On Mon, 20 Apr 2009, Barnett, Jeffrey wrote:
> Joe, I read the slashdot article, and while it mentions at least two
> MySQL forks, it also makes clear that while forks *can* survive they
> clearly weaken the community.
I'd argue the "clearly weaken" statement. Have OpenBSD, NetBSD and
FreeBSD "clearly weakened" each other?
Have the many forks of Linux?
Honestly, we'll never know for sure, as it's impossible to test
empirically. However, you have to look at _why_ things fork do determine
if it's good or bad.
In some cases, it's with a clear goal in mind; eg, OpenBSD favors security
over all else; NetBSD favors cross-platform compatibility. PicoBSD was
just damned tiny.
As each of these projects test out different concepts, we can see what
works and what doesn't. Just look at Linux -- would it still have been as
strong if Slakware stayed on top, and we didn't have varients that took
the headaches out of package management?
The problems with forking is when it's done for personal reasons (eg, two
core developers can't get along, but it's not just a software philosophy
lines so other people don't know which group to follow), or when they
split the development community to the point where there aren't enough
people working on each of the branches to maintain it.
Sometimes forks merge back together, taking the best of each development
branch, and other times they remain distinct but share code between the
two (KHTML/WebKits come to mind).
Forks help the greater software ecosystem when they have clear goals in
mind, and when they realize that they're not their own isolated island --
they take the improvements from other forks that further the goals of
their fork.
Yes, they effectively compete against each other, but in a free market,
competition is good -- it forces them to differentiate themselves by
providing better features or services. (as obviously, they can't really
compete with each other on price).
...
All of that being said, the case that you specifically mentioned was in
Oracle trying to slow down the open source projects run by Sun:
>> I haven't heard
>> anything quotable recently from Larry Ellison, but he has in the pass
>> been an ardent opponent of OSS and I find it hard to imagine him not
>> taking advantage of this opportunity to place roadblocks and/or booby
>> traps in the way of the OSS community.
And in that case, the fork is obvious -- Oracle maintained with whatever
roadblocks, vs. a more open fork. There might be multiple forks at first,
but unless each one has a clear identity, they'll most likely coalesse
into one or two 'main' forks as alternatives to the Oracle-controlled
ones.
...
I don't know if you're seriously worried, or just trolling and mailing out
the same message to code4lib and oss4lib. My advice is not to be worried
-- all of the projects that you mention have a large enough community to
survive. Yes, there might be some rocky patches, but there are enough
groups who use them that I don't think they're going to be left to die.
At the very least, take a more constructive approach, rather than the
Chicken Little approach -- ask what it is that _you_ can do, or how you
can help to organize the library community to ensure that we can continue
depending on these tools.
-----
Joe Hourcle
Principal Software Engineer
Solar Data Analysis Center
Goddard Space Flight Center
[insert disclaimer about these being my personal opinions here]
|