Post Syndicated from Bradley M. Kuhn original http://ebb.org/bkuhn/blog/2010/03/03/jeremy-on-sun.html

Leslie Hawthorn referred
me to
an excellent
article by Jeremy Allison about Sun merging with Oracle. It was a
particularly interesting read for me since, while I knew that Jeremy
worked for Sun early in his career, I didn’t realize that he started in
engineering tech support.

The most amusing part to me is that it’s quite possible Jeremy was on
the UK tech support hotline during the same time frame when I was
calling USA Sun tech support while working for Westinghouse. I probably
would have had a different view of proprietary software if Jeremy had
answered the USA phone calls. One of the major life experiences that
led me down the path of hard-core software freedom beliefs were my many
calls to Sun tech support, who would usually tell me they just weren’t
going to fix the bugs I was reporting because Westinghouse just wasn’t
“big enough” (it was ironically one of the largest employers
in Maryland in the 1980s and early 1990s) to demand that Sun fix such
bugs (notwithstanding our monthly Sun maintenance fees).

But, more fascinating still is Jeremy’s analysis of why Sun failed as a
FLOSS
company. Specifically, Jeremy points out that the need for corporate
control over all software technologies that Sun released, specifically
demanding the exclusive right to proprietarize non-Sun contributions,
was a primary reason that Sun just never succeeded as a FLOSS
company.

Meanwhile, I’m less optimistic than Jeremy on the future of Oracle. I
have paid attention to Oracle’s contributions
to btrfs
in light of recent events. Amusingly, btfs exists in no small part
because ZFS was never licensed correctly and never turned into a truly
community-oriented project. While the two projects don’t have identical
goals, they
are similar enough that it seems unlikely btrfs would exist if Sun
had endeavored to become a real FLOSS contributor and shepherd ZFS into
Linux upstream using normal Linux community processes. It’s thus
strange to think that Oracle controls ZFS, even while it continues to
contribute to btrfs, in a normal, upstream way (i.e., collaborating
under the terms of GPLv2 with community developers and employees of
other companies such as Red Hat, HP, Intel, Novell, and
Fujitsu).

I have mostly considered Oracle’s contributions to btrfs (and to Xen,
to which they contribute to in much the same way) as a complete fluke.
Oracle is third only to Apple and Microsoft in its predatory,
proprietary software marketing practices and mistreatment of users.
Other than these notable exceptions, Oracle’s attitude generally matches
Sun’s long-ago roots (and Apple’s current attitude) in this regard:
non-copyleft FLOSS without giving contributions back is the best
“Open Source” plan.

Software corporations usually oscillate between treating users and
developers well and treating them poorly. Larger companies are often
completely self-contradictory on this issue across multiple
divisions. Microsoft and Apple are actually unique in their
consistency of anti-software-freedom attitudes; I’ve typically assessed
Oracle as roughly equivalent to the two of
them⁰. I don’t really
see Oracle’s predatory proprietary licensing models changing, and I
expect them to try to manipulate FLOSS to bolster their proprietary
licensing. Oracle was never an operating system company before the Sun
acquisition, and therefore contributing to operating system components
like btrfs and Xen were historically a non-issue. My pessimistic view
is that Oracle’s FLOSS involvement won’t go beyond what currently exists
(and I even find myself worrying if others can pick up the slack on btrfs if
(when?) Oracle starts marketing a proprietarized ZFS-based solution
instead). In short, I expect Oracle’s primary business will still be
anti-FLOSS. Nevertheless, I’ll try to quickly acknowledge it if it
turns out I’m wrong.

---

⁰ Contrary to the
popular receptions at the time, I was actually quite depressed both when,
in
1999, Oracle
announced first that they’d have a certified version of Oracle’s database
available for Red Hat Linux and when, in
2002, Oracle
announced so-called “Unbreakable” Linux. These moves were
not toward more software freedom, but rather to leverage the availability
of a software freedom operating system, GNU/Linux, to sell proprietary
licenses for Oracle databases. Neither event should have been heralded as
anything but negative for software freedom.