a positive
>> meaningful way. How to we make OpenIndiana&  Illumos viable alternatives and
>> brands that people can trust.
>>
>> The 33 people who left OpenOffice is a step toward disaster if we don't
>> build these brands what we'll have is another OpenVMS....
>>
>> We need to find a distribution model that works, and a software stack that
>> people can trust. The fact that we already have an AMAZING platform is a
>> better starting point then most have. We have what was once dubbed 'The most
>> advanced operating system in the world' we want to keep it that way.
>>
>> In my very humble opinion OpenIndiana should be used as our 'cutting/
>> bleeding edge' and Illumous as our stable production. When they eventually
>> both become that way. Similar to the old SXDE/Solaris model except
>> unrestricted. People dislike change, so lets follow a formula they are
>> currently familiar with.
>>
>> Once we have something going I'm sure a big company who will remain nameless
>> at this point will look at our work and sponsor our little community.
>>
>> Mind you this is just an idea I had, let me know what you think, I enjoy the
>> feed back.
>> _______________________________________________
>> OpenIndiana-discuss mailing list
>> OpenIndiana-discuss at openindiana.org
>> http://openindiana.org/mailman/listinfo/openindiana-discuss
>>
>
> _______________________________________________
> OpenIndiana-discuss mailing list
> OpenIndiana-discuss at openindiana.org
> http://openindiana.org/mailman/listinfo/openindiana-discuss