Will Sun look to use the GPL for offerings it has already open-sourced, such as the OpenSolaris version of its operating system? That was the topic of discussion in this exchange between Jonathan Schwartz, Sun’s president and CEO, and Rich Green, the company’s senior VP of software at an event to formally open-source Java. “We will take a close look at it,” Green said, adding that it was possible that the familiarity and comfort level developers have with the GPL may result in Sun adopting it for OpenSolaris.
Should be interesting. I see GPL will be used.
Edited 2006-11-14 23:31
Does it matter? In case if you would like to transfer code from OpenSolaris to Linux, I guess it does. But, personaly, I will use Solaris for its features and linux for its features. Best tool wins the job, and since jobs differ, tools will too.
This means OpenSolaris will be able to ship without binary blobs. (or better, can’t even ship with them)
It looks like Sun was listening to people. And if this comes true, they certainly got all the yesterdays bling and fame from Novell. Novell got points in community eyes for the same reason, releasing a bunch of quality code in short time. But, I doubt Sun would be stupid to follow them in the second example they gave to the world.
This week competition result
Sun: clean 10.0
Novell: -1.0
p.s. No I don’t blame Novell for getting in business with MS, I don’t blame them for providing indemnification like agreement. I blame them for putting FUD stickers on other people software, including mine. It seems they think they are more suitable to deploy my own software than me.
Edited 2006-11-14 23:43
“””Does it matter? In case if you would like to transfer code from OpenSolaris to Linux, I guess it does.”””
Question for those with experience working with OS kernel code. Does this matter? I always figured that Solaris and Linux differed so much in kernel internals that the kind of compatibility that would allow direct transfer of code… the type of transfer that would be covered by copyright, would be irrelevant. Am I correct? Or off the mark?
Its not untenable for things like drivers and filesystems, for example, as well as specific independent subsystems or data structures, depending.
Hmmm. Device drivers. Could Solaris use some of those?
ZFS would never make it into the Linux kernel. “Rampant layering violation” was the term Andrew Morton used, and he *likes* ZFS… its features, anyway. I suspect that Linus would be apoplectic if a patch were presented to him.
But still, if broken up into enough independent patches, it might just avoid giving Torvalds a heart attack.
This new Sun (if they are sincere, Deanlinkous) might be very interesting, indeed.
Let’s hear it for ponytails! Rah! Rah! and all that sort of rot.
Linux doesn’t need ZFS. It has NILFS. ( http://www.nilfs.org/ )
It’s not production quality yet it has a big company behind it just as ZFS.
Have you mod-downers read this?
NOTE: This feature is to report comments that are offensive, inflammatory, off topic, or otherwise in violation of the OSAlert forum rules detailed on this page.
Please do not use this feature to vote down comments that you merely disagree with, or even ones that contain factual errors or misinformation. Rather, use the reply feature and enlighten us all with your opinion or correct facts.
You’re wasting your time; if your post consists of more than ‘Linux r00lz’, its instantly marked down – perish the thought of actually contributing something substantial to the coversation given the fanboy kiddies around this place.
As for GPL; its a stupid idea to change to it; just stick to to CDDL and make changes where necessary to keep the noisy ones in the cheap seats happy; at the end of the day; even if Sun were to GPL all their software, you would still have the whiners from the Linux quarter whinging about something.
As for drivers; there are plenty out there already which OpenSolaris can ‘poach’, for example, the WPI driver from OpenBSD which supports the Intel 3945abg chipset (<rant mode> something that the Sun coders have failed to port – I raised the question and they replied “file a RFE” – excuse me, its a bloody obvious piece of hardware, you shouldn’t *need* to file a RFE to get hardware support for a bloody common device!) or FreeBSD or NetBSD; which are actually close to Solaris than Linux; and better still, how about improving the sound API; its terrible; geeze, just grab the FreeBSD OpenSound implementation, and we’ll all be happy.
NILFS looks interesting, but it’s no ZFS. I don’t see any pools or checksums.
Is there no end to the search for alternatives to the good or even the excellent?
I really don’t care if one or the other file systems comes out the ‘strongest’ in the sense that
‘strongest’ = the most freely implemented or the most conveniently licensed ….
I use the file system that’s right for the job…and I’d love to have the choice of ZFS on Linux as I love having Solaris as an alternative to Linux…
I think Sun has learned the hard way that software evolution, as exemplified by the Linux kernel, kills off, in effect, certain choices. It’s foolhardy to take choice for granted.
Of course, either people will appreciate the opportunity that a GPLed OpenSolaris as the opportunity that it is, or choice will suffer.
Probably depends on “when”.
If you look at their release of the HotSpot JVM, it’s under a “GPLv2 only” licence. And if they decide to GPL Solaris before GPLv3 is out, Solaris will quite likely be released like that, too. But be aware that this does not mean that they wont change to GPLv3 in the future. Sun is part of the GPLv3 process, and has stated that the main reason for not using it, is that they didn’t want to wait:
http://www.sun.com/software/opensource/java/faq.jsp#g24
They will require a transfer of copyright for every patch that goes into their source, much like the FSF does, so they will be able to legally and easily change to GPLv3. And given their involvement, it seems likely to happen.
I still think it’d have been better with the “or later” text, but at least they’re not going hit “developer permission hell” when/if they’re upgrading the licence to v3
I still think it’d have been better with the “or later” text, but at least they’re not going hit “developer permission hell” when/if they’re upgrading the licence to v3
Of v2, v3, “or later”, etc. I would consider “or later” to be the most radical condition for a business to go with. Their developers should read v3 now and make noise about what they dislike before it is set in stone in a couple of months.
Don’t forget about IBM.
Sun has undergone a major attitude adjustment to go with any GPL. The CEO Schwartz and others are already on record as being positive about GPLv3 and working with the FSF. The *technical* improvements of v3 over v2 cannot be lost on Sun and would be the public justification in any announcement. But I bet the strategic interests matter more in this decision.
Consider the big company identified with the Linux kernel (“GPLv2 only”): Sun’s arch rival IBM. IBM had been, in effect, wanting to bum Sun contributions in the Java realm. GPLing Java on Monday stymied IBM’s non-GPL wish.
IBM would benefit at Sun’s expense if OpenSolaris went v2 because goodies such as DTrace licensed this way could be ported more readily (from a legal POV) to the Linux kernel. Thus, Sun would want to go v3.
It’s a two-way street, of course: OpenSolaris could benefit from the driver work in Linux. Moreover, the only real complaint a group of leading kernel developers could come with against v3 in particular is that v3 would, in effect, prohibit DRM being forced upon kernel users. How would this “benefit” of enforcing DRM matter to the market segments (enterprise) of Sun?
It is not so obvious to me that porting *drivers* would be made significantly easier by being compatible with “v2 only”. Note that driver porting goes on quite a bit among the free kernels of differing licenses.
Finally, even though everyone knows it would be tough to move the Linux kernel to *any* other license from sheer paperwork, it was easy enough to for them to sound their alarm over v3 and how there would be a “balkanization” in the community of v2 vs. v3. Chances are the v2 only folks will yawn at OpenSolaris. The majority will be in between, and then a v3 fringe. With an aggressive v3 move OpenSolaris could recoup a bit of mindshare the Linux kernel stole from Solaris. Clearly, technical factors will be weighed, but this Sun could make some hay out of this. There are other possible motivations for v3: a greater sense of security in that Novell/MS types of circumventions will be harder. There are real improvements in v3, after all!
Now this mindshare thing by going v3 may seem a bit silly, but it is real: free software people will gain yet more respect for Sun’s leadership; and there will a lot of users out there who don’t know the GPL from the LPG but will wonder why Linux is v2 and OpenSolaris is v3. (Maybe I’ve seen too much in support!)
It seems to me that Linux has hit Solaris hard and Sun should make the most aggressive move and go v3 if Solaris is to matter in the long term and they are to gain from IBM here.
Edited 2006-11-15 01:30
“””It seems to me that Linux has hit Solaris hard and Sun should make the most aggressive move and go v3 if Solaris is to matter in the long term and they are to gain from IBM here.”””
This is starting to sound uncomfortably like the UNIX wars. Just with licenses instead of code.
Let’s not use GPL license versions as weapons.
Well, I am talking survival of *Solaris and wish to take nothing for granted. I also have nothing particularly against the Linux kernel.
Yes, the Unix wars would be bad, but much of the proprietary elements are at least missing here. Indeed, “wars” of some kind are ongoing but among free kernels with differing licenses. It all seems rather benign compared to the Unix wars.
One could also argue against a Linux kernel monoculture. You could see the alternative as wars or as diversity. As an optimist, I see it as diversity.
One thing we can count on is that corporations will use GPL license versions as weapons if it benefits them. I believe Sun used the GPL against IBM with Java and would have been specific about the version if they thought it would help them. (Of course, it was moot in the Java case.)
Forgive my clumsy, aggressive language, but I suppose that is the way some of the corps look at it. In fact, I’m just a conservationist of not-quite endangered species.
Edited 2006-11-15 02:10
b3timmons,
I enjoy reading your posts. What you say makes sense. I’ve always been ready to jump ship from the Linux kernel at such time as it was no longer fitting to my needs (and tastes), though I’ll admit to a certain amount of resistance to change. (OK, I’ll admit to being a Linux fan!!! )
Your suggestion of GPL being a pawn that Sun is using against IBM… is a bit of a downer. Damn you for making so much sense!
But… over all, it’s probably better than fighting over closed binaries.
I still think that RMS would have been better leaving well enough alone with his GPLv2 masterpiece.
But this is not the place for that argument. And I’m not the best debator.
Take care.
-Steve
I’m just summarizing a bunch of stuff I’ve read in various places as a way to better invest money and time. I feel the need to say that free software gets a bad rap because some see it as anti-business. I don’t buy that nonsense: we are on the verge of all kinds of businesses that will depend on free software.
Steve, I too am a fan of the Linux kernel. I installed Solaris x86 on a headless sparc workstation years ago in the process of getting some certs only to use it for a year and realize how much I missed GNU/Linux, in particular the goodies provided by apt-get (Debian). I know enough of the kernel to appreciate that few innovate technically as well as those guys. I think you will be optimistic and see diversity with OpenSolaris out there.
Have you read the latest v3 draft? I have read it and have tried to read the v2 repeatedly in the past. I am OK with both in principle, but for all of its merit, I am afraid v2 is becoming legal swiss cheese. I see it as a historic masterpiece. I can empathize with the kernel developers who after all are looking after their own interests as we all do. Empathy does not imply agreement, however.
There are so many wasted words out there over this when in fact the whole mess can be boiled down to differing assumptions. If the only thing we did was simply to state a few of our own assumptions, then a lot of bitterness would be avoided. But, right, this is not the place.
“Steve, I too am a fan of the Linux kernel. I installed Solaris x86 on a headless sparc workstation years ago in the process of getting some certs only to use it for a year and realize how much I missed GNU/Linux, in particular the goodies provided by apt-get (Debian). “
Oops–lame! I confess that I am not skilled enough to install Solaris x86 on a headless sparc workstation! An honest conflation of lovely memories. Time for sleep.
Steve, I too am a fan of the Linux kernel. I installed Solaris x86 on a headless sparc workstation years ago in the process of getting some certs only to use it for a year and realize how much I missed GNU/Linux, in particular the goodies provided by apt-get (Debian).
Well, you are very strong if you ran Solaris x86 on a headless sparc workstation. =)
Edited 2006-11-15 07:15
Well, you are very strong if you ran Solaris x86 on a headless sparc workstation. =)
Yikes–mixing memories silicon or otherwise is a dangerous habit.
Edited 2006-11-15 15:15
Your well written comment is appreciated, I feel exactly the same way. I thought I was the only one who thought of this (obviously not
.) The GNU(v3)/Solaris kernel would be an excellent “new” replacement for the GNU(v2)/Linux kernel in the eyes of true free software advocates, fans, and developers.
Not to mention the thought of saying GNU’s not Unix Sun brings tears to my eyes.
I watched the Sun netcast and was floored. Even RMS seemed shocked during his speech. When that happens you know Schwartz is on the right track and that his company means “open for business”. Countless numbers of times I heard them reference the next version of GPL (v3) and that they like the direction that it is heading. There is some serious information to share and a lot of situations to settle legally but this will be the finest day in free software (Solaris GPL v.3).
People called me crazy when I said that a “mistake” in a previous release of Red Hat Fedora which forbid tainting the kernel was a feature. You see, lads and lassies, free is the way to be. Schwartz realizes this as does the rest of the “information should be free as in freedom” hackers out there.
There’s no stopping us.
Thanks for the compliment, but the real inspiration can be found from EFF and FSF type of sources (see below). Yeah, I saw a bit of FSF reaction to the Java thing and they looked a bit disoriented–priceless!
Thanks also for speaking up about Fedora–you, they, and their users are stronger for it and better-prepared for changing times. Your example firms my resolve!
Profound change continues and Sun is lucky enough to be here now with an accomodating business model and enlightened leaders. Companies like Sun force me to re-evaluate my views on business, force me to believe that free software and business go together like milk and cookies. There is no communist conspiracy!
Anyone who has ever got a thrill out of writing their own little program or even just identifying with someone like that must identify on some level with these changes. An old tradition of code sharing is being restored to its rightful place. But these thoughts pale to the real agenda of the likes of the EFF and the FSF. What few people realize is that free software is not necessarily the main point–“free culture” is the more profound shift. Lessig and the Creative Commons are great, but look to EFF/FSF for unmatched eloquence.
Please listen to the speech (in English with a German title) “Die Gedanken Sind Frei” at http://emoglen.law.columbia.edu/
if you have not yet! Even if you want to outlaw free software, you can find out about the other side and learn how to speak properly.
That German title is from the following neat little poem (read it out loud!): http://www.cs.rice.edu/~ssiyer/minstrels/poems/1185.html
Edited 2006-11-15 03:16
Linux stands to gain a lot from this move, they’d get easy ZFS/Dtrace/etc ports.
OSOL stands to benefit too, mostly from the driver perspective. I wouldn’t want drivers to be copied “stock” over (not that they could) because of stability issues, but it would give OSOL devs a head-start when it came to writing device drivers for OSOL. This is pretty much the only area that OSOL is lacking right now.
Oh, a lot of the OSOL userland is antiquated. Some people like this, but personally – I’d prefer a lot of the GNU chain. If Sun would go the full mile, add whatever optimizations they can (without IP issues) from their compiler into GCC, and be done with it, it would make a LOT of people happy. Since they are “considering” GPL, they might as well just go the full mile while they are at it.
This is just ridiculous:
$ /usr/sfw/bin/gcc –version
gcc (GCC) 3.4.3 (csl-sol210-3_4-branch+sol_rpath)
Linux has NILFS, Reiser4(?), SystemTap, OProfile, and so on.
Both DTrace and ZFS are so tied to Solaris that a nearly full reimplementation would be needed for “porting to Linux” anyways.
Linux also has GGCC.
Dtrace in OS X 10.5, and it did not take them long at all to get it working.
What do you find ridculous? The the version of gcc shipped is 3.4.3? The reason for that is that’s the version used to build the OS/Net consolidation and it took quite a bit of work (and some gcc changes) in order to get it all compiled.
Upgrading to a GCC 4 base would be wonderful and it’s on the list of things to do but it will require some amount of work and regression testing before the switch is made.
If you’re interested in participating, please join the [email protected] mailing list or the
http://www.opensolaris.org/jive/forum.jspa?forumID=8
forum.
I’m quite aware of the reasoning behind the “cruft” in userspace. Both the Sun/UNIX derviced tools, as well as the /usr/sfw repository. Your second paragraph demonstrates what I was speaking of. To be specific, “would be wonderful”. Yes, it would. Improving GCC in the process, and making those improvements with (typical for Sun, which is a great thing) good judgement in terms of what to add/modify/etc. I know it’s not a priority, and the userland (especially the sfw stuff) shows it.
I’m already a member of a bunch of osol lists, I don’t really have time to dedicate to another one, nor is my area of expertise helpful. I’ll leave it to you fine folks who are capable of what I am not.
I was speaking in terms of OSOL becoming GPLd, and in that instance, it would make sense to start playing more ball with the GNU toolchain, and possibly some of the userland apps. Right now, I also understand Sun’s position. I *also* am quite aware of backwards compatibility requirements, as well as the extensive testing required. There’s a reason all of my servers run Solaris!
It’s pretty well known and accepted (check osol-discuss, there are weekly discussions concerning it) that a lot of things in the userland/especially sfw are fairly outdated, and it would be “nice” if they were not. It’s just a resource crunch at this point in time, but if OSOL does get GPLd, you’ll probably end up with an influx of people willing to help out, and GNU projects being more receptive to patches from a GPL friendly company. Not that I agree with this, but that’s how things operate right now.
To sum it up, your points are noted (and understand, and have been understood) – but I still think my thoughts are valid, even if unachievable at this point. Your “would be wonderful” pretty much validates that sentiment. I wish I was in a position/had the capability of making it a reality, but unfortunately all I am capable of doing at this point is providing my viewpoint, which may or may not help anyone!
Cheers,
David
Your points about the userland stuff (both legacy and sfw) is appreciated. A number of us are actually looking at this problem and adopting more/most/all of the GNU tool chain is definitely an option. The matter of resources is an issue and the first thing is to get the sfw repository pushed out into the open which is coming soon. At that point, getting things updated and more things integrated will be a lot easier.
My point around GCC is that it’s a bit special given it’s nature in the tool chain. Of all the pieces in the GNU tool chain, I suspect it will require the most care and feeding over time.
I see your point about GCC, and I apologize for my usage of the word “ridiculous” with the version output of gcc. I should have chosen another highly out-of-date program, instead.
You’re probably spot on concerning GCC and the difficulty in getting synchronized with the rest of the OSS “world”. I suspect it’s a worthy undertaking, though, because once you finally get OSOL building from a current revision of GCC, patching GCC where prudent/necessary, and OSOL when the problem is with the code itself – you’ll be ahead of the game. That is, assuming GNU/the GCC project will take your patches. I don’t know how restrictive they are, but I would *hope* anything that “fixed” issues in GCC or improved GCC in a positive fashion, while maintaining “correctness” to the standards would be accepted. Time will tell, I suppose!
All these people circling around solaris like vultures around a dying man in the desert. Hoping to steal the juciest morsels for themselves.
But the real question is why bother porting them to Linux at all just run Solaris. It has more cool features, and is more stable than Linux. Not just in system stability, but platform stability, you can run binaries that were built 10 years ago on the latest Solaris Express. Solaris also has the best testing so it doesn’t get slower from relase to release. All code that it gets included into Solaris is also reviewed by multiple people so things don’t break when you need it the most.
Of course if you want to run your company on the OS solaris also has the lowest support starting at just $120 per cpu socket. Of course service is provided by a company with decades of experience supporting enterprise customers.
Vultures of all kinds could be imagined, but I contend seeing those from IBM is the first step. Adapting your story accordingly is the best hope, IMHO.
With Linux permanently stuck at GPLv2, could it possibly end up getting largely replaced by OpenSolaris? (for the hardcore free software guys anyway)
This is a very smart move for Sun. If both OpenSolaris and Java are GPLed Sun gains developers (Linux kernel developers, gcc developers, classpath developers) instantly. All the work that goes into these projects can now benefit Sun and vice versa. Now all Sun has to do is create the best platform to deliver OpenSolaris and Java on.
I imagine there will be a huge upsurge in Java programming when everything is released under the GPL. It will be considered seriously by the GNOME project and other FOSS projects. Commercial developers will have a free, full-featured, and unencumbered language to work with. Imagine Java actually fulfilling its destiny as the basis of all web applications and imagine Sun delivering it. This is perfect timing too as Web 2.0 becomes all the rage.
Indeed. There are longstanding incentives of various kinds to get Java into GNOME. The big obstacle will be gone–just a matter of time now. GNOME is part of the GNU project and in light of various events so far this year I would not be surprised to see ever more favor shown to a GPLed OpenSolaris. I like your vision.
First off, I will openly admit that I am impressed with Sun’s embrace of OSS, in that we’re seeing action now instead of talk. GPL v2 for Java was a substantial paradigm shift for Sun, one that I frankly never expected to happen.
BUT, I’m seeing a lot of “free” love for Sun now as people seem to be implying that Sun is embracing the principles of free software while linux remains a corporate slut. Let’s be real here, they’re both corporate sluts. Sun is saying the right things and in some cases even doing the right things to win more and more respect from the FSF-type community, but I fully believe their moves are for pragmatic and strategic reasons, as were Linus’ from the beginning.
Java being GPL’d is recognition of linux’s strength in the datacenter, and is frankly a smart move that carries little risk for Sun. Solaris being CDDL’d and possibly GPL v3’d is an equivalent recognition of that, but for different reasons. Java in a manner needs linux (as linux in a manner needs Java), but Solaris competes tooth and nail with linux.
More to the point, linux development is anarchy. It’s not owned by any one or any organization. Certainly some people are decision makers and empowered to accept or deny changes and submissions, but it’s an environment where the organizations and individuals involved contribute on neutral ground, as much or as little as they choose, but to the ultimate benefit of all.
I simply can’t see Sun running the risk of becoming part of that type of model, any more than I could imagine Apple. That’s not the way these companies operate. People can cry freedom all they want, but at the end of the day all of these commercial companies investing in OSS platforms are doing so in order to reduce development overhead in an effort to maximize profits either directly or indirectly. Money is driving this, and their position and commitment can and will easily change depending on the money coming back. Nothing wrong with that, mind you, just pointing it out. The community may think they’re participants in this, but the reality is they are, for the most part, beneficiaries. Again, not that there’s anything wrong with that. I’m willing to participate in any model that encourages mutli-billion dollar organizations to pour money and development into a software platform that delivers the requirements I’m looking for, available at absolutely no cost to me aside from the time involved in learning, using and supporting it. And if my time is more valuable than that, I always have the option of paying them to do it for me. Basically I can be the beneficiary of high quality free software while the big money companies still manage to find revenue elsewhere for it.
There are incompatibilities between Sun’s approach and the “linux” approach that will keep them seperate entities with little ability to cross-pollenate and share code and development. Sun requires copyright assignment, and will continue to require it regardless of the future licensing for OpenSolaris. Linux by it’s nature cannot offer that unless individual developers (or organizations) step up and specifically grant it to Sun. And conversely, let’s face it, Sun will not choose v2 for Solaris. They’ve said all along they would consider v3, and since they hold “ownership” of OpenSolaris the patent issues in v3 that are giving HP (and assumptively IBM) the heebee-jeebies aren’t an issue for them. So a GPL v3 Solaris and a GPL v2 kernel are still, effectively, license incompatible.
Which is a shame because I do think the strength of the GPL has been the ability for competitive organizations to collaborate on development. But at the same time I’m not opposed to competition in the OSS space either, as ultimately it does drive innovation (witness KDE and Gnome, for instance).
Of course the OpenSolaris distributions may find a way to incorporate linux code since they don’t have the copyright assignment requirements Sun does, but if those improvements can’t be incorporated into the “official” project it just makes more work for them maintaining patches and addons each time a new release comes out.
And, putting on my cynical hat for a second here, there’s always the chance Sun could pull back. Certainly they can’t withdraw Solaris 10 now that it’s been released to the community, but there’s nothing forcing them to keep future versions and releases equally open, other than their goodwill and business requirements. It’s not inconceivable that some sort of prosperous community could be built around an open Solaris only to see Sun pull out with their next release if they feel they’re not seeing enough benefit or competitors are benefiting more. Again, just pointing it out. The nature of linux makes that far less likely to happen which creates more secure ground; IBM could decide they’ve had it with linux and pull all development tomorrow which would certainly hurt the project but at least the remaining participants could build on and continue. In a similar scenario with Solaris, it would essentially have to be forked if Sun decided to go a seperate way which would be counter-productive.
At any rate, I applaud Sun’s overtures. I think the OSS community will benefit immensely, regardless of Sun’s motivation. I just think we should hold off on granting them FSF citizen of the year; as with Linux’s embrace of the GPL, their actions aren’t necessarily motivated by the four freedoms. But again, nothing wrong with that as far as I’m concerned. Just adding perspective.
My 2 pennies.
Interests just happen to coincide for the moment. Sun can still talk freedom, but hopefully anyone would see that as PR, not as the FSF view of freedom as an end in itself.
Perhaps we could see OpenSolaris under GPLv3 and a “cathedral” model. Early next year several important projects will be under that particular arrangement–the bazaar is not everywhere. A dual-license is possible. Sun would presumably benefit from lessons learned in their OpenOffice project.
In contrast to Firefox logo issue, Sun will open source Duke (the Java mascot) along with this.
Duke reply in his blog.
http://blogs.sun.com/duke/entry/open_source_me%3F
To be honest i quite like the CDDL, it provides a nice middle ground between the GPL and BSD after all, FreeBSD for example can incorporate code (DTrace) and not have to worry bout relicenseing their entire codebase as would be case with GPL. Likewise for Dragonfly and ZFS.
It might however be a good idea to dual license certain things. For example if Dtrace was also available under GPLv2 then you would get a scenario of one system becoming standard across at least 4 *nix OS kernels (Solaris, FreeBSD, Linux and Mac OSX). That could only be a good thing imho.
How they can release source code that come from BSD as GPL one? GPL nazis would be happy to cry out loud.
I thought older versions of SunOS were based on BSD but newer ones were based on SVR4?
SunOS is in fact BSD based, Solaris is SVR4. This might help, but it is confusing:
http://en.wikipedia.org/wiki/Solaris_Operating_Environment
GPLv3 is germane to this news item. To help counter the FUD generated earlier this year, I challenge critics who have not yet read the current draft to do so:
http://gplv3.fsf.org/wiki/index.php/GPLv3_Draft
It’s a surprisingly short and pleasant read, much easier on the eyes that v2, IMO. Improvements can still be made and the following side-by-side comparison of the previous and current drafts might suggest further kinds of improvements to go into the next one:
http://www.fsfeurope.org/projects/gplv3/diff-draft1-draft2.en.html
Why not leverage the open source community to help sell your hardware?
Schwartz and Co are delusional if they think they can get driver developers to port to Solaris for free. There are other GPL’ed operating systems out there that haven’t benefitted from the Linux/GPL effect – why would you think that people want to contribute their time and effort to a OS controlled by a company.
Solaris may be GPL’ed but it’s still under a Cathedral like control – Sun has the ultimate say in the direction the kernel and apps go. Linux took off because there wasn’t any redhat or novell in the early days – there were just a bunch of cowboys winging it as they went along.
Finally – to assume that developers are falling over themselves trying to program for some GPL system – sorry that’s the biggest piece of crock that’s been propagated by the GPL zealots. There aren’t enough programmers to make Linux good enough to beat MacOS and Vista, you think there are programmers left to work on Solaris?
It’s called being delusional – in 1 year of Open Solaris – we still don’t have Virtual Consoles and I thought 4million copies of Solaris source code have been downloaded. In 1 year of Open Solaris – not one new driver contributed by the community has made it in. Only people paid by Sun are churning out stuff. I still can’t get USB wifi adapters to work in Solaris.
Jonathan – you want people to contribute to Solaris, how about trying to open your check book – it’s worked wonderfully for Google, Microsoft and other companies.
“Schwartz and Co are delusional if they think they can get driver developers to port to Solaris for free. There are other GPL’ed operating systems out there that haven’t benefitted from the Linux/GPL effect – why would you think that people want to contribute their time and effort to a OS controlled by a company.”
Schartz and Co don’t expect it for free, they have allready invested millions, and if they wanted to make driver porting go faster they have the power to do it, and they have in some cases, did everyone forget how fast Linux was ported to the Ultra T1 chip after a couple t2000’s were donated to the right people. A few simple programs and they could have all the drivers they want. One program could be port a driver get a Sun Ultra 20m2 large config. (retail value about $3000), want to guess how many drivers would be ported in 6 months under that program? 100 or so drivers would not seem difficult and even at retail cost its still only $300,000 with far more value added to the goodwill collumn in the anual report.
“Solaris may be GPL’ed but it’s still under a Cathedral like control – Sun has the ultimate say in the direction the kernel and apps go. Linux took off because there wasn’t any redhat or novell in the early days – there were just a bunch of cowboys winging it as they went along.”
But how much farther would Linux be if there wasn’t such a cowboy coding style? Go back and count how many times major sections of Linux has been rewritten, Just from memory the networking code has been rewritten a dozen times, and scsi and ide are probably closing in on a dozen times each that is millions of lines of code that have been generated that should of been redone maybe twice if done right. Also Sun doesn’t control what code people write, they enforce the quality of code that is accepted. If Sun/Solaris doesn’t accept your code you are free to fix it or realease a patch or even fork your own distro.
“Finally – to assume that developers are falling over themselves trying to program for some GPL system – sorry that’s the biggest piece of crock that’s been propagated by the GPL zealots. There aren’t enough programmers to make Linux good enough to beat MacOS and Vista, you think there are programmers left to work on Solaris?”
True coders are not sitting on the fence because of licensing issues, perhaps some users are, and most of them are on the fence because of PR issues, so this whole move to GPL is more a PR stunt than anything else. As far as coders are concerned Sun has enough coders allready on the payroll to maintain Solaris and grow it at a moderate pace, and as long as its being addopted by the corprate world at its current rate they will have money to continue doing it long term.
“it’s called being delusional – in 1 year of Open Solaris – we still don’t have Virtual Consoles and I thought 4million copies of Solaris source code have been downloaded. In 1 year of Open Solaris – not one new driver contributed by the community has made it in. Only people paid by Sun are churning out stuff. I still can’t get USB wifi adapters to work in Solaris.”
Who cares if Solaris doesn’t have virtual consoles its being worked on, not sure why? because Solaris x86 has had them in the past but it was decided that it wasn’t worth the effort to maintain them. 99% of users could care less about virtual consoles, that is what ssh, screen were made for, on a server they are of little use and on a typical users desktop its of little use, the secratary aren’t going to be hitting ctrl-alt-f1 to do anything. Do you really think its easy to get up to speed at porting drivers to Solaris it is over 8 million lines of code, and with todays commodity driven computer hardware prices its far easier just to buy supported hardware than to port drivers anyway. How many coders are going to think “hmm I need networking support, well I have this old NIC i could port a driver for it would only take 200 hours or so, or perhaps i will just drive down the computer megamart and get a supported card for $20” which would you pick?
“Jonathan – you want people to contribute to Solaris, how about trying to open your check book – it’s worked wonderfully for Google, Microsoft and other companies.”
he has openned his checkbook do a little research on java.net and sun.com and see how many programs have been ran that gave away free hardware and price reductions for developers.
> did everyone forget how fast Linux was ported to the Ultra T1 chip after a couple t2000’s were donated to the right people.
When developers can trade t2000s at the local grocery store for a gallon of milk or talk their landlords into taking a t2000 as rent money, then I’ll say it was generous of sun. Look these machines cost nothing to make – it’s not like these machines are going to be used by the guys who ported Linux to run their personal enterpise accounting package or their 200GB database.
Whenever someone donates hardware to get some software ported, it’s a freaking cop out. Tape 2000 C-notes on top of each t2000 and then it becomes interesting.
If you missed it, see the Java announcement (550MB .mp4 file) on
http://suntv.feedroom.com/
to get a better idea about OpenSolaris and the current Sun leadership.
Although you cannot escape the Java hype here, Schwartz nicely sums up the reach of Java. Spreading the radios” (GPLed stuff) and selling the “radio towers” (servers) was a metaphor. They made a convincing case that the move was a ton of work and fundamental change of culture for the firm. Green has a special attachment to (Open)Solaris.
Nonetheless, the Linux kernel is a tidal wave and GPLed Java is best seen in that light. I think the most productive assumption is Sun will adapt its business model to put OpenSolaris under GPLv3. How can one exploit this event?
Will the best “progress” happen with a single herd of cats? Need there be but one path to developing a kernel? I say no–competition such as GNOME vs. KDE is healthy. The problem of a competing kernel is not a scarcity of developers but a leadership problem. The right kind of person or team of people needs to convince developers of non-Linux kernels that unless they can organize their disparate agendas, their hard-won ideas and activities will have no viability.
That is, no viability unless coopted by Linux–toys in a tital wave. Isolated ideas are nothing without a viable organization.
Should “OSAlert” eventually be renamed to “LinuxNews”?
This year has been special with good and bad events, seeds sown to ripen in a spring harvest. Will the right harvesters be there?
Edited 2006-11-15 20:53
Sun has shown leadership. Anyone involved with *BSD kernels should think hard about the big changes developers at Sun are going through and figure out how to promote their own agenda if and when OpenSolaris is GPLed.
Is it better to have one GPLed kernel with a bunch of other kernels fading into obscurity?
Or can licensing pride and qualms be set aside for this one GPLed project for the sake of having more than one free kernel that matters?
Why?. There are benefits to BSD style development. Plus there is MORE opensource software that doesn’t use GPL than there is GPL software.
Are you going to develop for Solaris once it goes GPL? If not then STFU and stop handing out your free advice.
stephanem,
I intended no disrespect. If you thought my point was to disparage BSD licensing, then we clearly are on different pages. The growing popularity of the Linux kernel is undeniable and necessarily takes away from development of other kernels. If you disagree with this, we will have to agree to disagree.
Another thing to consider is that, just as Java will be dual-licensed, Sun may dual-license OpenSolaris. The picture does not have to be gloom and doom.
Edited 2006-11-15 22:37
“If consumers even know there’s a DRM, what it is, and how it works, we’ve already failed” – Disney Executive.
More on this at
http://www.defectivebydesign.org
If you are 100% certain that DRM has no potential to be abused, stop reading.
The Linux kernel has been and will increasingly be used to promote systems that force DRM upon the user. Faulting the kernel developers for not stopping this next year by changing to the GPLv3 license is not a good idea, IMO.
Instead of complaining, the next step should be to take advantage of Sun’s likely GPLing of OpenSolaris next year. If and when OpenSolaris is under GPLv3, a unique opportunity arises: it will be the only kernel that prohibits DRM from being forced upon the user. Anyone who does not want DRM to be forced upon them should figure out how to get involved in a kernel that cannot be used to promote that type of coercion.
Such involvement will not discourage DRM by itself of course, but the involvement will have value outside of your own agenda.
Edited 2006-11-15 22:13
If you are 100% certain that DRM has no potential to be abused, stop reading.
Not that this hasn’t been discussed ad nauseum, but GPL v3 has absolutely nothing to do with DRM implementation. A v3 kernel of any flavor does nothing to prevent closed/non-GPL restricted applications from imposing DRM. v3 will do nothing to force iTunes to open up, or stop the proliferation of BlueRay and HD. It will also do nothing to prevent hardware manufacturers from implementing DRM (a la HDCP) that bypasses software enforcement altogether and should be a bigger concern. DRM is strictly under the control of the content providers, and circumvention provisions are up to the legislatures in various jurisdictions. FSF has no say. The only way to make DRM go away is for consumers to stop buying DRM protected products, and that doesn’t seem likely to happen.
v3 is ultimately about preventing Tivoisation, nothing to do with content control, and you can bet that if Sun didn’t hold copyright and have the ability to modify their licensing at will, they wouldn’t be giving up that control to v3.
Besides, this argument about DRM being inflicted upon consumers sidesteps the issue that consumers still have choices in the products and services they choose to purchase. If I want to play my media anywhere I choose, I’m not going to do my shopping at iTunes. I’m not a big fan of DRM in principle, but I see it as simply a tool that can be abused as any tool can. I would prefer that DRM existed and consumers had more choice if the alternative is an abolition of DRM and a retreatment from digital content. I like to take responsibility for my choices and I prefer to educate myself and make balanced decisions. I’m not a big fan of allowing others to make arbitrary decisions in the name of protecting the “uneducated” public in order to protect them from themselves. And I don’t believe freedom is served by removing choice. I wouldn’t choose to use drm-enforced media products like iTunes or Media Player on linux, but I don’t agree that others shouldn’t be able to if they choose to spend their money that way.
On a final note, it might be worth mentioning that Sun themselves were driving their own DRM standard, an open sourced one that would at least prevent users of alternative platforms from not even having a choice of whether to choose DRM or not.
edit: typo
Edited 2006-11-15 23:18
I should have been more precise about the GPLv3 preventing Tivoization specifically. I am confident that the more people educate themselves about DRM, the more will be done to prevent it from being abused. The implications of Tivoization go beyond some notions of content delivery; it threatens a way of making software, software licensed with the intention of protecting fundamental freedoms. The GPLv3 stops this particular threat and therefore, as I have maintained, makes a v3-licensed OpenSolaris preferable to v2-only Linux for people who oppose enforced DRM. I will STFU about it for now.
Edited 2006-11-16 00:06
We have barely touched on the pros and cons of having a free kernel other than Linux that is reasonably popular going into the future.
Things like the Novell/MS debacle are already raising litigation FUD problems and might raise litigation itself. Having the alternative kernel would clearly spread the the risk of both of these kinds of problems. OpenSolaris will have the particular big advantage of having been vetted by Sun. A more popular OpenSolaris would immediately improve perception in this area and make investment in platforms based on free software more attractive than with just the Linux kernel.
OpenSolaris will still likely bring these kinds of risks, since Sun is not perfect. However, the risk is still spread because a worst-case scenario–all (i.e., both) popular free kernels stymied at the same time by legal actions–becomes much less likely than Linux only being stymied.
Edited 2006-11-16 02:07
OpenSolaris will still likely bring these kinds of risks, since Sun is not perfect. However, the risk is still spread because a worst-case scenario–all (i.e., both) popular free kernels stymied at the same time by legal actions–becomes much less likely than Linux only being stymied.
Yuk. I did not express that well at all. Consider the two cases in the future:
Case A. All Popular free kernels = {OpenSolaris, Linux}
Case B. All Popular free kernels = {Linux}
The likelihood of “litigation paralysis” is much less in Case A than in Case B. I hope that’s clearer. I.e., don’t put your eggs in one basket.
P.S. Yes, I know, Case B itself is much more likely than Case A itself. But since no one can show that the probability of Case B is 100%, OpenSolaris still matters.
Edited 2006-11-16 06:03
The power of assumptions is so neglected in these discussions. We should try harder to clarify our own assumptions (mentally if not on the keyboard). Then play those assumptions out as thoroughly as we can.
Sadly, most people fail to understand the following: the cool part about assumptions is not in their truth or falsehood but in their usefulness!
Of course, the assumptions differ in potential. The most neglected assumption I see is the commoditization of OS kernels by Linux. Influential tech giants such as Microsoft, IBM, and Google base their strategies on this and a few other assumptions. Others–whether they love or hate Linux–would do well to do the same.
Edited 2006-11-16 03:02
Warning: this comment might come off as condemning non-GPL licenses. Careful readers will see that the main point has nothing to do with this. (After all, the only piece of software in the topic of the discussion is just a kernel, not the entire OS!)
The following article and discussion about a Java project describe concerted anti-GPL efforts inside and outside of Sun:
http://lwn.net/Articles/209274/
I never knew the opposition to the GPL for Java was so intense, organized, and petty. This puts the future of OpenSolaris in a more sobering perspective. Pettiness is something all of us suffer from in various ways, but it was a major element here, IMO. Yet we failed to acknowledge this factor in our discussion. Note that I am not claiming anti-GPL = petty. There are excellent reasons for not using the GPL. However, pettiness was apparent among some of the attitudes in this case.
The only popular free kernel in the future will be Linux unless some kernel developers conquer their own pettiness.
So, I could be petty: five years from now I would be using and propping up an obsolete, once-great kernel (clearly not Linux!) that maybe only a thousand (and dwindling) other holdouts use, but that’s OK dammit because my license WXYZ is not the GPL!
No thanks. If the leadership and mature attitudes are not there for OpenSolaris, Linux will be the choice.
Edited 2006-11-16 08:10
Sorry, I missed the point you’re trying to make via the article you cited. It seemed to concern IBM’s “disappointment” that Sun released Java under GPL rather than donating the code to the Harmony project.
Where do you see a lack of leadership or mature attitudes on the part of the OpenSolaris project?
Where do you see a lack of leadership or mature attitudes on the part of the OpenSolaris project?
Sorry, I should have been clearer and please bear with my emotion. I am merely assuming that a GPLed OpenSolaris will be immediately “positioned” as THE Linux contender, and I am trying to figure out the biggest obstacle. I wrote:
“No thanks. If the leadership and mature attitudes are not there for OpenSolaris, Linux will be the choice.”
I am referring to a future, diverse group whose disparate agendas would be best served by realizing that OpenSolaris is the way forward. In particular, the *BSD fractures in the face of Linux dominance is the kind of wasteful problem that sufficiently enlighted and diplomatic Sun project managers could turn around to their advantage! It’s a fascinating management problem, I suppose. Could Sun pull this off?
I confess that this implies the accelerated demise of the *BSD projects. I have nothing against any *BSD–I just consider that the divisions, especially in contrast to Linux, are ultimately a tragic waste.
Pardon me for a little brain-storming: I am not sure if Sun realizes what GPLing Java has really done–the enthusiasm these kinds of actions will unleash among geeks, particularly the younger, more liberal crowd. Sun needs to come up with more stories where they can milk this for all it’s worth. A great story (even if told explicitly in only the mind of a leader) would help immensely if they end up GPLing OpenSolaris. Don’t yet write this off as flakey, hippie nonsense–Sun is getting into a new game and needs some humility.
Edited 2006-11-16 18:31
Sorry, I missed the point you’re trying to make via the article you cited. It seemed to concern IBM’s “disappointment” that Sun released Java under GPL rather than donating the code to the Harmony project.
The point I was trying to make was simply an intuition from not just the article, but also the discussion on LWN and a bit of development mail activity.
Of course, that intuition could be totally wrong–that’s one reason why I kept the claim of pettiness general, not aimed at any entity in particular. Anyway, it’s not the kind of thing I would want a project to be surprised by!