Linux leader Linus Torvalds has finally found something that could convince him that the forthcoming version 3 of the General Public License is worth adopting: open-source Solaris. “If Sun really is going to release OpenSolaris under GPL 3, that may be a good reason“ to move Linux to the new license, Torvalds said in a posting to the Linux kernel mailing list on Monday. “I don’t think the GPL 3 is as good a license as 2, but on the other hand, I’m pragmatic, and if we can avoid having two kernels with two different licenses and the friction that causes, I at least see the reason for GPLv3.”
I was worried about the community splitting so this is good news
There is one problem with this though. It’s unlikely that Linux can move to GPL 3 even if Linus wants it too. Because remember, that Linus put a clause in the GPL 2 that Linux uses that excluses the “or at your option, any future version of this license”. In otherwords, unlike the normal GPL 2, Linux code, cannot automatically update to GPL 3. Updating Linux to GPL 3 will require the explicit permission of anyone who has ever contributed code to the Linux kernel. Overall, it would be logistical nightmare trying to move Linux to GPL 3.
That’s not completely right. Many parts of the Linux kernel are licensed GPL v2 or later, but not all. Still, nobody said that moving a project like the Linux kernel to a new license woule be easy. And while I prefer the GPL v3 over v2, I have no problem with Linux staying GPL v2 for now.
And it’s not like it’s urgent for Linux to upgrade to GPLv3, even if the licence is applied to Solaris. They’ve lived without Solaris code for this long, I think they’d be able to survive until the GPLv2/only parts have been gradually exchanged for new code, or changed to GPLv2/later.
Edited 2007-06-13 11:34
http://gplv3.fsf.org/wiki/index.php/Compatible_licenses#GPLv3-compa…
If Solaris moves to GPL3, and GNU software will certainly move to GPL3, then you will be able to have a GNU/Solaris distribution, but you will not be able to have a GNU/Linux distribution any more. At least, you will only have an old (read unsupported) set of GNU software to pair up with you GPL2 Linux kernel.
GNU utilities include GCC (which outputs ELF binaries) and binutils (which includes the ELF loader) … there is no point in having a kernel if it can’t load anything to run.
So, if Solaris goes to GPL3, then the Linux kernel will have to also, otherwise it would become irrelevant.
Er. No. Not unless gcc and the other GNU software hook/link directly into the kernel somehow. Which they don’t, and which would be quite silly, I think.
GPLv3 applications are not prohibited from running together with, or on top of GPLv2 code. You just can’t interlink GPLv3 and v2-specific code.
This would be a problem if you want to use a GPLv2-only library in GPLv3 code. And I’ll bet most GPL libraries around uses either LGPL or GPLv2-later licences. And in those cases, it’s a non-issue.
That’s false. There’s no reason you can’t have software with multiple incompatible licenses running on the same computer. That’s like saying I can’t run GCC on windows because the kernel is owned by MS. As long as they don’t link together it would be fine and nothing directly links into the kernel except for drivers and kernel modules.
unless the software is using glibc which is under LGPL. but that may change, so future glibc will be under GPLv3. and any software using glibc cannot run under linux kernel.
Where did you hear that glibc would be relicensed under GPLv3?
Actually, he may be right about that… HOWEVER, if that is to happen, it will be done with a GPLv3 with added permissions to make it functionally equivalent to the LGPL. I believe there were musings about constructing a sort of “new LGPL” that way.
I fairly often see people who appear to underestimate the FSF a bit. The FSF may have strong principles and ideology, but that doesn’t mean they’re stupid. They compromise, they are realistic, and often practical about what they do.
They’re quite aware that making the GNU stack incompatible with the Linux kernel (either legally or technically) is probably on the list of the “10 most stupid things the FSF could do”. Likewise is putting glibc under a pure GPLv3 licence (thus forcing all code compiled using glibc to become GPLv3. Cool in theory, but a really stupid move in practice).
“””
I fairly often see people who appear to underestimate the FSF a bit. The FSF may have strong principles and ideology, but that doesn’t mean they’re stupid. They compromise, they are realistic, and often practical about what they do.
“””
Sometimes those of us who “appear to underestimate the FSF a bit” (I sincerely like your choice of wording, there) do it to help ensure that the FSF does not act stupidly out of a feeling of overconfidence.
If they weren’t willing to compromise, and even leech a bit, they’d have executed another item on the “10 most stupid things the FSF could do” list, and put their idealistic weight behind the GNU HURD.
It’s easy to collect on sites where FSF sentiment is strong, and soak up all the “feel good” vibes about us taking the world by storm.
But once one ventures out of that fantasy world, and into the real one (Gnew what?), compromise and practicality become vital necessities again.
The FSF, despite their rhetoric, understand that idealism alone is worth little. If they do not have a system that can compete with others in the real world they’re sunk.
But anyway, last I looked, the glibc guys and the FSF were not so chummy.
Edited 2007-06-13 19:28 UTC
“””
unless the software is using glibc which is under LGPL. but that may change, so future glibc will be under GPLv3. and any software using glibc cannot run under linux kernel.
“””
No. “system software”, software which may generally be expected to exist on a particular platform, is exempted. Otherwise, you wouldn’t be able to run GPL software on Windows, etc.
I think that the Linux kernel and glibc both qualify as “system software” on most Linux systems. 🙂
Edited 2007-06-13 18:19
Why did this post get modded down?
It doesn’t offer any insults to anyone.
It points to FSF’s own words about GPLv2 being incompatible with GPLv3.
The conclusion may or may not be incorrect, but that is no reason to mod it down, surely?
http://www.osnews.com/rules.php
There was no offensive language or personal attacks, therefore modding down was not appropriate.
Yeah, it probably shouldn’t have been modded down. It might be incorrect, I don’t know I don’t keep up with such things, but it didn’t rate a -2
OS threads are rough places. Certain groups of people are territorial.
Call the mystery modding the digg effect, I guess.
Linus want to move the Linux kernel to GPL 3 in order to take some code from the Solaris kernel.
So despite what the linux zealots are saying, there is unmatched features in the solaris kernel (ZFS, Dtrace, etc…)
Now that Linus himself acknowledged that, I have something to show each time a linux zealot is trolling about Solaris 10.
I don’t think anyone in the Linux camp ever denied that Solaris 10 had some very nice code and features, and that being able to use it in Linux might be neat.
What has been somewhat discussed is the relative merits of porting over Solaris code straight vs. using it as inspiration for a re-implementation of the relevant features in Linux. The former having the advantage of being potentially quicker, the latter having the advantage of the resulting code potentially being better integrated with the kernel and more ‘Linuxy’ which could ease maintainability.
“””
Now that Linus himself acknowledged that, I have something to show each time a linux zealot is trolling about Solaris 10.
“””
I’m replying to this post because the parent post, from which the quote is taken, is already at -5, as it should be, and cannot be replied to.
First, I would like to ask Solaris and Linux advocates alike to really consider whether a “turf-wars” attitude really helps their cause. In particular, if your real cause is allowing FOSS to maximally benefit the people of the world, does it really make sense to treat the OS Wars as a zero sum game?
I generally like GPLv3. (B3timmons’ jaw drops to the floor.) 😉
And sharing a license with the excellent Solaris kernel would have some real advantages in the long term.
But a move to it would definitely come at a substantial cost. It’s likely that it would involve rewrites of a good bit of code, not for technical reasons but for political ones. Linux’s resources are pretty impressive, these days, but not unlimited. So one should expect resources to be diverted from other desirable work in order to accomodate the move.
Also, as I have said before, it would introduce the potential for the ongoing promotion of fear, uncertainty, and doubt surrounding the Linux kernel. (Note that I spell the words out to indicate that they are literally what I mean.)
Microsoft, and/or other threatened parties are unlikely to stop treating the OS Wars as a zero sum games any time soon. And I forsee a lot of insinuations out of Redmond that Linux is improperly licensed. And since copyright law defaults to 3rd parties having no rights to a copyrighted work, any rights beyond fair use being granted by a valid license, should one exist, then if Linux were found to be improperly licensed, everyone would automatically lose the right to use it, effective immediately. At least until *everyone* could agree upon a new one. And the GPLv3 drafting process shows us that such a new agreement might take a very long time, indeed.
So a move of the Linux kernel to GPLv3 is *not* something to be taken lightly, regardless of how one feels about the advantages, or the desirablility, of a GPLv3 Linux kernel.
In my heart, I find that I would be for such a move. But my brain keeps wondering if my heart is right.
Just some things to think about…
Edited 2007-06-13 14:32
The preparations for a move could be pretty simple. All Linus & Co have to do is to stop accepting “GPLv2 only” contributions from this year on.
Since the Linux kernel systems seem to get overhauled more often than in comparable kernels, there would come a day that the Linux kernel would be mostly “GPLv2 or any later version”. At which point the switch could be made to GPLv3 or GPLv4 relatively easy.
“””
The preparations for a move could be pretty simple. All Linus & Co have to do is to stop accepting “GPLv2 only” contributions from this year on.
“””
Yes, I should think that would make sense as a minimum measure, coming at little cost. Although it is not as uncontroversial as one might, at first, think. Should “GPLv2 or later” submissions be accepted, or only “GPLv2 or GPLv3” submissions.
Don’t forget the reasons for not including “GPLv2 or later” in the first place. I made this related post recently:
http://www.osnews.com/permalink.php?news_id=18064&comment_id=246814
Also, don’t expect Linux kernel subsystems to continue to be rewritten from scratch. Linux, a relatively young OS, is entering a period of maturity. If the LK devs are doing their jobs properly, the need for complete rewrites should be steadily dimishing.
Edited 2007-06-13 16:31
OSS challenges the assumptions on which copyright is based, and therefore its legality will always be in question to some extent. We must realize that new models for collaborative content creation are largely untested in the courts and represent a considerable legal gray area. The effectiveness of these models, as well as the social and economic value of the resulting works, is the only real mitigating factor.
It turns out that there is incredible value in sharing content, just as there is value in restricting content. The GPL is at the forefront of the conflict in that it represents a powerful combination of sharing and restricting covered works. It’s a relatively simple and intuitive idea that is rather complex to describe in the context of copyright.
The issue of relicensing a collaborative work is just one of the complexities, and in the case of OSS, it’s exacerbated by the limited foresight of the pioneers of various OSS licenses. One can argue that the entire justification of the GPLv3 is to codify the intention of the GPLv2 with the benefit of hindsight. But others argue that the perceived loopholes in the GPLv2 are pragmatic and beneficial.
There is no legal resolution stating the rights of a contributor in influencing any future efforts to revise the licensing terms of a collaborative work. These rights are established as terms of contribution and may be a condition of the copyright license itself. This creates a bit of circular logic.
Any attempt to relicense some or all of the Linux kernel, no matter how carefully it is executed, would be no more legitimate than the steady-state legal status and probably less so. There are no certainties in the legal nature of copyright in the context of collaborative works, especially when it comes to the GPL.
We can debate the legality of various means of relicensing until we are blue in the face, but we will never know how a court will rule until it is too late. Furthermore, it’s well established that PR spin is a sword at least as mighty as the legal system itself.
So we’re back to my point about the social and economic value of the software mitigating the risk of legal action. As we’ve just seen in the case of Microsoft’s patent covenant campaign, this is not an effective deterrent for those with either a lot or very little to lose.
Microsoft “insinuations” amount to squat. Copyright law provides near-unlimited discretion to the authors of code over what they permit or deny others to do with that code. There is no doubt (aftter being open to public view for many years) that GNU/Linux code is indeed written by the authors who claim it as their work. Therefore, those authors have the discretion to decide what other people may or may not do with their code, and Microsoft has no say in the matter.
If Microsoft did want to have a say, and took the matter to the courts, such a move would subject Microsoft’s own code to examination by the FOSS authors … who would without doubt find copyright/patent violations in Microsoft’s code.
IMO Microsoft would not win in such a scenario.
In any event, Microsoft rumblings about “Linux is improperly licensed” have next-to-zero relevance anywhere but in America, so the point is largely moot to most of the planet (the sane part of it). I dispute the contention that “everyone would automatically lose the right to use it, effective immediately”. The maximum scope for your scenario would in reality be: “Americans would automatically lose the right to use it, effective immediately”.
So despite what the linux zealots are saying, there is unmatched features in the solaris kernel (ZFS,
Some developers are opposed to ZFS, because it breaks layering.
Dtrace
Did you ever hear of systemtap?:
http://sourceware.org/systemtap/
Now that Linus himself acknowledged that,
Actually, the most interesting mail from Linus on the subject is missing from the links in this article:
http://lwn.net/Articles/237905/
Quotes:
Ergo: they’ll not be releasing ZFS and the other things that people are
drooling about in a way that lets Linux use them on an equal footing. I
can pretty much guarantee that. They don’t like competition on that
level. They’d *much* rather take our drivers and _not_ give anythign
back, or give back the stuff that doesn’t matter (like core Solaris:
who are you kidding – Linux code is _better_).
So the _last_ thing they want to do is to release the interesting stuff
under GPLv2 (quite frankly, I think the only really interesting thing they
have is ZFS, and even there, I suspect we’d be better off talking to
NetApp, and seeing if they are interested in releasing WAFL for Linux).
I am not saying that I necessarily agree, but you are clearly misrepresenting his opinion here.
Edited 2007-06-13 12:17
Yes, and it is nothing like DTrace. SystemTap isn’t mature, isn’t complete, and is pretty much a joke and will be for a while.
Come back in a few years, maybe SystemTap will be worth looking at then.
DTrace is also more than just ‘tapping the kernel’ it is an all pervasive tool, it goes into Xserver, libraries, kernel, even applications (Firefox Dtrace capabilities anyone?).
Those who dimiss features in Solaris need to do a bit of reading before opening their mouths; then again, those who bash Solaris, bash Windows whilst not having the slightest clue about the technology itself.
Sometime back I read that systemtap’s design itself has flaws. I think that was from a DTrace developer.
If that is true, then systemtap is not comparable to DTrace that has already proven to be stable and very useful to developers.
License compatibility is important and Linus knows that. For the sake of opensource, isn’t that one of the main reasons for doing so? Linus is saying that if Solaris is GPL v3, then he will make the effort to move the kernel to the same version so the code can be compatible. Who benefits from this? Everyone. Solaris gets Linux drivers, Linux gets Solaris’ technology, both communities merge and now there are two kernels to play with.
No Solaris doesn’t. I am sick and tired of people citing this over and over. Drivers are not somehow magically portable between operating system. In many cases, the drivers would have to be completely rewritten from scratch to work on Solaris and vice versa.
That is one of the reasons our community (the OpenSolaris community) hasn’t fallen into the “GPL driver trap” yet.
The argument does not hold water.
We would be almost no better off tomorrow if we changed our license to GPLv2 today for drivers.
We can use the knowledge contained within the GPL drivers to write new ones if we wanted to right now.
“””
We can use the knowledge contained within the GPL drivers to write new ones if we wanted to right now.
“””
A naive question, perhaps. But… why don’t you, then?
Compared to Linux and BSD, OpenSolaris driver support is still a wasteland.
I would encourage OpenSolaris to utilize all available resources.
Edited 2007-06-13 20:05
Its all about “scratching itches.” Most of the people that can write drivers already have working hardware or are working on other things instead of drivers.
People like me who would probably work on drivers if they had the hardware, don’t have the hardware…
The hardware support is actually not as bad as people think. I have three completely different systems that all “just work” once I download and install the drivers from a manufacturer’s website (for my network chipsets) just like Windows, or it just works out of the box (nVidia Video Card, etc.).
In fact, my Intel Core Duo 2 system actually was able to boot Solaris Express *before* I could boot Fedora or Ubuntu on it (they had support for the new hardware first).
That’s the problem. Right now, most of the people working on Solaris (over a thousand) are Sun employees, and as such, their time is allocated by Sun who is naturally interested in primarily supporting their own hardware or hardware that their customers have directly asked to be supported.
For hardware that their *paying* customers don’t care about or that an engineer doesn’t have personal interest in, you probably won’t see too much work being done. Sun has commendably done some work in this area, and has even worked out a deal with 4Front Tech to bring OSS to Solaris which should substantially increase hardware support as far as audio is concerned.
In fact, it will put audio on Solaris at a level above what Linux currently has.
Edited 2007-06-13 21:48
I’d encourage all OS’s to try and share driver code and info. I’m sure I’m not the only person whose hardware and OS choices are based on what works on what OS.
Too bad their not. It would solve the driver deficiency on most alternative OSes, and Minix would recognize my old Xircom PC Card.
You’d think that most people who come here would know of the vicious cycle that is userbase and hardware support.
Anyway, just out of curiosity, can’t Solaris import drivers from the BSDs fairly easily, OpenBSD specifically. Not just because of the license, but because it was originally a BSD then it had System V stuff merged onto it. I thought I heard that somewhere.
Merging of NIX operating systems is an unavoidable necessity. Things are already happening on the user interface level (XOrg, Gnome), and GPL3 is a nice platform and opportunity to unite. We need hardware vendor support right now, hope they get out of their IP holes as soon as possible. If users and developers could get serious and stop comparing their dick size, that would also be great.
Nice thought. Yes Solaris has BSD ancestry. SunOS was BSD based, which is Solaris 2.5 and earlier iirc, 2.6 onwards still contain SunOS, but its based around AT&T and SysV.
Porting drivers between BSD and Solaris is probably about as hard as porting between AIX and Solaris. Its probably that 80/20 rule that bites when you get to the ‘flavour specific’ bits.
What needs to be remembered is that Solaris is a server OS, always has been, always will be. I agree, it would be nice if it supported all hardware, like your Zircom PC card, but then why should it? Solaris on SPARC certainly doesn’t need to, and Solaris on Intel is really for cheap x86 headless (or built-in graphics) servers to compete against Linux in the enterprise where a support contract is the currency. I honestly don’t think Sun has an agenda to ‘destroy linux’ as others think – Sun doesn’t give a rats what you run at home for free, they care about what your company buys (and they hope its Sun!).
There’s space for both, Linux is a jack of all, master of none, but since its free, that’s ok. In fact, its great as it turns out to be a benefit not a problem because it makes it ubiquitous. Its good enough as a server, and not bad as a desktop, oh yes and did I say its free?
Solaris is simply a server workhorse, even as a Solaris goon I wouldn’t want JDS as my workplace!
Thanks for clearing that up. I know pkgsrc runs on Solaris, so I guess I was thinking the porting of drivers would be easier.
I don’t think they do either. I really think they are more of a hardware company then they are a software company, and all they want is to support servers for a fee. I know the Sun guys where I work don’t care about the OS the hardware is running. If the hardware is running Solaris, there are nicer tools for it though.
Nice thought. Yes Solaris has BSD ancestry. SunOS was BSD based, which is Solaris 2.5 and earlier iirc, 2.6 onwards still contain SunOS, but its based around AT&T and SysV.
No, Solaris was always Sys V-based. It was a very controversial move.
I honestly don’t think Sun has an agenda to ‘destroy linux’ as others think – Sun doesn’t give a rats what you run at home for free, they care about what your company buys (and they hope its Sun!).
I don’t know if I would go so far as to say that Solaris wants to destroy Linux, but they definitely want to take a good chunk of Linux’s user base.
And they do care what you run at home for free; they want you to run Solaris/Open Solaris. Linux got a strong foothold in IT because the “opinion makers”, the guys passionate about technology and ran Linux at home, said “screw these expensive Sun boxes, I run this stuff at home, and it works great. Let’s replace our infrastructure with x86 boxes running Linux”. And it works, and it worked great.
Sun wants people to start developing in Solaris again, for distributions to pop up using Solaris, for appliances to pop up that run Solaris, for cheap dedicated server providers to start offering Solaris.
They know that will have an expansive effect, just like it did for Linux. Sun wants to be the new Linux.
I disagree; but the simple fact is, Sun has very limited resources, do they focus on a small niche (Xircom pccard) or those who make up the majority (Intel based wireless cards)? Sun has limited resources, and therefore, has to be prudent about where they put it.
For the vast majority of there, their hardware will be supported – Intel centrino (before their 4295 n based wifi) – 2100/2200/2915/3945 are all supported – Atheros which is a very popular wireless chipset, again, that is supported.
Sun is focused just on servers – that is patently incorrect – they’re focused on workstations, servers and developer desktops/workstations – have a look at Solaris Express Developer Edition 05/07 – it is the perfect desktop for developers of web applications, java developers and so forth.
For me, I’m running SXCE and updated it to B66 today – rock solid and reliable as a desktop – if you have a look at my blog, it does into details over what I’ve done so far – heck, I’m compiling Pigin 2.0.1, and it is awesome. It might not support *every* piece of hardware, but there is nothing stopping you or your friend with the Xircom card from analysing existing sources out there and writing a driver or contributing to the code.
OpenSolaris is an opensource project, one can’t expect Sun to do all the work – the whole point of opensource is for those outside the corporation to work on area’s that directly affect them. Heck, if it is really urgent, and you know enough people, why not get together, find the specifications, and put up a bounty on getting Xircom (or any hardware) support for Solaris.
Edited 2007-06-14 16:21
Linus could publish a very public notice of his intention to relicense the kernel under GPL 3. If the notice was filed worldwide in significant news outlets asking the heirs to any past kernel contributor to raise any objections that they may have before a certain date, I see no reason why the kernel could not be relicensed.
Linus has a wonderful long-term view and understanding of the IT industry and he has rightly seen that GPL3 brings together a wealth of code contributions from all walks of the IT spectrum, by removing license incompatibilities with licenses such as the Apache license.
Also kudos to SUN for avoiding the temptation to create their own separate playground. The minute that SUN releases Solaris under GPL3, I will begin using it commercially. For now, I am testing it and learning more about it.
If the notice was filed worldwide in significant news outlets asking the heirs to any past kernel contributor to raise any objections that they may have before a certain date, I see no reason why the kernel could not be relicensed.
I’m not an expert, but I’m fairly sure copyright law doesn’t work that way. Lack of objection doesn’t constitute acceptance.
At any rate publishing large ads in major newspapers and magazines around the world would cost A LOT of money.
Edited 2007-06-13 12:01
Lets not put the cart before the horse. Linus did not say that he would move the kernel to v3. He only said that if Solaris moves to GPLv3, it “that may be a good reason.” This says nothing about moving to GPLv3. All it says is that he (Linus) might have found a good reason to move to GPLv3. If Sun puts Solaris under GPLv3 and if Linus decides the kernel should be moved to GPLv3, then he will. Logically, if Sun chooses not to put Solaris under GPLv3, then the “good reason” fails. Also, remember, GPLv3 is a double edged sword. With GPLv3, Sun could keep all the neat stuff (Dtrace and ZFS) for themselves. Linus says (from the article), “To Sun, a GPLv3-only release would actually let them look good, and still keep Linux from taking their interesting parts, and would allow them to take at least parts of Linux without giving anything back.”
Remember, Linus still thinks that the GPLv2 is the better license. Even if Sun does move to GPLv3, that is no guarantee that Linus will move to GPLv3. This article is speculation at best.
So. it is even possible for Solaris to give up GPL v3 release, if Linux kernel goes GPLv3…. ?
Yes it is possible; it is also possible that Sun will give away all of their IP. I think both are improbable. Sun will make a decision based upon what is good for them, not necessarily what we would like.
Not as general a focus as OSNews, but a worthy read indeed: http://lwn.net/Articles/237905/
Ingo Molnar:
http://www.ussg.iu.edu/hypermail/linux/kernel/0706.1/1123.html
“And that’s i guess what OpenSolaris lacks
and which i suspect it is mostly interested in: lots of nice Linux
drivers 😉 XFS, the largest Linux filesystem is 100K lines of code –
and ZFS (i’ve never seen it) is very likely smaller than that. Linux
drivers on the other hand, as of today, are _3.7 million_ lines of code
and enable Linux to run on 99% of the hardware that is produced today.
Guess which one has the larger strategic significance? ;-)”
More from Linus:
http://www.ussg.iu.edu/hypermail/linux/kernel/0706.1/1722.html
“So they want to use Linux resources (_especially_ drivers), but they do
*not* want to give anything back (especially ZFS, which seems to be one
of their very very few bright spots).”
Oups… seems like many others posted almost the same thing while I was fixing an account.
Edited 2007-06-13 12:52
http://blogs.sun.com/jonathan/entry/one_plus_one_is_fifty
Edited 2007-06-13 13:13
Jonathan Schwartz says…
Be sure to read Theo Deraadt’s comment after that:
http://blogs.sun.com/jonathan/entry/one_plus_one_is_fifty#comment-1…
It’s easy to talk about openness and freedom. And Sun has certainly done more than that by opening OpenOffice, OpenSolaris, and the JDK. But there is still some way to go. They *do* seem to (want to) compete with the community. If they wanted to make ZFS the defacto UNIX filesystem, there would be more interesting licensing options. If they wanted to coorperate between all communities, they would have provided specifications for their hardware.
Sure, it’s their own right to do as they do. Sun is a company, and is acting in their own interest. That’s what companies do. But it is good to realize that Sun is not a community project. Even OpenSolaris isn’t in some senses.
Edited 2007-06-13 13:44
You should read Jonathan’s blog. He writes, losly cited, “Companies compete, communities simply fragment.”
Sun, under Jonathan’s lead, does not try to compete with the community. If so, he woulnd’t have pressed Sun into releasing Java and Solaris as OpenSource. Jonathan knows that the future of the computer industrie lies within services, not selling software.
At least I hope he does.
Theo de Raadt saying that? No.
Edited 2007-06-14 16:29
I’m sorry, after reading this I’m left with a smarmy aftertaste.
SUN caused the rift, and now they are trying to repair it, this I applaud. However, Jon seems to want to put some of the blame on the OSS community for Schwartz’s mouth?! Amazing. SUN pulled the sword, not the OSS community.
And this reads like a veiled press release for SUN. “Look at the things we did.” How about a sincere sorry for the wrongs SUN committed, like talking trash about Linux and OSS and being a shadow partner to the SCO saga. Hey, there you go, if you want to show how much you’ve changed, come clean about SCO.
Actions speak louder than words. I’d back off myself, and see what SUN does for awhile before I would start trusting them again. But that’s just this cynic’s opinion.
Sun was under different managment back when they were trying to fight Linux, as I recall. Nowadays, under Schwartz, even Java is going OSS, which is MORE than enough of an apology to the community I’d say. Schwartz is also smart enough to realize their fight anymore is with Red Hat, not Linux. And that’s just the reality of fair competition.
Sure it reads like a press release, but if Linus farts it gets plastered all over the internet, and since this latest comment from the Benevolent Dictator includes a few plugs at Sun, it’s only fair for Schwartz to defend his pride and joy a bit. Linus would’ve done the same. Except for inviting anyone over for dinner
I meant Scott McNeally’s mouth, not Schwartz’s mouth. McNeally was the reason for the OSS community/Sun community split, not Schwartz.
“””
McNeally was the reason for the OSS community/Sun community split, not Schwartz.
“””
Ah, Scott! Aggravating, opinionated, outspoken, and in a strange way, lovable, Scott!
One day he’s insulting the hell out of Linux… the next, he’s on stage in a penguin suit. (Which, I hear it rumoured, still hangs in his closet.)
I agree that he had much to do with the negative feelings that Linux fans have toward Sun and Solaris. The serious case of multiple personality disorder that Sun developed in the final years of his reign were likely indirectly related to his presence… but who knows.
I’ve always had a lot of respect for him and Sun Microsystems, though.
You see, while Linux is riding high today, things were not always so. There was a lengthy period over which Linux existed, but was not “ready” in the way that it is today.
And during that period, while SGI was off selling Windows NT and HP/UX was churning in the doldrums, Scott and Sun stood, steadfastly, by Unix’s side.
I have no doubt that had it not been for Sun and Solaris on the high end, and SCO, with their OpenServer and Unixware on the x86 low end, OSS unix-like OSes would not be where they are today, because we’d be trying to revive the idea of using Unix on servers, instead of the easier job of simply proving that we can do what the existing Unix boxes can do, and nibbling Windows installations on the side.
The move to Schwartz was necessary, and has been beneficial.
But we owe Scott, Sun, and even “Old SCO” a debt of gratitude for bearing the torch before we were of an age to do so.
Edited 2007-06-13 15:47
I have no doubt that had it not been for Sun and Solaris on the high end, and SCO, with their OpenServer and Unixware on the x86 low end, OSS unix-like OSes would not be where they are today, because we’d be trying to revive the idea of using Unix on servers, instead of the easier job of simply proving that we can do what the existing Unix boxes can do, and nibbling Windows installations on the side.
The move to Schwartz was necessary, and has been beneficial.
But we owe Scott, Sun, and even “Old SCO” a debt of gratitude for bearing the torch before we were of an age to do so.
I’m not sure old SCO (or BSDI) was much of a factor. The enterprise market was mostly Sun and a smattering of other operating systems.
Sun did create the Linux movement by refusing to acknowledge that SPARC rapidly became uncompetitive on the low and mid-end. In 1998, Sun’s slowest SPARC processor (the one in the Ultra 5) was about as fast as Intel’s fastest processor. In just a few years later, however, Sun’s fastest low and middle processor was far slower (and far more expensive) than Intel/AMD’s offering. Sun continued to think $25,000 was a reasonable price to pay for a low-end web server, and there was a giant sucking noise from the air rushing in to fill the void of Sun users that jumped to Linux and other open source operating systems running on x86.
“””
I’m not sure old SCO (or BSDI) was much of a factor.
“””
Think “business accounting” and “point of sale”. Strong areas for OpenServer. Critical for businesses, but not much talked about because, face it, it’s totally unsexy.
The company that I work for sells and supports business accounting and point of sale software. And let me tell you… if it had not been for SCO, my Linux efforts over the last 11 years would have come to nothing, and *all* our clients would be running Windows.
Edited 2007-06-13 16:16
Wow Linux truly is evolution and perhaps a revolution too as it brings a cooperative, OSS, way of doing things.
Now pardon me, do we not hear that that BSDs are better than Linux in so and so way and better in a general sense.
So does this ‘ready for the market place’ mean. Were not the BSDs also ready for business. If they are better they should have been ready before linux was. Or does ‘production quality’ also involve marketing or word of mouth advertisement so as to speak. If it means support then Red Hat has been there for a long time.
Please, I am not trying to bad mouth any OS out there, I respect them all the same. I am just asking, that does an OS really need to be good or does it just need endorsement from a multi billion dollar company to be ready for the market place?
“””
So does this ‘ready for the market place’ mean. Were not the BSDs also ready for business. If they are better they should have been ready before linux was.
“””
None of the BSDs have official vendor support for any of the accounting and point of sale software that my clients run.* RHEL does. End of story.
I did not intend to put BSD down in my previous post, but official vendor support makes Linux the appropriate OS for the job.
* No suitable FOSS exists to replace these apps.
Edited 2007-06-14 01:57
“””
I’m sorry, after reading this I’m left with a smarmy aftertaste.
“””
Smarmy? That implies a certain two-facedness. Just read what he says, without reading in any more. Sun loves open source and they hate Red Hat. Jonathan is very upfront about it.
Sun wants to ride out the shift to open source, and they want to end up the ones on top after the revolution is over.
Just like everyone else in the Linux business.
It’s not “smarmy” at all. It’s just plain, old fashioned, “distasteful”.
“It’s why we freed OpenOffice, elements of Gnome, Mozilla”
Say what? Elements of GNOME? Since when was GNOME a Sun project? Mozilla? Uuuuuh…not a Sun project either. Am i missing something obvious here?
Linus is basically daring Sun to release Open Solaris in GPL3.
There’s a good case to be made that Sun specifically chose GPL3 and their own license in order to keep their prized features and code from going into Linux.
With this, he’s saying “OK, we’ll go GPL3 too, and we can then use some of your code”.
After all, Sun sees Linux as their number one competitor. Most of the users and businesses they hope to attract to Solaris and Open Solaris are running Linux now.
Sun is trying to have its cake and eat it to. They want to open up their code, but they don’t want any of it going into Linux. It’s kind of hard to have one without the other.
Thank you. I thought I was the only person holding that opinion.
Linus’ comment about Sun was in the context of it being an example of a reason for switching, as opposed to the FSF babble that was being raised earlier in the thread. And I do see it as a dare. He went on to say after that he doesn’t believe Sun really intends to do it.
Frankly, I really don’t see what Sun has to gain from moving to v3. They’ve already said that a move to v3 would be a dual-licensed CDDL/GPL type thing, which means that they wouldn’t be able to backport linux code anyways, even if that was their intent. On top of which, upstream contributions to Sun require copyright assignment, something the linux kernel does not, so I can’t imagine that linux kernel devs would be willing to provide that and grant Sun the right to relicense their code at will.
I really don’t understand how v3 is being considered internally as a *business* decision. I really cannot see a gain, they’re basically giving a gift to the GPL-or-death community because any improvements made will not be transferable either to Sun or the rest of the openSolaris community. I wouldn’t be surprised if some of the devs in the openSolaris community refuse to release dual-license code, and stick to CDDL only as well. They’re not exactly happy at the prospect of GPL licensing at all. Solaris will inevitably become fragmented, even if Sun changed the terms of licensing to do away with copyright requirements (not likely to happen) and tried to pull patches upstream.
I do applaud the moves Sun has made, I don’t think anyone believed that Java would ever be opened to the extent that it has, for instance. But I can’t help thinking that the v3 talk around Solaris is more about muddying the waters for enterprise linux than actually embracing a GPL development model. The timing of “musings” from Sun about a GPL solaris coincided nicely with the controversy that started with the kernel devs drawing a line in the sand and saying “No” to v3, after all.
I think Linus has definitely given Sun something to think about here, I’m just not sure how they’ll really interpret this, or how serious Linus really is. I’m cynical enough to believe Sun won’t follow through with GPL licensing, and I’m not really sure there’s a legitimate business reason for them to do so. The one clear advantage for the linux devs, though, is if Sun does, then the patent protections that Sun attached only to CDDL-licensed code will be opened.
Time will tell, and at any rate I’m sure it will be several months after the release of v3 before Sun confirms an intent one way or the other. A lot can change between now and then.
Since when was Linux development dictated by what Solaris was doing?
I’m not aware that it is. Linus seems to be saying that Solaris on GPLv3 (and all that comes with it) might be a reason TO try to switch.
But that wouldn’t be the only reason to switch. Heck, read the rest of these comments; there are plenty of reasons to NOT switch, first and foremost being the difficulty of doing so, and that Linus still doesn’t like the GPLv3 that much… though he does seem to be slowly warming to the more recent drafts.
If the arguments FOR outweigh the arguments against, then we’ll probably see movement. But I sincerely doubt it’ll be for Solaris alone- that would be a stupid reason.
In the meantime, other than possibly trying to determine how difficult a switch would be, I don’t expect anything to really change with Linux development.
…unless the linux kernel becomes another fragmented *BSD situation.
Read the thread. Major contributors like Al Viro and Greg KH won’t be contributing under GPL v3.
Don’t try to hide your code with the GPL3.
And Schwartz said:
Hide behind the GPL3? What? No! We like the GPL3 because… because… because we want end users to feel safe that we won’t sue them over patents!
…on why the GPLv2 or later clause is not good. IImagine if Linux were licensed v2 or later, and somehow, the FSF decided to relicense to a v3, and started accepting contribution with v3 or later. That would mean there were now 2 version of Linux, one which could not accept code from the other because v3 or later code can’t be used in the v2 or later kernel.
I think Linux is being smart about this. There is no pressing need to move to a new license right now, but I do think that a creation of a non profit organisation responsible for the licensing of the kernel will be a good idea. Perhaps one thing they could do is to get some Joint copyright status with contributors so that relicensing the kernel can become a bit easier if required.
On the other hand, they do have to be sensitive to what contributors think about relicensing because they might just discourage some people with strong views. However, some pretty big projects get by fine e.g. the BSDs where you can take code and use it pretty much wherever you wish, so that might not be as big a problem as could be imagined.
It seems he’s very peeved Linux wouldn’t be able to use ZFS if Solaris is ever licensed as GPLv3. According to his mail, it is very hard to move the Linux license to GPLv3 from GPLv2.
It would appear then, via the effect of Microsoft’s attempt to attack FOSS via patent saber-rattling and the GNU/FSF response to that threat, that the Linux kernel with its insistence on “GPL v2 only” licensing has backed itself into a very uncomfortable corner.
A licensor permitting distribution of a licensed entity under an updated (and therefore different) set of terms does not imply automatic retroactive changes to the distribution terms (when said updated terms become available).
http://www.osnews.com/permalink.php?news_id=18064&comment_id=247569
Live from the Linux Collaboration Summit
http://www.consortiuminfo.org/standardsblog/article.php?story=20070…
The sections entitled: “State of Linux Kernel Roundtable Discussion — What’s New with Core Development” and “The Legal Protection of Linux – Patents and Licensing, GPLv3 and the Future” are both particularly relevant to this thread.
Finally, in respect of the apparent issue between the Linux kernel and Solaris, perhaps the future efforts in this scenario should be as per the section entitled: “How Do We Work Together as a Collaborative Community?”
Re: Turf wars.
Linus is lighting the matches for a turf war, with those “Slowlaris” like jabs. Then again, he’s an asshole and even admitted to it. Doesn’t help however when he has a following that takes everything he says without a grain of salt or forming their own opinions.
Re: Server operating system only.
Cue “Yeah right.” and a rolleyes smilie.
Re: Code snatching.
Obviously, Linus wants the interesting bits for his kernel. After all, it’s his brainchild and wants it to have (more) success. This includes wanting all the great stuff, like ZFS, and pushing the competition away. Open source is all nice, unless you have someone else standing on your turf.