[OpenIndiana-discuss] Solaris 11 Express

[Richard L. Hamilton]

On Nov 2, 2010, at 3:05 AM, Edward Martinez wrote:

> 
> On 11/01/10 09:32, Apostolos Syropoulos wrote:
>>> this discussion is pointless - imo.
>> 
>> I totally agree. IMHO, we should not care about what happens
>> to Oracle.
>> 
>       I  second that. Oracle did not really care about the Opensolaris community, so be it. if  it was not
>     for the great man Garret D'amore for starting  illumos, and the openindiana team releasing a great OS, then all  of non -oracle employess, like, myself, would of been screwed, since they have access to
> new solaris developer builds  and source code, we dont.   now lets wait and see
>          if oracle does indeed release the new  solaris source code, since they kick us out from reading and downloading it.

Well...they said they would.  Of course they've said a thing or
two that they didn't follow through on (like that there'd be at
least one more OpenSolaris (distro) release after 2009.06).
I kind of think that there was always the implication though that
after that, they'd be transitioning to getting a commercial release
ready, and not necessarily bothered much more about keeping the
open distro going.

The code is another story.  But I do understand that they want the
money, or at least the publicity, to flow first to them before everyone
else is doing stuff with it.  If it takes them 6 to 8 months after
they have most other things done to get the new IPS installation and
upgrade functionality ready for both small and large-scale deployment,
that would be 6 to 8 months that Nexenta would be ahead of commercial
Solaris.  I can see how they might not like that.

The problem I have with it is not so much that I need my bi-weekly
fix of free updates, but rather that, except for those already
engaged in participation (the ksh93 effort, and presumably Intel
hardware support, and maybe one or two others, but not many),
not seeing most of the code and ARC cases as they happen effectively
means not being able to participate even to the point of making
reasonable suggestions while things are early enough along that
they might be adopted readily.

OTOH, there is the possibility of one thing I _would_ like still
remaining in what they've said.  What I've always wanted was to have
code that actually matched what I was running (preferably without
going through the hell of building the OS myself).  If they're only
going to release code after they've released binaries, well..."after"
could be a short time.  Having code that matches what one's running
has always been helpful for troubleshooting, or clarifying what the
documentation implied but didn't clearly state.  It's more important
with dtrace, so that one can design smart D scripts for situations
where there isn't a stable provider (how the heck are you supposed
to know what to do with FBT probes _without_ source?), or where
you think you may have already seen enough to isolate a problem.

There would be value to the vendor in that too - imagine if a nontrivial
portion of bug reports included enough troubleshooting to identify
the code at fault.  That could speed fixes and cut costs considerably.

So while I'm quite disappointed in what Oracle has done thus far,
and am glad that enough talent is now outside Oracle and still
interested in working on the code base, remember, that's still
a small part of the talent that remains inside - by no means
all of which necessarily is thrilled with the present arrangement
either, but presumably they'd rather stay than leave over it.

That leaves me with wait-and-see whether they'll do what they
said they would this time...