Slashdot Mirror
Sun Mulling GPL for Solaris
comforteagle writes "According to this article in InfoWorld, Sun Microsystems is considering open sourcing Solaris by changing licenses to the GPL. What kind of impact would this have on those of you considering opting out of Unix for Linux? Red Hat and others have openly targeted Solaris users to switch." By the end of the article, the change seems rather unlikely to happen, but it's still interesting to see what changes this could bring about.
Would this inculde Re-GPLing the part they licensed from SCO?
I just posted this on an OpenBSD story, but it fits quite well here. I only use Linux because it's the easiest way to get myself a KDE desktop.
Really, if *BSD or a Free Solaris or anything else come up with live cd's or start-me-up installers, I might as well try them to test for performance and stability. Since KDE runs in any Unix-like system, "switching" is not quite a problem for me.
I just want the best desktop environment available today and that's KDE. What it's running on top of, I don't care.
Im not sure how this would affect the business world, but here at least it would most likely spread more understanding of *nix. Most of the apps we use here in classes, various Programming/Asic/Chip design programs, are only run on solaris boxes. If solaris were available for free, i have a feeling many students would install it on their system, just to more easily use these apps if for nothing else.
I though there was a lot of System V code in Solaris. How can SUN ever GPL that?
No way. They wont gpl java, but they'll gpl solaris? Highly doubt it.
http://wwws.sun.com/software/solaris/binaries/get. html
It's pretty fussy about hardware etc, though, and very obviously not the equal of Solaris/Sparc.
Okay, last time: 5 is May. I think you meant this to be on 4/1.
Getting sued by SCO?
KFG
1. Sun would have to replace all of the UNIX code. They can't put that under the GPL, period (unless SCO and Novell agree it's ok ;-)
2. Solaris includes many products that Sun has incorporated over the years. Most of them would likely have to be replaced, since I doubt the contracts involved allow Sun to just GPL the whole mess.
3. They would just be asking to have SCO add them to the list of companies targetted for a "tainting" suit, though honestly Sun may not care.
In the end, I think it would make far more sense for Sun to open source their SMP code by working with IBM on modifications to Linux. Sun+IBM could probably get Linux deployed on both of their very-high-end boxes in short order.
The SMP stuff is, as far as I know, most of what's left that Solaris does better than Linux, so what's the point in open sourcing the whole OS anyway?
I'm a Solaris user. OK, I'm a Red Hat Linux user too. But all of my important stuff happens on Solaris. It's just part of my reality at work.
I wouldn't mind if Solaris opened up. It wouldn't be a huge deal for me - I'd still pay Sun for "premium" support, and I'd still only use official Sun versions of things. Heck, I need someone big to blame if and when things go really wrong. I pay Sun to be that target.
I use Sun/Solaris because (1) I have the budget to, (2) it works, (3) I only have one vendor to deal with, and (4) there's no compelling reason to change right now.
If Sun can get something out of opening Solaris - great! If open source developers can improve the world by the opening of Solaris - great! But at least in terms of my current position, it won't have direct impact on me.
1. pretty nicely on expensive hardware, but I wouldn't choose it over Linux. (And in my job, I *do* basically have that choice.)
2. not really, no. The x86 version was available "free as in the cost of media" for a while, and it was a sad, sad joke.
Anyway, the article is pretty full of silly FUD, like this choice tidbit:
That's ridiculous of course, but more importantly, the only way in which it really makes sense is when you're still thinking about developing *propriatary* apps. If you can just recompile, pretty much *anything* that runs on one Linux distro will run on another. But if you're stuck in a shlepping-around-binaries mindset, yeah, there may be some difficulties.
Basically, they're still as clueless as ever. And we're certainly not going to see what Sun really needs -- an open source Java.
Because it works.
I think it's obvious what would happen if Solaris were released under the GPL.
They would be able to dip in to all of the device drivers that Linux has today. That would happen first.
What would happen second is the standard cross-polination; anything that's substantially better in either one (think scheduler, VM etc) would be copied from the better one to the worse, improving it. How much Linux might take we won't know until we've seen the code. But I'm certain that Sun would benefit the most from access to those device drivers.
It is about time *ALL* *nixes were reunified into one humongous tree, out of which various branches (which will have a common base) will exist for various uses.
We use AIX (AIX Isn't uniX) at work, but Big Blue keeps making noises about replacing it with Linux.
Too many flavors.
I do think this move is benefitial to all.
Please don't flame me for bad-mouthing Linux, I'm a diehard Linux admin myself but I still think Linux has much to catch up in enterprise computing.
We've a Linux cluster which has a critical bug in mounting the share disk which has filesystem. Sometime when one node down the mount point is not released to another node which is supposed to take up the process, thus result in critical failure.
This is all kernel problem(or limitation), and we don't have problem with non-fs type disk(raw disk). Therefore we must use raw disk where possible in cluster, but we don't have choice when some apps require a filesystem(e.g. like infracture database in Oracle's forsaken Real Application Cluster (RAC). Good name huh)
The engineer who diagnosis this problem told me they've no such problem with similar setup with solaris so they THOUGHT it's okay in Linux. Ahem, there goes millions dollars for paying their great product(*cough* Oracle RAC *cough*).
I expect more of such problems could be solved when those companies specialized in enterprise bringing back good stuffs to Linux, and GPL.
(ring ring) ..."
scott: "hello?"
bill: "what the fuck are you doing?!"
Scott (sweating, scared): "well I, err, I just, thought.
bill (slow calm and menacing): "yes?..."
scott (smiles): "I'm going to fire the fucking idiot who aired this dumb idea in public."
bill: "have a nice day."
scott: "yea, thanks bill. Thanks."
(ring ring)
scott: "have someone in marketing killed and his hands delivered to Bill Gates"
lackey thug: "It's done."
I think Sun would do better digging up some of their older code. NeWS, in particular... now that Apple has shown that you can be successful with a non-X UNIX GUI based on Postscript, Sun's own networked Postscript display system is ripe for a comeback. Remote desktop performance for a NeWS-based environment using current processors would be a killer, and they could incorporate Java as well as Postscript applets in the GUI.
He's quite right there, though.
(a) Businesses don't want more source code. Well, to be strictly accurate they don't want to have to manage, worry about, maintain, etc more source code than they have to. I think that given the choice, most intelligent CIOs would definitely say "sure, we'd love the source and rights to use it" - but would probably prefer not to have to actually do much with it, and face the burden of maintaining changes etc.
OSS alleviates this to some extent by permitting changes to be submitted back upstream, but this only works if you have the resources to engineer you changes "properly" and not break anything else (even stuff you don't use or care about).
(b) If you write for RH9 or RHEL3, your app will not run on a stable release of Debian. Not if it's a GUI app that uses any GNOME/KDE libs, needs a recent QT, etc. It can be made to run by either packaging it with a lot of extra libs (see Ximian's RH8 builds of Evolution for an example of this approach), spending a lot more time to make it handle varying versions of libraries, or forcing the user to update their distro or libraries themselves. None of these are attractive.
I see this as a real issue, but not a distro one. It's actually more about _versions_ - the rapid change of OSS, including APIs etc for major libraries and toolkits, is the root of the issue. OTOH, the same thing keeps "ugly" decisions from hanging around, and permits much more rapid advancement.
I'd like to see a cleaner way of running multiple versions of things in parallel (within the package management systems), as a work-around for this issue.
(c) Also quite correct. Many open source apps do not follow established standards, and often the file formats, protocols, etc are defined largely or entirely by the source code of the app. While these protocols/formats are definitely open, they're not open standards and there's usually not much chance that other apps will work with them.
It's true that you do have more chance of enchancing other apps to work with the formats/protocols, time and money permitting, or enhancing the OSS apps to work with the protocls/formats of your choice. It's also true to say that many apps don't support standard protocols or formats because there is no standard in that application domain, or it's crap. These things do not change the truth of his statement.
I can't see Solaris OS being majorly profitable for Sun either - they sell hardware too and if an open-source Solaris led to more end-user interest in their hardware it's easy to see it leading to an increase in revenue for Sun.
Essentially, Apple's strategy. And not really surprising, since Sun essentially does what Apple does--sell proprietary hardware with a tailored OS.
Question is: has that strategy paid off for Apple? And Sun has more to lose: they have a strong position in the server room, that Apple never had, so Sun would be trying Apple's consumer level strategy out but with their own Enterprise products--results may vary.
--
$tar -xvf
Now RMS can go on crusades to let the world know it's not Solaris, it's GNU/Solaris.
Karma: Non-Heinous
As others have mentioned, Sun can't GPL code they licensed. (Remember the first open source mozilla code?)
Thus we'd be given a nearly useless, incomplete operating system. If the Sun-owned Solaris code is truly GPL'ed, the Linux folks would pick all the good bits out of this carcass and discard the rest.
Thus nobody would use OSS Solaris, but Linux might be improved here and there. So, I highly doubt Sun will truly GPL their code.
(Apologies to Linus Torvalds for comparing him to a vulture.)
As opposed to the original AT&T UNIX license which thankfully prevented companies like Sun, HP, IBM, SCO, SGI, DEC, BSDi, etc. from forking...
Remember, they aren't pulling in huge $$$$ for licenses for their OS, the way Microsoft is. Sun sells hardware. They have been known to give away their OS at zero/low cost in the past - it's their hardware that makes them the money.
Linux is good, it is very good. But it is not as good as Solaris in a lot of situations - Solaris has been in high end trenches and mission critical situations for a lot longer than Linux. An open sourcing of Solaris under GPL means several things:
a) Linux can benefit from Solaris
b) Solaris can benefit from Linux
c) Extensive code review of Solaris by the world probably won't hurt efforts to further improve security.
TREMENDOUS positive PR for Sun from an often ambilivant open source community, and a rush to make sure all important open source software runs flawlessly on Solaris (harder to test now since fewer people use it)
Problems to be delt with:
a) Making sure they have the legal rights to open source everything (of course)
b) Export restrictions? Not sure how this plays out for Solaris - since Linux is out there already I can't imagine the use of restricting Solaris (which is probably also out there, just not legally) but the government is known for a lack of common sense in such cases.
c) Fear of management that giving up ultimate control of all versions of Solaris will somehow be harmful.
Issue a) was one of the major problems when considering opensourcing BeOS - don't know how Solaris stands on such an issue.
But I think on the whole it's silly for Sun to try and compete with Linux head to head with a commercial OS - what's the point? Sun sells hardware and complete solutions, and generally does very well. If they can say "well, Solaris is GPL just like Linux, incorporates features X,Y, and Z that users generally cite as reasons they want Linux, and is proven and stable to boot" they get to just support Solaris again, and not have to worry about figuring out Linux. If that makes Solaris more widespread, what harm does that do Sun? It's not like Microsoft is going to pick up Solaris and incorporate it. Infighting among Unix like systems I think is fairly pointless in this day and age. Linux has made high priced commercial Unix licenses non-viable. So for companies like Sun, who sell hardware and solutions anyway, why not go with the flow on this one?
"I object to doing things that computers can do." -- Olin Shivers, lispers.org
An explanation that I've found very effective over the years:
...
Way back in the 1980's, I worked on a number of projects that had teams building their software on a number of different kinds of small computer systems. The teams using Sun workstations often got a bit of flak for using a system with a fairly high price/performance ratio. But the Sun-based teams invariably had the last laugh.
What happened was that debugging would frequently lead into parts of "the system", i.e., system libraries and/or the kernel. When we asked the vendor for details of the low-level software, the answer would reduce to "We can't tell you; it's proprietary". The proprietary, closed-source parts of the systems were brick walls that blocked progress.
With Sun (SunOS or Solaris), if we couldn't get an immediate answer from Sun, we would just ask on one of the Sun newsgroups. Usually an answer would come back within hours, most often from an engineer within Sun. Very often, they would include a chunk of the source code as an explanation.
The result was that the teams developing on Suns would get answers to their technical questions, and would have a functioning product long before any of the other teams. There's a real advantage to having a working, marketable product, even if it's more expensive than a competitor that doesn't work yet.
Over the years, this Sun advantage has evaporated. It has slowly become more difficult to get accurate details on the inner working of Solaris and other Sun libraries and tools. They have gone the protective, proprietary route. And their market is slowly being eaten by linux, for exactly the same reasons as above.
It's possible that what is happening inside Sun is that the people who understand this are starting to be heard again. If they can make the innards of their system as open as it was 20 years ago, they stand a good chance of recovering their business.
Alternatively, if the protectionist factions inside Sun prevail, they could also start up a linux-based line. This would be a bit of an expense, but no more so than their switch from SunOS to Solaris (i.e., from BSD to Sys/V) 15 or so years ago. If they did this, a Sun linux would probably wipe out Solaris over a few years, for the same reasons of faster development times on an open system.
The cheapest would be to open-source Solaris. This would get them back into the good graces of software developers, and would restore their earlier status as a system on which you can bring a debugged, reliable product to market very quickly.
And it would probably be better for all of us, since it would avoid the growing threat of a linux "monoculture". The unix part of the industry has always been better off because it isn't a monoculture, and thus isn't susceptible to the virus/worm-type attacks of the "market leader".
But there are those elephants hiding in the middle of the room: patent and copyright. Can Sun legally open-source all of Solaris? If they try it, can they withstand the legal might of an SCO with behind-the-scene Microsoft support? Stay tuned
Those who do study history are doomed to stand helplessly by while everyone else repeats it.
Another thing Schwartz doesn't get: the possibility of forking is precisely what makes "Open source" open. The CIO does not want open source code because he wants his internal IT team to make a fork. He wants open source code because he wants the assurance that if Sun drops the ball technically or goes out of business (both possibilities) IBM or Red Hat or some mythical "Solaris Support Inc.) could pick up the Solaris ball and keep running with it. As long as it is proprietary to Sun, it gets bought by whoever buys Sun and it may be in their best interest to kill it. Mitchell Baker said it best: Open Source is about the freedom to choose leadership. Solaris users should be able to follow Sun's variant of Solaris as long as Sun continues to innovate and lead. But they should be able to follow someone else's variant if Sun starts to falter. This is all true for Java as well of course.
I don't think, that there is a big thread for Linux to become a monoculture or proriety. (This would be the thread, Jonathan Schwartz was bashed here a few days ago, right?).
Yeah, you're probably right. The linux landscape is the metaphoric "herd of cats".
But still, we should be aware of the potential problem, and we should discuss it. It's similar to how we shouldn't be too smug about the linux (and *BSD) security question. In both cases, we're muuch better off if we constantly harp on such problems, and point fingers at potential problems.
In the case of the monoculture, there is indeed a real potential for problems in the business arena. The business world has always favored a monoculture, as a way to simplify decision making (which can be costly in both money and careers). In the corporate linux market, RedHat has a strong lead, and there's a serious possibility that they could end up ruling the linux business world.
RedHat deserves a lot of credit and support for what they've done. But "winning" and vanquishing their competition could make them a target for the virus/worm plague that has infected the Windows user community. Granted, writing such software for linux is much more difficult than with Windows, but it's not impossible that a single distro would have an exploitable hole. Then we could see half the banking system or half the credit industry going offline simultaneously.
So we should be preaching to the business folks about the dangers of putting all their corporate eggs in the RedHat basket. We should teach them that part of the reason for all the Windows problems is the monoculture. They should intentionally use different distros, configure them differently, run different DBs, etc. They should look for ways to tailor their systems to their environment, so that they aren't too similar to other computers.
And we should be on the lookout for other such developments. We want a herd of cats, not a flock of sheep, to help prevent the single points of failure that results from widespread use of a single distro. If we make serious sales to the non-tech world, we should fight the widespread desire to have a single "one size fits all" computer that everyone is pressured to buy.
Those who do study history are doomed to stand helplessly by while everyone else repeats it.
I definitely agree that Sun probably has people on both sides of the issue internally. For example, they provided the source code for Solaris 8 and the Java libraries but not Solaris 9. Also, the Solaris 8 source had some curious omissions, which are probably the parts that have licensing issues (OpenGL, SCO-cruft, etc.).
So, what would be very likely, based on prior behavior, would be for Sun to possibly release _most_ of Solaris under GPL minus the parts their lawyers are still unsure of. This would still be a big win for developers, who can benefit from debugging at nearly all levels. I've already benefited from having the Java source, GPL or no, and fully understand what those developers in the 1980's must have felt like.
Vote in November. You won't regret it.
Hey if you don't want to release the source you can (and probably should) do a static build, and not dynamic, so the end user doesn't have to have ur specific version of the libs installed, unless it has a fairly stable API, and is common on most distrobutions
I think they can, they've bought very extensive rights about SVR4 from AT&T years ago. And they got based for paying SCO some money some time ago. So I expect they have all the rights to open source Solaris, at least the SVR4 parts.
I beg to disagree - Solaris cotains a lot of code from entities other than AT&T/USL/SCO (even though they have unlimited rights to use the code, i.e. no royalties due to SCO - they don't have the rights to distribute the code to others). One example would be the PostScript code in xsun.
A Shadeless room is a brighter room.