> Huh?
>
> Okay, let me put a different way. Pick a subject for your proposal, be
> concise, and stick with it.

The main point of the second draft was to capture the discussion so far.

> Right now you seem to be acting as if a software consolidation and a
> reference distribution are the same. They are not, at least not in the
> terms I understand that are relevant to this project.

 Well without a reference distribution to test with how can we test
the new standards we are working on. (Don't answer this.)

Additionally there has been no formal project proposal. The scope of
the discussion is evolving as issues are brought up in the threads of
discussion.

> > > I think most people think of "OpenSolaris" as the ON consolidation,
> > > since that really is the primary things around which all other
> > > consolidations center.
> >
> > Clearly not everyone agrees on this. You say a majority consider ON =
> > OpenSolaris. I think actually you would be hard pressed to prove this,
> > as the view of what OpenSolaris is probably depends on which community
> > the participant is active in.
>
> Regardless, I think it is important to state specifically what you
> mean when you talk about OpenSolaris, since by your own admission not
> everyone agrees on what that is.


> I think my point behind that is important because the definition of
> what it is seems to be constantly changing. When the project first
> started, there were relatively few consolidations available, and ON
> was the primary one (though if I remember right DTrace was the first).
> At that point in time, I feel it would have been accurate to say that
> ON was the primary representative of what people "saw" as
> "OpenSolaris."



> However, as time goes on, and more consolidations become available and
> more projects are started on OpenSolaris.org, what exactly is a person
> referring to when they say "OpenSolaris"?
>
> Are they only referring to the official SUN consolidations? Are they
> referring to all of the projects on OpenSolaris.org? Are they
> referring to Solaris Express releases? Are they referring to *all*
> OpenSolaris derivatives?

 Yes.

> Given that each one of these distributions has the potential to be
> wildly different eventually (and some very important differences exist
> already), I think it is important to state explicitly what you are
> referring to.

I deliberately have left it vague, in order to bring to light the
self-contradictions within OpenSolaris

> Everyone may disagree with me; I may even be the lone voice in the
> wilderness, yet my feeling is that right now the word "OpenSolaris"
> has a rather vague definition.

There are at least 5 different official definitions, and at least one
of them is not even defined. "OpenSolaris Technology" is referred to
in many places, but I could not find a definition.

> > > JDS, etc. are part of the OpenSolaris project, but would not think of
> > > them as "OpenSolaris."
> > >
> > > So, you need to define exactly what you mean when you say
> > > "OpenSolaris" -- there is a difference between the project and the
> > > code, etc.
> >
> > I lost you on this one.
>
> See above. The OpenSolaris project has many communities with many
> software projects, some of which are not part of any Solaris or
> OpenSolaris-based distribution yet. So when you refer to OpenSolaris,
> are you referring to the project, or the technology platform itself?

What do you mean by technology platform? That is pretty vague.

> > > That still doesn't define "weirdness." For example, are you saying you
> > > don't want to distribute what is in /usr/openwin or /usr/ucb? Or are
> > > you saying they shouldn't be part of the default path, and so on?
> >
> > Can we get rid of the deprecated bits of OpenSolaris, that are just
> > there so Solaris can maintain backwards compatibility? (Those bits
> > could be viewed as distro specific) My question does not ask the
> > feasibility or desire to do this, but rather asks, *CAN* the community
> > decide to do this and implement it even though it is not in Sun's
> > interests Sun's interests?
>
> Sure, the "community" can. Nothing is stopping anyone right now. I'm
> not sure why you think someone can't start their own distribution or
> do their own thing.
>
> If you're talking about a reference distribution that is somehow
> semi-official or the like, I think that would be a horrible thing to
> do and would deprive it of one of the most valuable assets that
> Solaris has --- backwards compatibility. Sure you could do it, but I
> personally wouldn't support it and most of the folks I know of around
> here wouldn't either.
>
> You also have failed to justify why you want to break it. Throwing it
> out simply for the sake of doing so makes no sense given the precious
> asset it is.
>
> You need to document which aspects of it on a case-by-case basis are
> problematic, and then deal with them one at a time.


> > > What does "no single log" directory specifically mean?
> >
> > /var/adm/log/, /var/adm/, /var/log, /var/samba/log, etc..
>
> That's a matter of personal preference, though I am sure someone could
> always make the proper proposal.

I think you are very much misunderstanding the point of my
"questions". These were all specific examples, of how OpenSolaris
(whatever that means) very closely resembles Solaris and the Solaris
development process. They aren't actual proposals.

>
> > > > >
> > > > > There isn't much non-core stuff there anyway. And I got the impression
> > > > > that the aim was to enable the easy supply of much more software
> > > >
> > > > Ok let me restate. Can we remove all code that was not developed by
> > > > Sun or the OpenSolaris community. (I am not advocating this, it is
> > > > just a point of discussion)
> > >
> > > No, because some of that code is needed for a "minimal" system. I
> > > think what you're trying to say is can we remove everything that is
> > > not the "core" necessary for the base OS. Or, put in a different way,
> > > can we choose to not include all consolidations that are not
> > > "absolutely necessary."
> >
> > It doesn't have to be required for a minimal system. We could just
> > refactor and remove those dependancies. (Again a distro specific
> > requirement)
> >
>
> What you suggest makes no sense because your qualification is overly
> broad. For example, you said "was not developed by Sun or the
> OpenSolaris community." That would mean ksh, X, GNOME, some drivers,
> and so on.

Yes, all of the above.

> You need to be far less broad in your qualification of what to remove.
> I think most people would agree that removing some drivers just
> because they weren't written by Sun would be silly :)

Again you don't understand the point of my questions.

Just to clarify it again for everyone. Over the past couple of weeks,
I was under the impression that Opensolaris is solaris, just done in a
public forum where anyone can contribute. When I expressed this view,
many people from sun and from outside Sun "corrected" me. They told me
Solaris was just a distro, and that decisions made by OpenSolaris
would not match the decisions of Solaris.

As time passed I can to realize that despite that belief, there are
many definitions of OpenSolaris floating around. Many of them are
contradictory. Some are named and not describes. It is all very
confusing.

I chose to present questions from the point of voice agreeing what had
been argued. The chief arguements were that Solaris is just a distro,
and that no decisions in OpenSOlaris should be made to benefit Solaris
over other distros. Clearly this is not the case. (Why else would you
need to maintain backwards compatibility unless it was to support
SOlaris users?)

Now I am getting arguments basically saying that I am wrong (again).
You know it's a little frustrating when both sides of a position are
"wrong".

-Brian

Reply via email to