John Plocher wrote:
> Garrett D'Amore wrote:
>> Alan Coopersmith wrote:
>>> All software produced by the OpenSolaris Community shall be
>>> licensed to the public free of charge under one or more open
>>> source licenses approved by the Open Source Initiative.
>>>
>> By that token, Indiana can never meet that requirement. There are
>> closed source components in it, that are not likely to *ever* be
>> opened up, and for which it is not realistic to find a suitable open
>> source replacement.
>>
>> Removal of those closed source bits (or replacement with open source)
>> will cripple the distribution on certain types of hardware,
>
> That may be so, but I strongly doubt that any of the interfaces exposed
> by those closed-source drivers (etc) are in any way visible as part of
> an "OpenSolaris ABI for use by ISVs".
>
> That is, Oracle, QuickBooks, Acroread, Eclipse, NetBeans and MySQL
> don't give a hoot about the NVidia driver...
>
> Yes, you may need closed things to make an "interesting" or "high
> performance"
> distro on some platforms; No, you should not be required to have them
> just to
> be able to run applications found in an OpenSolaris-compatible repository.
So, this goes *right* to the heart of what I was saying.
That is, WHAT IS THE PURPOSE OF A REFERENCE DISTRIBUTION?
I think there are two camps here, and they are at least somewhat at odds.
Camp #1 believes that a "reference distribution" should be the open
source core components, which define the "platform". They believe that
ISVs should be able to target the "reference" and pretty much be assured
that it will work on all "compliant" distributions (FSVO "compliant".)
This seems to me much like what LSB, SVID, and other standards try to
achieve. (This, I guess, would mesh with the Solaris binary
compatibility guarantee.)
Camp #2 believes that a "reference distribution" should be the "default"
distribution that users download. That is to say, they believe it
should be stable, functional for the vast majority of users, and that it
should be what we point users to when they want to "download
OpenSolaris" or "just install OpenSolaris to try it out".
Resolving these different view points will be necessary before we can
even begin deciding whether a given collection of bits is suitable as a
"reference" distribution or not.
It is my belief that the community most urgently needs the product that
camp #2 wants. But it is also my belief that such a product is *not*
what I would call a "reference" distribution, but perhaps instead
"default" or "mainstream" distribution.
Further, it is my belief that what camp #1 needs is not really a
"distribution" so much as two other things... first a common core
base. (We have that today in the code ... the base does does not need
to be a binary deliverable.) Second, a set compatibility/qualification
criteria and supporting verification tests. Some of this exists, but
much of its not present.
Now, all that said, I think the community has kind of gotten a bit out
of touch here. There are some folks who feel, perhaps vehemently so,
that a perfectly level playing field is absolutely required for OpenSolaris.
Let me state without reservation that I believe such an attitude is not
in touch with reality. Right now there is only one major corporate
contributor to OpenSolaris. All the other contributions taken
*together*, barely amount to a drop of water in the ocean of
contributions that Sun has made (and continues to make!) to
OpenSolaris. Somehow, ignoring that reality, and saying that Sun
should have no more privilege than any other company seems a bit
quixotic to me.
Also, I content, that at this time, anyone who believes that the
community of both current and future users of OpenSolaris is going to be
more interested in any distribution that is *not* Sun-branded (and hence
also would lack all the support guarantees, etc. that come with the
Sun-branded distribution) is living in some other dimension... perhaps
the dimension where BeOS recently replaced OS/2 as the worlds most
popular desktop operating system, and where SGI purchased IBM in a
leveraged buy-out....)
So, I believe that it is in the interest of both Sun and the community
to have Indiana (or whatever it spawns after ARC'ing, etc.) become known
as "Sun OpenSolaris". This leverages both the Sun and and OpenSolaris
names.
The name (or names) of any other distributions are, quite frankly,
inconsequential in the grand scheme of things. (To those of you who
have worked on those other distributions, I'm sorry if you feel
disenfranchised by this statement. But I still believe this is the
state of the world today, and I don't see it as likely to change anytime
in the reasonably near future.)
-- Garrett
