[nycbug-talk] Sun Acquires MySQL
carton at Ivy.NET
Thu Jan 17 18:03:27 EST 2008
>>>>> "jv" == Jonathan Vanasco <nycbug-list at 2xlp.com> writes:
>>>>> "hg" == H G <tekronis at gmail.com> writes:
>>>>> "gr" == George Rosamond <george at ceetonetechnology.com> writes:
jv> From what I've been able to glean, Sun has pretty-much-been
jv> one of the largest proponents of PostgreSQL for a long time -
jv> and i've heard that they're one of the larger $ backers for
jv> development (though I can't find any real docs on that). I'm
jv> mostly worried about what this means for Postgres, as I prefer
jv> that to MySQL by quite a bit. If they lose clout or bonded
jv> development, that would suck.
^^^ what he said, exactly.
including the clout part. When Sun started releasing their own
Postgres port and announcing their intent to do some Postgres/Solaris
performance optimization, I thought, ``great---at least one clan of
wizards has stuck their staffs in the mud to say, `people who know the
difference use Postgres.' ''
Instantly after the MySQL purchase, Postgres lost massive clout.
hg> "stacked solutions" to their clients, instead of simply
if mysql makes its way into some stacked solution from Sun, Postgres
will lose huge amounts of further clout.
gr> it's not like the GPL'd license is going to change when they
gr> purchase it.
that's the thing that sucks about the buy-your-way-out-of-the-GPL
stuff---they can if they want to because they've retained single
copyright ownership. I'm much more inclined to spend my time working
on someone else's code if it's under FSF ownership. but yeah I think
MySQL's future licensing is only more secure by Sun owning it.
hg> MySQL, whether its better or worse than the alternatives,
hg> undoubtedly has a massive footprint in both the free and
hg> commercial domain.
It is crap, and using it may or may not cause problems for you, but
it's guaranteed to make you stupider, like learning BASIC.
But if you want to use this huge body widely-available
extremely-useful PHP crap, or tap into the pool of cheap PHP
developers, there's no choice, is there? As my old housemate used to
say, start swimming downstream.
gr> I need to read the Sun site a bit more thoroughly about their
gr> licensing, particularly with OpenSolaris, before I'd call it
gr> open source though. Most references note that their open
gr> source licensing isn't exactly that.
the CDDL is fine, even according to Stallman. The SCSL that Java used
to be under is the problem one.
The problem with OpenSolaris is that it's not ALL open-source, and
they're being deliberately misleading with the way they name things.
The realistic way to use it and do development on it is to install the
latest SXCE DVD, and then overwrite parts of it with OpenSolaris. And
there used to be this long explanation about how OpenSolaris isn't
actually a complete operating system---it's the name for a tarball of
source releases through a ``gate'', limited to ``core OS and
networking components,'' which can overwrite parts of Solaris Express
People kept getting confused and thinking Solaris became OpenSolaris
when Sun changed the license. About a year ago Sun just quit
correcting them. Since the error is in their favour, why bother.
from OpenSolaris source you can't:
* build your own install CD's like you can with BSD or Linux. The
only supported way to install is to download their DVD image,
install that, do a nightly build of OpenSolaris under that binary
environment, then use BFU to install what you built over top of
* check out the source for released/supported Solaris 10 Update 4
from Mercurial, change one line, rebuild, and BFU over your S10U4
system. The stable branches are not open-source AFAIK. The
binaries built from them are free-as-in-free-beer, but if you run
stable Solaris you can't get the source code for what you're
* build any kind of bootable operating system, make any forked
``Distribution,'' that includes open-source software only. Even
Nexenta includes lots of Sun binary bits. It's certainly not the
outdated throw-over-the-wall crap that Apple pulls---it's possible
to collaborate with them given what they realease---but it's still
missing key chunks that probably won't get magically replaced by
some industrious Linuxmonkey ever.
* buy any complete system from Sun that includes source code for all
the drivers. It takes some care to do this with BSD or Linux, but
(modulo 3D acceleration) is very doable.
There are three licenses:
CDDL: the open-source part of OpenSolaris that's in Mercurial
OpenSolaris Binary License: the binary stuff that's in Nexenta.
<forget the name>: the (binary) license of Solaris Express. You have
to register on their web forum and legally declare the number of
(free) copies you're using to download stable Solaris or Solaris
Express DVD's. They keep tweaking this one, but the license gives
them the sort of sneaky problematic control they had over Java.
as for Sun's contribution to open source history, though, Solaris is
huge and of high quality, and although the release is not as complete
as I'd like it's possible to get work done over there unlike on
Darwin, so even without OpenOffice and Java they are a big player.
I'm a little pessimistic that grumbling and agitation by individual
developers will ever lead to a more complete OpenSolaris release,
because anyone who matters and is doing enough work to get blocked by
missing source code can probably get a job there if he wants one.
unless they don't pay enough or one of their competitors agitates for
a more complete release, I don't think we'll get one.
And their claims of ``all _new_ work we're doin gon Solaris will be
open source'' are bogus because they do not apply to drivers. The
claim really should apply to drivers since unlike Linux zealots, they
are designing mass-produced motherboards and spec'ing the chips that
go into them. They hold all the cards and still deal crappy hands.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 304 bytes
Desc: not available
More information about the talk