Slashdot Mirror
Sun Joins the Free Software Foundation
RLiegh writes "Ars Technica reports that Sun has joined the FSF Corporate Patron program. The article explains that the FSF corporate program allows companies to provide financial assistance to the FSF in return for license consulting services. The article goes on to observe that this move is doubtlessly motivated by Sun's interest in GPL3's direction. Now that Sun has opened up Java and become an FSF corporate sponsor...could the move to dual license OpenSolaris under the GPL3 be far behind?"
Free Solaris for everyone!
GetOuttaMySpace - The Anti-Social Network
Is Linux has a new and very adept competitor. Solaris has some GNU pains, but they won't last long, and underneath the hood is some amazing work.. It is just just ZFS, and DTRACE either, just take a look at the main page for ifconfig on Solaris vs other systems. There is a lot of depth to Solaris that will start coming out, esp on SMP systems, but on any system really.. The great thing is, Linux will have Solaris to learn from now..
...or is Solaris looking more and more like just another Linux distro every day?
What happened to diversity?!
was to get rid of Mcnealy. I am betting that Sun will be back quite a bit stronger in about 2-3 years time. It sounds like the new CEO is not wanting to play games esp. with the OSS world.
I prefer the "u" in honour as it seems to be missing these days.
See subject.
If this is true, how come I can't ``apt-get install java'' and get the SUn Java on Debian default install?
Everything under the Sun must go!
"The article goes on to observe that this move is doubtlessly motivated by Sun's interest in GPL3's direction."
Or perhaps it was motivated by Sun's desire to buy their way into the "free" software community's good graces without fully embracing its approach. Remember how happy everyone is about IBM's contributions even though all their profitable software remains proprietary.
I just felt a tremendous disturbance in the Force. It was if millions of slashbots cried out in pain as their heads asploded.
With apologies to the late Sir Alec.
Stick Men
I'm saying that they didn't convert any of their profitable software to F/OSS.
Thank you Sun! Thank you RMS! Everyone is winning now.
I haven't contributed any free software.
My point is that you should beware of large corporations bearing gifts.
I've experienced Solaris and its predecessors from the early 80's. Their kernels used to crash from memory leaks, corrupt data, contain Trojans, use linear search in inappropriate places, crash on bad system call arguments, fold under load, and lots of other problems. It's good that after 20 years, they finally got most of the bugs out, but it's never going to be "amazing work". From a practical point of view, Linux has matured much faster, and I don't think Linux has anything to learn from Solaris.
In the end, the differences between the current crop of UNIX-like kernels won't matter much. All of them have roughly the same functionality, most of them are fairly mature and stable, and all of them give you performance close to machine. And under the hood, they're all ugly and messy.
So, personally, I'm sticking with Linux. Solaris might be slightly "better", but not in a way that matters, and far more people are contributing to Linux (in particular, drivers).
As soon as GPLv3 comes out, GNU/Solaris will have both kernel
and GNU userland licensed the same way.
This will make it more interesting than GNU/Linux from this point of view.
And Solaris is technically a great kernel btw.
"Most companies hold onto the source of their software far after it's making them any profits. Most companies don't even release their source even after they stop distributing it."
That's exactly what IBM did with Rational's Visual Test after they bought Rational. Not only won't they release the source, they won't even let you buy a EULA for it. Why? Because it competes with other testing tools they sell that are a lot more expensive.
Sun has promised that they will open up Java in Java 7, but current versions are still not fully available under an open source license. For better or for worse, I expect that Java 7 will be incorporated into Debian and other distros.
I would recommend against making any plans that depend on Sun actually delivering a fully open source Java implementation: even assuming Sun is being completely honest, there are still things beyond their control that might jeopardize it.
Honestly, even if Sun puts Solaris under the GPL (2 OR 3), it's not automatically going to be a serious contender to Linux.
I've worked on the Solaris O.S.. I've worked on Linux (totally the latter these past 8 years). I've put code into GPL'd projects and have worked on projects that require not only GPL licensing but that you transfer the copyright to the FSF. Here's my view.
One of the single biggest reasons why Linux is so popular among kernel developers is that they get to keep their own copyrights. It doesn't have to be transferred to the FSF, Sun, or anyone else. This is also the single biggest reason why no one is interested in doing work for the FSF's Hurd. There's a real distinctive pleasure in saying that you own the copyright to a certain piece of code. Forcing someone to transfer that is like ripping something away.
I work for my own pleasure. I don't work for the FSF, nor Sun, and I'll be d*mned if they end up owning my own code without paying me first.
My bet is that Sun is going to absolutely botch this effort, and force people to transfer all copyrights to either Sun (like they do now) or to the FSF (as RMS has wanted for years).
If they do that, Solaris will never catch up to Linux in popularity. There will always be more developers for Linux than Solaris, just like there is now.
Which, frankly, is a real pity. Linus has screwed up in a number of critical ways, and although Linux has come far, Solaris is still better in many aspects (if not most). I'll spare you my biggest nits, which cause me solid grief; ones I wouldn't have if I went with Solaris.
The bottom line though, is if Sun doesn't insist on the copyright, I'll drop Linux immediately. And I'll work my hardest to promote Solaris (especially if they are GPLv3). But if Sun and the FSF think they can use my talent and then take my code away, they can go piss off.
I really, REALLY hope Sun makes the right decision here. But from past experience, I would be surprised if they did.
[captcha: "creating". How appropriate]
I've experienced Solaris and its predecessors from the early 80's. Their kernels used to crash
You have experience with Solaris but don't realize that Solaris is based on a different code base than predecessors from the early 80's? Solaris is built upon SVR4 while SunOS 4.x and before were based on BSD.
The reason why Solaris was the OS of the dot com era was because is was so reliable. At the Brokerage firms I've worked at you always see Linux crash or hang and Solaris just keeps on running. That's been my experience.
And remember Solaris was designed from the beginning to support SMP, threading, and soft real-time. Things that Linux only later had hacked on (and soft real-time is still not part of Linux).
Solaris 10 is so far ahead of Linux that it's not even worth comparing the two but if you must just look at these New features.
I hope this extends to their Sun Ray line. I've been looking a various thin client solutions for a company a couple of friends and I are planning. I really like their thin client platform, but the custom protocol and closed server app (though it runs on both linux and solaris) give me pause. For example, I would like the flexibility to boot a PC as a Sun Ray client - not that I'd want to on a permanent basis, but in a pinch - say until I could order more Sun Ray terminals - it would be nice. With more open standards out there, I'm stuck between choosing the solution I really want - and tying myself to one provider - and one that leaves me the freedom of choice and the ability to support the platform whether or not the company is in business.
Leben Sie jetzt die Fragen.
IBM is currently involved in hundreds of open source projects, ranging from small stuff all the way up to stuff like Xen, Eclipse, and a huge amount of code for the Linux kernel. Claiming that IBM contributes nothing of value to open source is an outright lie.
I've heard some very strong opinions on Linus wanting Linux to remain gplv2. Some even suggest replacing Linux with Solaris. What I'm wondering is the same companies that helped Linux become what it is today make the transition? Open source is rarely run on kindness and love for humanity. Linux has reached critical mass, and I don't think GPL'ing even a product as good as solaris will derail it.
Both Solaris and Linux would benefit immensely from sharing with each other. But whos ever heard of two competing products helping each other.
You have experience with Solaris but don't realize that Solaris is based on a different code base than predecessors from the early 80's?
You misread that. What I was saying was that I have experience with Solaris starting from the early 80's, until now.
Solaris is built upon SVR4 while SunOS 4.x and before were based on BSD.
Indeed. And while SVR4 was slightly less buggy, it was a worse OS than BSD. The BSD/SVR4 switch was when many people (myself included) started seriously exploring alternatives.
Solaris 10 is so far ahead of Linux that it's not even worth comparing the two but if you must just look at these New features.
Quite right: Solaris has even more useless crap in it than Linux. That's a good reason to stay away. See, more features is generally not a good thing in a kernel or OS.
Remember how happy everyone is about IBM's contributions even though all their profitable software remains proprietary.
People are happy with IBM's contributions because IBM actually makes contributions. That is, they contribute stuff to existing open source projects without demanding control over those projects.
Sun isn't doing that; both Solaris and Java are going to be dual licensed and controlled by Sun. That means that while the code happens to be released under a nominally open source license, the projects are not run as open source projects, and the exchange of code doesn't work for them as it would be in an open source project.
On balance, it's still good for Solaris and Java to be released under an open source license, but there are good reasons to be a whole lot happier with IBM than with Sun.
and- promo7es our
I have worked with Sun products since 1994. When I first started using them they were
great, and the processing capabilities and hardware reliability were superb.
But since 2002, there has been a significant decline in their hardware offerings.
They're selling products like the V1280 to markets that have no place for them.
Why would I put my very large OLTP database with 20+ TB of data on a V1280 cluster
with just 5 PCI slots? Yeah the 12 CPUs is great, but I cannot even drive the I/O.
Yet Sun sales reps continued to push them at the time. I'm reminded of this when they tell me
to buy now the miracle T-2000 that will solve all my worldly problems.
Then, there's the mysterious cover-ups behind the fact that V1280s' system boards
are very faulty. The answers I get from Sun regarding hardware replacements are,
"Well our Parts Depots are fully stocked, so we probably don't have a problem"
Then there's the Fujitsu factor. They're supposed to be partners and sell each others
equipment, but do you ever hear a Sun sales rep say, "yeah you should probably get a Primepower
650 for that, instead of a V440", or you'll hear "Yeah T2000 is better than a Primepower 850,
you should get that"
Sun is using that relationship to get them to their next processor level. But I'm not holding
my breath for that. Because it will be another seemingly endless line of non-scalable "you should
buy 10 of these and cluster them" solutions.
As for me, I'm going Big Iron. And consolidating all of my stuff. I leave the name of the platform
out to protect the innocent. Hint: It doesn't use containers or zones, and it's been doing virtualization
long before Sun or VMWare.
I like Java, it's a good programming language full of possibilities. My advice, stick to software Sun.
Avg. Live Expectancy of a SysAdmin, 45 Years.
What I love is you avoid even commenting on the below because you know Solaris is more reliable and a better engineered kernel than Linux. Not hard to believe when Sun spent 500 million on Solaris 10 and have the best kernel developers in the world working on it AS A REAL JOB not part time hackers.
If Linux is so great than why do companies like Morgan Stanley or Goldman Sachs have linux crashing or hanging while Solaris just keeps on running? The ONLY reason people run Linux in the Enterprise is because until recently it was the ONLY OS other than Windows than ran on x86. Not that it's so great. No that it's free (RedHat is anything but FREE). It's that it runs on cheap AMD/Intel hardware. Of course Sun now fully supports Solaris 10 x86 so yet another reason to ditch Linux. I'd take Linux on the desktop over Solaris as Solaris is clearly a server OS but then I'd take Mac OS X over Linux hands down on the desktop.
* The reason why Solaris was the OS of the dot com era was because is was so reliable. At the Brokerage firms I've worked at you always see Linux crash or hang and Solaris just keeps on running. That's been my experience.
* And remember Solaris was designed from the beginning to support SMP, threading, and soft real-time. Things that Linux only later had hacked on (and soft real-time is still not part of Linux).
If you really think Linux is so great maybe you could give some examples of what makes Linux better than Solaris or Mac OS X? I've been around UNIX for almost as long as you and was using Linux before the 1.0 days.
"Sun ... have the best kernel developers in the world"
This is far from the truth.
What I love is you avoid even commenting on the below because you know Solaris is more reliable and a better engineered kernel than Linux.
Because even if it were true, it wouldn't matter.
Not hard to believe when Sun spent 500 million on Solaris 10
Yeah, too much; it's basically an Edsel.
and have the best kernel developers in the world working on it
And what evidence is there for that, other than unfounded claims about Solaris quality? Your reasoning is circular.
AS A REAL JOB not part time hackers.
Most Linux development is done by people who do it as their job.
The reason why Solaris was the OS of the dot com era was because is was so reliable.
Don't try to rewrite history. I was there, and I was one of the people who picked Solaris for dot com companies. People picked it because they knew it, and they knew it because 5-10 years earlier they were using it at university. And they were using it at university because it was cheap. Other than that, it was merely "reliable enough". If reliability had been the primary consideration, people would have picked AIX or Irix, both of which were generally believed to be superior to Solaris (a lot of their technologies and code have made it into Linux, incidentally).
And that's why people pick Linux: it's widely used, its development is open, and it gets the job done; that's all that matters.
And remember Solaris was designed from the beginning to support SMP, threading, and soft real-time.
Bullshit. Solaris wasn't designed at all, it evolved out of SVR3, BSD, and SunOS, and each of those all evolved from the original V7 UNIX. Trying to portray Solaris as the herculean design and implementation effort of some elite group of kernel hackers at Sun simply has no basis in reality.
If you really think Linux is so great maybe you could give some examples of what makes Linux better than Solaris or Mac OS X?
It's "great" in the same sense that a Ford Escort is a better car compared to a Ford Edsel.
a couple of weeks ago.
I'm still waiting for my install set though. Did anyone actually receive their discs?
F_T
This is so true.
Solaris is outstanding technology.
What drove Linux's success was its x86 support, freeness, openness and the ability for anyone to innovate with it (thanks to GPL). These enabled Linux to grow.
However, Linux has its issues.
A big one that's not obvious to the casual user is backwards compatibility.
ABI backwards compatibility in Solaris is simply outstanding; take old Solaris 2.5 code and run it on Solaris 10 and it will most likely just work without rebuilding it.
They even have little tools to warn you what might not work *before* you trash your old OS image, and they work on any binary.
You come to expect that when you use Solaris a lot, and then you get a rude shock when you find Linux hasn't got the same level of "just works".
I think GNU/Solaris could shape up to be really stiff competition for GNU/Linux (GNU wins regardless).
Let's not forget just how gutsy and forward looking Sun are being yet again; they are the biggest corporate contributer to OSS by far. And it's good stuff.
IFF Solaris goes GPL it would really shake things up.
In fact, if they did, why would Apple need Darwin????? After all they just borrowed ZFS and DTrace... why not borrow the lot?
Why wouldn't mobile phones use Solaris (not Linux), especially because they ALREADY use Java -- combining the too would be a great way for Sun to drive adoption.
Interesting possibilities everywhere.
Let the birds out of the cage, and the Linux community will absorb the best parts into Linux.
Vista is proof that closed source is a bad idea, a big let down compared to Windows XP Pro.
(and that isn't saying much...)
I was going to convert a financial site from PHP5 to WebObjects (Java JBoss) because of the soon to be widely known bugs in PHP5 thanks to some jackass who wants his 15 minutes of fame; however, open source Java is not good enough if Sun is going to join a bunch of communists in New England who evolved out of a self-serving academic beuracracy and now cater to Fidel Castro and others. I'm going to look at some solution involving Mono or PERL-based SOAP responder until the self serving jackass' "hardened PHP" project gets airborne. Also going to trash the idea of buying a SunRay 2 this week and instead go with cheap PCs that boot off flash-drives. What a complete bunch of morons, adopting GPL3 and giving money to fuckhead RMS.
.... become the package mantainer and do it.
Otherwise, frankly STFU, and wait like the rest of us, greateful that there are actually people doing instead of whinning.
IANAL but write like a drunk one.
http://www.google.com/search?hl=en&safe=active&q=p redecessor&btnG=Search
How can you call bullshit on a statement so vague that it doesn't say anything?!
Look, if Solaris and java are released under the GPL, if you disagree with what they are doing, you can actually for the project and create your own.
How is that for openess?
I wish people would stop the unnecessary whinning when the licenses used take good care of that.
Once Sun, or any company, GPL stuff, they will remain in control only as long as they continue doing good, sane decisions. The day they stop doing so the community goes and plays elsewhere.
What else do you foxy need?
IANAL but write like a drunk one.
Apple needs Darwin so it has something it has legal control over so it can keep as much (or as little) "open" as possible. Hacks like the Don't_Steal_MacOSX.kext thing wouldn't be viable if Apple had to release the source code to it under the GPL.
I do agree with your wider point though, I can see Solaris being the base of a more unified set of *ix operating systems in the future.
You are not alone. This is not normal. None of this is normal.
Not only did I qualify my statement with "and its predecessors", you're actually wrong, since Sun retroactively renamed SunOS to "Solaris".
Really, I have been using Sun operating systems since the early 1980's, and the notion that Solaris is some kind of super operating system designed by highly skilled developers, or that Sun shows how much better commercial development is than open source, makes me laugh. It's nice that they finally got most of the serious bugs out; it sure took long enough. And the only reason Sun managed to survive the first 15 years was because open source projects like GNU and X11 made the platform semi-open and made up for the crap Sun was shipping.