Submitted by “Amid falling sales of its bread-and-butter servers and mounting pressure on Schwartz to cut more jobs and boost a stock price that’s dropped more than 22%, to USD 5.26, since early February, Sun is considering its most radical open-source move yet: releasing Solaris under the love-it-or-hate-it GPL. The move could reinvigorate Sun by putting one of its crown jewels into the thick of the open-source movement – or it could diminish the worth of one of Sun’s most valuable pieces of intellectual property.”
This is all too little too late – Linux has reached critical mass in the desktop world and is poised to be competitive with Windows by 2010. Had Sun released Solaris earlier, before adoption of Linux became so widespread (especially amongst enterprise and education centers), it’s chances would be much, much better and it wouldn’t be in the predicament it’s in now. This is the helium flash before Sun goes nova.
I tend to agree that “GPLing” Solaris won’t have a huge impact – there’s only so much room for GPLed *nixes. Shame really, I quite like Sun, and wish them well.
I don’t think it’s going to have an immediate effect, as a lot of people seem to think it will. More of a long-term effect.
You have to remember the advantages of Solaris over Linux (insert random distro.) It’s stable – you develop for it – it’s going to work. Across releases. Companies such as Nvidia have learned to love that, and individual developers might come to love it as well.
It’s also got a lot of funded development to bolster it. Sun spends an enormous amount on R&D alone, a good portion works towards Solaris. ZFS, Dtrace, etc – all the new “cool” stuff being trumpeted is a direct result of that. I’ve not really seen all that much groundbreaking development come out of the Linux camp for some time.
The system isn’t fragmented. This is one thing I worry about with the GPLing. We’ve already got Nexenta and some other OSOL distros out there, and no offense to their projects – but the last thing I want to see is Solaris become Solaris + random userland stuff, like Linux. The whole stability thing goes out the window.
I could go on for ages, but I think the key is just the fundamental differences in design/organization of the project. Right now, there are a few valid (on some level) complaints concerning Solaris.
The biggest – some people are GPL only. I don’t like GPL, personally – but I realize some people won’t touch anything that isn’t GPLd. GPL-izing Solaris will allow them to utilize the OS. Most people I speak with in the IT world know the advantages of Solaris, and of all the developers I know – the only ones who don’t develop for it/like it are the GPL-only crew. This licensing change should make it a lot easier for them to give it a go, and enjoy the stability when programming.
I agree solaris [random distro’s] will actualy distroy opensolaris growth.
Linux distro’s have the problem that new developers arnt working on their existing projects why? Because they are starting their own.
I agree there is a need for Free alternitive products like Centos and the like but solaris 10 has that.
Sun Microsystems have the means and the community to vendor trap some users. This is what they need and microsoft have done a great job. By Vendor trap I mean release some low spec sparc systems. I would buy a Solaris 10 Sparc server for 800USD like a new netra x1 @800-1000MHZ dual SATA 2 single Gigabit and LOM @ no graphics. But also on that note I wont pay double for their support either.
Linux is not in a good position on the sparc platform. ZFS everywhere is actualy bad for solaris10.
Sun Microsystems have also not seen development in embedded systems and might be taking drastic steps to improve the chances of such development.
Oh random live CD’s of open solaris have helped to some extent but are always suspect to GPL issues.
I wonder what Ben Rockwood thinks…..
That’s where Solaris has a big advantage over Linux. Linux is a kernel, not an OS so there’s no definative Linux OS to devlop for (personally I think that’s what Debian should be). Solaris, on the other hand, will always be the definative version of Solaris. A lot of people’s problems with Windows come from having a load of userland stuff piled on top. There’s always userland crud breaking stability. That’s OK as long as there’s a stable core to develop for.
There’s always userland crud breaking stability. That’s OK as long as there’s a stable core to develop for.
In Linux land that stable core would be the LSB. So, I don’t think stability would more of an issue in Linux than it is in Solaris.
The interesting thing is what is going to happen with driver development. With Solaris going GPL there would be one more company to put pressure on hardware vendors to open up their specifications. That would be good for all free OS:es.
Another thing to hope for, if Solaris goes GPL, is that Sun starts to keep the stuff in /usr/sfw up to date and perhaps add gcc as a compiler. Sure Sun compilers generates better code, but a lot of free software out there requires a lot of tweaking to be compiled by it.
Sun has provided gcc since the FCS Release (3/05) of Solaris 10:
SunOS helios 5.10 Generic_118833-36 sun4u sparc SUNW,Sun-Fire-V210
helios:/export/home/escuer $which gcc
/usr/sfw/bin/gcc
helios:/export/home/escuer $gcc -v
Reading specs from /usr/sfw/lib/gcc/sparc-sun-solaris2.10/3.4.3/specs
Configured with: /gates/sfw10/builds/sfw10-gate/usr/src/cmd/gcc/gcc-3.4.3/configure –prefix=/usr/sfw –with-as=/usr/sfw/bin/gas –with-gnu-as –with-ld=/usr/ccs/bin/ld –without-gnu-ld –enable-languages=c,c++ –enable-shared
Thread model: posix
gcc version 3.4.3 (csl-sol210-3_4-branch+sol_rpath)
I suppose someone will start complaining about the “ancient” version, but it is there. All you have to do is add /usr/sfw/bin to your PATH.
it’s too bad really, it seems like that has happened too many times to Sun: eg., with Sun Java’s restrictive policy (even quite recently) toward Linux vs. Microsoft’s burgeoning .NET, and even Mono. They could have released both Java and Solaris a long time ago and come out strategically better.
BTW this quote from the article is so melodramatic and biased; nice way to portray the FSF as irrelevant when in fact Moglen was there really for the GPL3 process (which is now ending), and is going to the FSF-affiliated SFLC as well as teaching at Columbia again, instead:
“Further muddying the waters is the chaotic state of affairs at the GPL license administrator, the Free Software Foundation, which is struggling to write a new version of the license and has taken fire from Torvalds for threatening the intellectual property of companies that might use it. On Apr. 25, Eben Moglen, who ran the foundation’s legal affairs, resigned.”
Yet I thought the whole thing was to license OpenSolaris under GPL3 instead of GPL2, so Linux CAN’T take ZFS, Dtrace, etc changes into it.
Edited 2007-05-01 00:01
I agree about it being “late” but disagree about it being too late. More later.
This kind of commentary has been made for years, replacing 2010 with X where X == random_date_in_future(). Also, Solaris isn’t exactly the desktop OS of choice, regardless of license. It’s a workstation/primarily server OS.
Wow, quite the pessimist, are we? While I think Sun should have gone forward with open-source plans at an earlier point in time, I don’t think it’s over/done with. They still have a lot of traction in the government, and with major enterprise companies. You might be correct about the desktop – I personally don’t feel UNIX (or clones) fit well as “normal people” desktop OSs – but this applies equally to Linux. I don’t know where these dates keep coming from, but they are nothing more than conjecture.
If any UNIX/clone OS has a chance at being “competitive” with Windows in the future, it’s OSX. The whole “has to be free” deal isn’t earning Linux any friends at the moment – even if the fault lies with HW manufacturers/software manufacturers.
The one thing Solaris has going for it that Linux doesn’t is relatively rigorous testing and QC. I’ve yet to have Solaris N+1 break on me because of bugs introduced between RC and gold releases. Can’t say the same for Linux.
Wow, quite the pessimist, are we? While I think Sun should have gone forward with open-source plans at an earlier point in time, I don’t think it’s over/done with. They still have a lot of traction in the government, and with major enterprise companies.
I agree with you, ormandj, and you’re right about the date being more or less random (at least I’m not spouting that “2007 is the year of desktop linux!” rhetoric.) I work at an educational institution and we use a mixed Solaris/Linux (RHEL4) environment on our servers. What Sun has going for it is hardware support, quality control, and enterprise features. I don’t expect Solaris to just up and die anytime soon (thus helium flash), but all their licensing shakeups over the last few years seem to be a desperate “HEY, LOOK OVER HERE!” cry for attention. They’ll be competitive in the HPC, SAN, etc arena, certainly, but not as much as they could have; and GPL’ing Solaris isn’t going to make inroads into the desktop arena. Solaris is powerful but the learning curve compared to linux is steeper and unless that’s resolved we’ll see LAMP and not SAMP servers. I may be narrow-minded here, but as far as taking over the current linux marketshare goes, this is too little too late.
I say meh to OSX as a chance for competition. Apple isn’t showing any real interest in offering it on non-apple hardware and they seem to enjoy their “hot and sexy alternative lifestyle” image. The notion that everything Linux related should be GPL and FOSS is a falacy of idealists who don’t live in the real world.
Every year I hear this prediction, and every year I see the Linux world fall flat on its face. It will be several years before Linux becomes a viable mainstream desktop. Until game companies and manufacturers widely support it, it will remain a niche platform.
Because every year people (fanboys) seem to think Linux will explode like a stripper from a cake and that the entire world will be on it like frat boys on a keg. Linux of course wont be mainstream for years. I said competitive – if major manufacturers are offering it as an alternative to Windows, I’d call that competitive. If more companies make a move like Dell (might) towards Linux, more hardware companies are likely to consider Linux support. Linux based desktops are already widely used in education and eye candy via Beryl/Compiz/XGL/etc are getting Linux much needed press. I fully agree that without the gaming industry linux will never ‘hit it big’, but I wouldn’t call it a niche platform: it’s extensively used as a desktop, server, and embedded OS in a variety of areas.
It’s not too late for Sun. A new battle is emerging between Linux and Solaris, and it’s going to fun to watch. Let’s forget about the desktop for a moment. Desktop readiness has more to do with buy-in from hardware vendors that anything else.
Linux has made incredible headway in the entry-level server market. But in more demanding environments, Linux has three main problems: binary compatibility, enterprise storage, and RAS. Solaris is losing traction in the high-end space and needs to seek refuge in the midrange. But in emerging markets, Solaris has a problem with limited mindshare.
Sun sees these strengths and weaknesses as complementary, and therefore all signs point to a near-term marriage of a Solaris foundation with a Linux-style userland. It will be Debian-based, similar to the promising community project NexentaOS. The result will look and feel like Linux, with APT package management, a GNU toolchain, and a Linux syscall layer. But it will also support Solaris packages, Sun’s toolchain, and a Solaris kernel. Although it might only support Solaris drivers initially, it will provide a stable driver ABI. It will also feature ZFS, DTrace, and Zones.
Still think it’s too late for Sun? Linux might have a short-term advantage on the desktop due to driver support, but Sun can quickly gain marketshare in the bottom half of the server market with a free software platform that combines the best of the Solaris and Linux in one product. The relationship will help Linux nearly as much as it helps Solaris, as the work that Sun puts into the Debian-based userland will easily translate into Linux-based Debian distributions. If you want a bleeding-edge desktop, go with Linux. If you want long-term stability and enterprise features, go with Linux. Neither route has much in the way of lock-in, and migrating requires minimal retraining of IT staff.
The ensuing war between Sun and IBM at the kernel level will be the most interesting part, and we’ll all benefit. Sun will start in the lead, but IBM has enormous resources to throw around when the doing gets tough. Intel and HP will be caught in the middle. I’m not sure which way they’ll fall, and they might play an important role in this next generation of the great UNIX wars.
Sorry, I’ve been using Linux since about 1996. Every year I hear this prediction, and every year I see the Linux world fall flat on its face. It will be several years before Linux becomes a viable mainstream desktop. Until game companies and manufacturers widely support it, it will remain a niche platform.
Your choice of words, while amusing, is merely a flash in the dark without some substance behind it.
“This is all too little too late – Linux has reached critical mass in the desktop world and is poised to be competitive with Windows by 2010.”
Haha, right, I don’t see that much which has changed since back in 1998 or whatever, and I doubt much more will happen in say 10 years, so maybe somewhere around 2020?
An OS needs good apps to be succesful, and Linux still lacks severly, and I don’t see that change in 2.5 years time …
Linux has reached critical mass in the desktop world and is poised to be competitive with Windows by 2010
hahhahahahahahhaha! Rich!
I love these random Pulled-From-My-Ass dates people come up with.
I use Debian a lot, so I’m not linux hater, but this is truly amusing.
The fact of the matter is, ‘linux’ /already/ *is* ‘competitive’ with ‘Windows’. It just doesn’t have a vary large mindshare nor a large market share. And probably won’t for awhile yet. It will slowly increase, just like Apples and MS will slowly errode. *very* slowly.
… and please mod me down as well as the other guy; Since we are both pretty off topic. Me a bit more so than him
wow, I need to proof read…
I thought every year since 1999 it’s been the “year of linux on the desktop”.. now it’s 2010?
Very confused.
This article is full of several inaccuracies. Not only is it wild speculation based on a post made by Schwartz a *year* ago; it is little more than pandering by a media press member for web traffic.
A few examples of inaccurate and complete false statements from the article:
This is *not* true. CDDL licensed code does not give any special rights to Sun. Signing a joint-copyright agreement such as the Sun Contributor Agreement is the *only* way to give any special rights to Sun.
Releasing code under the CDDL does NOT give Sun any more rights to it than anyone else receiving the same code under the same license.
Also, the article says that:
This statement is also incorrect. The GPLv2 would not work, only GPLv3 due to specific patent rights that need to be granted that ZFS and other technologies Sun has are to be used.
The GPLv2 has several holes and that is why even the FSF is releasing a new version to fix them.
Not only that, even the article itself quotes a Sun member saying that it is very unlikely given that many of the 30,000 people that have worked on or contributed to Solaris are against it.
Edited 2007-05-01 00:06 UTC
Yes
This is *not* true. CDDL licensed code does not give any special rights to Sun.
Re-read the article. It does not say that the CDDL is responsible for giving rights to Sun on any code you write regarding OpenSolaris. It’s by virtue of Sun holding the copyright, which begs the question why the CDDL was even necessary, if not to purely make CDDL code incompatible with code licensed under other open source licenses – specifically the GPL.
Signing a joint-copyright agreement such as the Sun Contributor Agreement is the *only* way to give any special rights to Sun.
Which you have to do to contribute to OpenSolaris, which means that the article is correct.
This statement is also incorrect. The GPLv2 would not work, only GPLv3 due to specific patent rights that need to be granted that ZFS and other technologies Sun has are to be used.
Sorry, but that does not make the statement in the article incorrect. If ZFS and various other things were under the GPL then they would show up in Linux and there’d be a lot more cross-pollination – which is what Sun needs to keep Solaris development and ideas afloat. That’s the way the open source concept really works. As it is, one is entitled to ask why on Earth Sun felt the need to patent various things that made ZFS and Solaris code far less portable.
The GPLv2 has several holes and that is why even the FSF is releasing a new version to fix them.
The GPL version 2 is a perfectly reasonable license, and the FSF is trying to improve the GPL in the ways it sees fit. That doesn’t mean that the existing GPL version 2 has any holes at all in it – and none that are relevant here.
Not only that, even the article itself quotes a Sun member saying that it is very unlikely given that many of the 30,000 people that have worked on or contributed to Solaris are against it.
OK. The relevant part of the article is this:
“The license was “perfect for Java,” since its requirement that users republish their modifications to Java’s source code prevents proprietary versions”
Which apparently is OK now for Java, when it most certainly hasn’t been in the past. However, when it comes to Solaris the tune abruptly changes:
“More than 30,000 programmers have worked on OpenSolaris projects, and things won’t change without their say-so, Phipps says. “They’re pretty skeptical about using the GPL,” he says, “no matter how enthusiastic Jonathan is.””
So how are the positive effects of the GPL OK for Java, namely the reasonable prevention of proprietary versions, and somehow not OK for OpenSolaris?
WRONG. Read page three of the article:
Sun holding the copyright does not mean a bloomin’ thing. If you modify Solaris code, your contributions are your own unless you give Sun joint copyright ownership via a contributor agreement.
Wrong again. The article claimed that the license itself caused changes to “revert to Sun” — which is wholly incorrect.
Also, you do NOT have to sign an SCA to contribute to OpenSolaris. Some third party contributions can be integrated without one and have been if they have an appropriate license.
Sun already is getting cross-pollination from FreeBSD, other BSDs, Mac OS X and more. Linux is the only left out in the cold thanks to their license.
Just as the BSD was the right license for BSD, and GPL the right license for Linux and Java, the CDDL is the right license for OpenSolaris. No one license is right for every project.
WRONG. Read page three of the article:
On the other hand, OpenSolaris’ Common Development & Distribution License has been criticized since the rights to any changes users make to Solaris’ code revert to Sun (see BusinessWeek.com, 11/14/06, “Sun’s Surprising Openness”).
It’s worded badly, but it does not say that the license is responsible for that. It specifically says ‘Solaris code’ (look at the referenced article). The CDDL is then used to keep that code within the domain of OpenSolaris due to various license incompatibilities.
Sun holding the copyright does not mean a bloomin’ thing. If you modify Solaris code, your contributions are your own unless you give Sun joint copyright ownership via a contributor agreement.
Wrong. Sun holding the copyright means everything, and for Sun fans to continually try and paint over this is utterly laughable. When you contribute to OpenSolaris you have to transfer copyright of your code through the agreement. It means that they can take your code, put it into proprietary products and change and modify it. Congratulations, you work for Sun for nothing.
Wrong again. The article claimed that the license itself caused changes to “revert to Sun” — which is wholly incorrect.
Wrong. The article did not say that the license was responsible. You didn’t read it correctly, nor the quoted article.
Also, you do NOT have to sign an SCA to contribute to OpenSolaris. Some third party contributions can be integrated without one and have been if they have an appropriate license.
The effect is the same. Code gets integrated into Solaris that Sun can do whatever it wants with, without Sun showing the same courtesy to you. That is the biggest criticism of OpenSolaris as a project, and the CDDL allows that to happen because any code transferred must be licensed in a way that allows that to happen.
Sun already is getting cross-pollination from FreeBSD, other BSDs, Mac OS X and more.
Sorry, but Darwin is an extremely dead project and the BSDs have amounted to nothing in comparison to Linux’s usage. In the open source world, if it happens first it happens on Linux.
Linux is the only left out in the cold thanks to their license.
ROTFL. Dude. Have you seen how much code makes up Solaris’ userspace that is licensed under that license, and how Sun actually benefits from that being the case?
If it happens first it happens on a Linux system generally. Look at HAL and various other things. I also laugh at that statement, because the GPL in Linux solves a dozen criticisms of OpenSolaris. Namely that Sun can effectively use you as a free employee and use the CDDL license incompatibility to stop your code from getting out and using share and share alike licenses like the GPL. No one can do that with Linux because everyone must put their cards on the table – even competitors.
Just as the BSD was the right license for BSD, and GPL the right license for Linux and Java, the CDDL is the right license for OpenSolaris.
Why? I believe that was my question originally, because the GPL certainly wasn’t the right license for Java before. You also don’t grok how important the concept of “I modify this piece of code and everyone benefits, someone else modifies te same code and I benefit” nature of the whole thing. Not a lot of people do, surprisingly.
No one license is right for every project.
*Shrugs shoulders*. Cop out statement.
Edited 2007-05-02 18:48
English depends upon wording for meaning. In this case, the statement in the article is wrong regardless of its wording.
If I make a derivative work of copyrighted material, the only use allowed of it is under the agreement which I am subject to. Making a derivative work of Solaris code does not given Sun any more rights than anyone else receiving the same code under the CDDL. This is the same regardless of license. No, they cannot take any derivative works and put it into proprietary products unless a joint-copyright agreement has been signed.
If that was true, by your logic, any modifications I make to Solaris code that is under the GPL would also be able to be incorported into proprietary products. We all know that is NOT true.
You apparently are not aware that the CDDl is merely an updated version of the MPL (Mozilla Public License) and is not special nor geared towards Sun. In fact, it is a generic, reusable license recognized by the Free Software Foundation as a “Free Software License” and classified as “copyleft.”
By your own statement it was worded “poorly,” so I fail to see how I can not call it wrong.
WRONG again. Sun can only do what the license the code is under gives them the rights to do. Otherwise, by your logic, Sun could take any GPL code they want, integrate it, and do whatever. Stop spreading FUD.
And that isn’t?
Edited 2007-05-02 18:54
English depends upon wording for meaning. In this case, the statement in the article is wrong regardless of its wording.
It’s poorly worded for people who skim read it ;-), but it does not say that the CDDL is responsible for giving Sun rights to your code. That’s beside the point though, because if you contribute to OpenSolaris that will be the case, and the CDDL then sees to it that your code then doesn’t end up in a GPLed project like Linux – but it is in the hands of Sun. That’s the point the article is referring to.
If I make a derivative work of copyrighted material, the only use allowed of it is under the agreement which I am subject to. Making a derivative work of Solaris code does not given Sun any more rights than anyone else receiving the same code under the CDDL.
You’re making a cop out statement there. If you contribute to OpenSolaris that is exactly what will happen by virtue of the JCA. By virtue of CDDL licensed code, it also means that that code will not be able to be used in CDDL incompatible licensed projects, limiting its scope still further.
A lot of Sun people uncomfortable with this fact keep frantically pointing to the CDDL and saying “Oh, it doesn’t give Sun any special rights”. They also refer frantically to derivative works in order to divert attention from the real point. Yes, you can create a derivative of Solaris where you don’t sign your code over to Sun as part of OpenSolaris. Who fancies maintaining a whole version of Solaris though, just as you wouldn’t fancy maintaining your own entire Linux fork? At least with Linux, the project and copyright ensures that no one takes your code and re-licenses it in a way that means you get zilch back.
No, they cannot take any derivative works and put it into proprietary products unless a joint-copyright agreement has been signed.
Which is what you sign to contribute. The CDDL license is then designed to stop the code from flowing back, especially into GPLed projects (not so now with Java). The whole balance of power in the OpenSolaris world is uneven, and that’s why Sun has such a small OpenSolaris community.
If that was true, by your logic, any modifications I make to Solaris code that is under the GPL would also be able to be incorported into proprietary products.
As copyright holder Sun would be able to relicense your code as it sees fit (yes, even if it was GPLed – how do you think dual licensing works?). In the case of Solaris, the obscure CDDL license is then there to ensure that this code does not move into GPLed projects like Linux as a result either.
We all know that is NOT true.
That is true. Go and look up how this works.
You apparently are not aware that the CDDl is merely an updated version of the MPL (Mozilla Public License)
Why create a new license?
In fact, it is a generic, reusable license recognized by the Free Software Foundation as a “Free Software License” and classified as “copyleft.”
Read what the FSF thinks of it here:
http://www.fsf.org/licensing/licenses/index_html#GPLIncompatibleLic…
It is not classified as copyleft, nor are you encouraged to use it. I don’t know why you think the FSF has endorsed it.
By your own statement it was worded “poorly,” so I fail to see how I can not call it wrong.
It’s poorly worded, but nevertheless, it does not say that the CDDL is responsible for giving Sun rights to your code.
WRONG again. Sun can only do what the license the code is under gives them the rights to do.
As the copyright holder for Solaris code, and what gets accepted into the project, Sun can do whatever it wants with the code. Is it really that difficult to understand? Publicly speaking, as it is licensed under the CDDL it is prevented from being shared in incompatibly licensed projects. That isn’t an accident.
Otherwise, by your logic, Sun could take any GPL code they want, integrate it, and do whatever.
If you’re the copyright holder, yes.
Stop spreading FUD.
I’d do a bit more research if I was you, and I would also stop trying to defend what Sun did with Solaris and the CDDL. You have to sign a JCA to contribute to Solaris, have to have code licensed under very permissive licenses for it to be CDDL compatible (i.e. Sun can do what it wants with it) and they don’t want any code going into Linux or vice versa.
No, you do not. Since you are not actually a contributor the project, I doubt you actually have any experience. Read the mailing lists. This topic was just recently covered.
There is a lot of code in Solaris for which no contributor agreement was ever signed.
You are not twisting your original argument to try to make your statements correct.
The point is that the CDDL does not give Sun any special rights and that is what the original article implied, said, or otherwise indicated. Period.
We will have to agree to disagree. However, since I am actually a contributor to the project, I feel am far more qualified than a non-contributor to speak of exactly what the processes involved are.
Since I know you are not, I will leave others to judge who really knows what they’re talking about here.
There is no reason for SUN to go GPL.
1. Solaris comes with a lot of GPL-ed software already
2. GPL people are already putting their efforts in Linux
What should SUN do: create UNIX based desktop OS that runs on PC hardware. Sort of MAC OS for PC. It is not necessary to be free of charge, but it should be cheaper than MAC, and not more expensive than Windows. It should have no limitations with proprietary software and drivers written by other vendors and with DRM.
I agree.
plus sun should cooperate with adobe to create an (open) desktop enviroment + graphic tech, like the old postscrip disply, but following apple quartz based on pdf (1.7) +flash+ jdf…
and joining their forces thy can have also a nerar complete software offer from day one.
> What should SUN do: create UNIX based desktop OS that runs on PC hardware.
Yeah this AMD64 Sparc processor in my machine is really slow and expensive. At least Intel’s new CPU’s and chipsets has been released under the GPL license. I should be able to get out the soldering iron, a digital camera and some sand, and built a real Free PC and run Free OSX!
“What should SUN do: create UNIX based desktop OS that runs on PC hardware. Sort of MAC OS for PC.”
No. Sun is in the business of servers, not desktops.
“No. Sun is in the business of servers, not desktops.”
In the past, Sun delivered high performance computers for desktop use, so called workstations (Sun Sparc and UltraSparc).
Sun is comparable with SGI. They built workstations and servers, lost the workstation segment and do not have a high market share in servers, if I remember correctly. Allthough they built excellent machines (e. g. SGI Octane 2 with IRIX, or the Tezro rack servers), they lost because of the high price…
Now that they offer a free OS along with support (which you pay for) they can get popular, at least in the server world. Even their centralized workstations (Enterprise server + Sun Ray workstaions) are still great. Sun desktop systems are raraly to be found. Maybe this will change with a GPLed Solaris which attracts developers.
I hope it will be dual licensed then so we don’t have to live with the shit that GPL is and that no GPL-retards will fork it into their own GPL-only version.
Thought I don’t know how much more free their current license are… I’d much rather see something like a BSD license, sucks to switch to GPL just to make those guys happy, I hope it’s because Sun themself wants to integrate GPL-stuff they can’t use if they don’t switch license.
I think that sun would be doing themselves a favor in getting into the GPL open source game. Their software isn’t linux, and if they can get a true community around them, and not ham string their users in anyway, they might be able to get some good open-source applications written for them. And, business likes Linux not because it’s linux, but because it’s free, flexible and available. If Solaris really is better, than business will get on board. I’m a linux user, probably won’t abandon ship, but I’m more for the best tech, then just one tech.
Just assign copywrite to Linus! He knows how to run an OS project!
…or just make a better Open UNIX Desktop.
> Just assign copywrite to Linus! He knows how to run an OS project!
> …or just make a better Open UNIX Desktop.
I think that Sun makes very well UNIX Desktop if not the best one. I consider Linux another thing, because it’s not a real UNIX. Anyway, Solaris has got a long way to go to catch up with Linux on desktop area.
I recently installed Solaris (SXDE)on my laptop (first time on a real x86 hardware!). It’s fast. It’s different and it’s difficult to maintain for newcomer from Linux.
As I was saying a few times before, Solaris lacks package management tools, like Synaptics or apt. I mean it should have something like that out of the box, because efforts like Blastwave or pkgsrc on Solaris are not officially supported by Sun, neither repositories are maintained by them. And what is the crap with registering Solaris Express to get security fixes? It’s development version, damnit
. Upgrading to another build – does it have to take downloading 5CDs or depend on BFU archives (which can be risky as they say).
Another thing is that Solaris has got too much in common with Windows. I mean, management tools – java based SMC… It might be useful in enterprise environment, but in SOHO?…
My advice to Sun would be to take Debian’s way in development. Divide it into stable (Sol 10), testing (Solaris Express) and unstable where real development takes place. Put the repos on mirrors. Give people easy to use tools, apt would be fine so they could just ap-get update && apt-get upgrade to switch from build 61 to build 62 easily. They will love it, improve it and everyone will be happy. This is not the case with servers, where Solaris “stable” is used anyway, but would help newcomers. And it is realy better infrastructure than what Sun does now.
One more thing, Debian or Canonical have their official repositories, like “main”. But they maintain unofficial ones as well so that software that is not part of official distribution for many reasons, can be easily installed and updated – this would be great in Solaris too.
Sun wants to build community of users and developers around Solaris and IMHO they sould take Ubuntu’s way – they did it well.
> I recently installed Solaris (SXDE)on my laptop (first time on a real x86 hardware!). It’s fast. It’s different and it’s difficult to maintain for newcomer from Linux
Just like using Linux is a struggle for Windows users. As a daily Solaris user, it is actually piss easy to maintain. As with all things. You are comfortable with things that you know.
> Another thing is that Solaris has got too much in common with Windows. I mean, management tools – java based SMC… It might be useful in enterprise environment, but in SOHO?…
I always used to hate SMC. I starts making sense and becomes useful when you use Trusted Solaris.
Can sun do that without permission from IBM or HP?
Does sun have the right to expose proprietary SysV code? There is stull sysV code in Solaris, and sun does not own sysV, sun only licenses sysV.
Has anybody been following the scox v ibm case? Supposedly that is about scox suing ibm because – according to scox – ibm exposed proprietary sysV code by putting that code into linux. If it’s such a crime for ibm to expose the code, then why is okay for sunw to do so? Once the code is open – it’s open.
Of course, scox has no case, because scox doesn’t own sysV. SysV is owned by novell, sun only sells sysV licenses, and keeps 5% of the sales. So sun having scox’s blessing is meaningless.
“””
Can sun do that without permission from IBM or HP?
“””
It’s a little late to be asking that question since they open sourced Solaris last year under the CDDL.
This is just about the possibility of dual licensing it under GPL as well.
I believe that Sun has been careful to retain control of their OS’s IP. At least Scott said that they had. Remember that they were one of the SCO licensees, and I believe that the reason was to make sure there was no question as to their rights.
Compare that with OpenServer, which, from what I hear, was so full of third party licensed code that most of the customer’s purchase price went to paying royalties.
Sun is a very independent company, and I can’t imagine them willingly being beholden to anyone else regarding IP. Or at least, no more than they absolutely have to.
Anyway, while I’m posting, I suppose I’ll go ahead and say that just because Sun was late to the party does not mean that they cannot become valuable members of our community.
I love Linux, and don’t see myself switching. But I welcome OpenSolaris and find it odd that others in the community sometimes seem to feel threatened by it. (I’m not talking about you walterbyrd, just slightly hijacking your post.)
Edited 2007-05-01 02:47
>>It’s a little late to be asking that question since they open sourced Solaris last year under the CDDL.<<
No. OpenSolaris != Solaris.
>>I believe that Sun has been careful to retain control of their OS’s IP.<<
I believe there is still sysV in Solaris. And sunw does not own sysV. Sunw licensed that from AT&T, then novell bought the rights to sysV from at&t.
>>At least Scott said that they had. Remember that they were one of the SCO licensees, and I believe that the reason was to make sure there was no question as to their rights.<<
Except that scox does not own sysV, novell does. Why do you think that scox has to pay novell 95% of all unix license sales?
Also, sunw has signed a confindentiallity agreement. When HP, and IBM, bought their sysV licenses, HP and IBM were assured that the sysV code would be kept confidential – just as hp and ibm had to agree to keep the code confidential.
It would make absolutely no sense to sell a license to one client, and let another client open source the code. The first client would be completely screwed.
>>
Sun is a very independent company, and I can’t imagine them willingly being beholden to anyone else regarding IP. Or at least, no more than they absolutely have to.
<<
Who do you think developed UNIX? Hint: it wasn’t sun. Sun does not own sysV – and has never claimed to own sysV. Sun licenses sysV. Maybe not as independant as thought?
“””
>>It’s a little late to be asking that question since they open sourced Solaris last year under the CDDL.<<
No. OpenSolaris != Solaris.
“””
Yes. I know that. But are you implying that OpenSolaris has no SysV in it but Solaris does?
That’s extraordinary.
>>Yes. I know that. But are you implying that OpenSolaris has no SysV in it but Solaris does? . . . That’s extraordinary.<<
Yes that’s what I’m claiming, and it’s not extraordinary at all. There is no proprietary sysV in FreeBSD or Linux either, and very little in IRIX or AIX.
SysV is old. Most of the functionallity of sysV has been replaced by better technology.
It could be that sun plans to replace all the sysV in Solaris (there probabably isn’t that much sysV left by now). In that case sun could legally gpl solaris.
Or it could be that sysV is already more open than most people know, since the att vs bsd decision remains closed. In that case sun could legally gpl solaris.
i disagree – sun is in a good place.
hardware:
They reinvigorated a tired line of servers. x86 helped a lot. knowing their servers are tested and meet higher quality standards than commodity hardware is what their target markets need. driver support.
A series of good acquisitions also helped plug their previous weaknesses – storage for example.
OS:
Solaris is solid and being developed actively. Its matched to their hardware. You know is been tested when official releases come out. Its supported.
Innovation is incredible – ZFS, DTRACE, zones, smf, …
Software:
Java is still being developed and is used. its ,arket penetration is very respectable.
Large organisations with mild to high IT requirements are still enthusuastically purchasing SUN hardware and using Solaris. Its not a downward trend.
I think Sun managed the balance between open-source, community participation, transparency, protecting their IP, innovation … very well indeed. Well done.
If only more companies were liek Sun.
oh – remember OpenOffice – that too came out of Sun.
… actually started as StarOffice (from Germany).
Sun just bought into it and forked (free) OpenOffice and kept (non-free) StarOffice.
On one hand, this is bad news. Sun used to be a world leader in software development, and they were responsible for many good things. However, licensing first Java, then StarOffice, and now their OS under the GPL — bald-faced attempts to steal and profit from other’s code through the copyleft GPL — is a sign of desperation, an indication that even if Sun survives the year, the culture in Sun has changed so much that they’ll no longer be the leaders they once were.
On the other hand, this is good news for the operating system industry. By releasing Solaris under the GPL, that is putting a new *nix into an already-overcrowded field — but this one already has a notable number of users, software, and support. And because it’s licensed under the GPL and everyone in the *nix community wants to be king, this means we’ll see dozens, perhaps hundreds, of new Solaris distros. Hopefully, this will serve to increase the discord within the GPL/*nix community, and along with further education, prevent the copyleft disease from growing any further. This is just like setting a small section of the woods on fire to prevent the forest fire from destroying anything more.
Edited 2007-05-01 01:33
I’m sorry Shayna but sometimes your hatred of the GPL just becomes mindless… What is happening is showing that even big companies are having to start playing by the rules of the Free software crowd, eventually Microsoft will have to do the same. This won’t be stopped.
Will Apple have to also play by the same rules? I know Darwin is OSS but I am referring to everything on top of it which is pretty much everything the user and application developer uses.
Apple has easy to use and pretty interfaces, oh and ‘trendy’ people seem to like it.
With the rate that GNU/Linux is moving at though, OSX won’t be able to catch up either, they certainly couldn’t do it all on their own, hence they used a lot of the work done on BSD.
I know, this anti-GPL obsession borders on the irrational. I wonder what would happen if MS was to release some piece of software under the GPL…I think his head would asplode!
This would be a desperate move… Sun *hates* Linux. They’ve had a love-hate relationship for years. The Sun desktop was Linux based for a long time (is it still?) because there weren’t enough drivers to get Solaris running on x86 desktop hardware. They hated that they had to use Linux, and they were trying like mad to get it off Linux and onto Solaris. Solaris is starting to shape up on the desktop, but it’s still got a ways to go.
Linux started eating Sun’s lunch after 1999, and Sun never seemed to think Linux deserved it. But, Sun’s problems were largely self-inflicted. Pushing expensive and slow SPARC hardware on the low and mid end, where x86/Linux was just a way better alternative. They’ve got a lot of ground to make.
Based on what evidence? who is coming up with this doomsday crap claiming that the world is goign to end and Sun is going to die? this sounds like the same doomsday that is being told about Apple each year – its dying! its dying! but instead of it being a chant from the Windows fan base, its from the Linux world.
To me it looks very stupid when you have people speculate on things they don’t have a bloody clue about – and thats when you have so-called ‘analysts’ giving their 5 cents worth about an industry they’ve never been involved with let alone worked in.
Nobody at Sun *ever* said that sales would rocket and keep rocketing upwards for ever and ever – every manager at Sun said it is going to be a long hard slog to get things working; I also think it is bloody irresponsible when idiots come out claiming that the number of employees need to be slashed and yet give no incite as to how the hell things are going to be designed, produced, code written and all those other things which need to be done in the production of a product.
These analysts some how think think that you can magically cut employee numbers without having an impact on productivity. Analysts quite frankly contribute as much to the economy as money traders and housing investors (look up Adam Smith on his views of those sorts of people).
Do we really need another linux?
Once opened under gpl all the features that make solaris powerful and unique will quickly be incorperated into linux (if they dont lunge at zfs and dtrace then id be VERY surpised). Solaris is a means of selling servers, same as osx is for apple. I don’t feel the loss of IPR will affect the company value to any great degree but will increase mindshare (hopefully) and this in tur will help sales
“This is the helium flash before Sun goes nova.”
Good one!
I’ve been working with Slowlaris 10 for a month now and I can officially say its a pain-in-the-ass. The install takes forever. Setting the network is fun – not! The GUI is useless unless you like a cracked version of GNOME the company `developed` or `innovated`. CDE is a mature desktop on 10 but use it only if you must.
I would much rather use a BSD or Linux without a GUI rather than use Slowlaris. Since Oracle is moving toward a more GUI-centric environment I must adapt.
Edited 2007-05-01 11:45
I’ve been working with Slowlaris 10 for a month now and I can officially say its a pain-in-the-ass. The install takes forever. Setting the network is fun – not! The GUI is useless unless you like a cracked version of GNOME the company `developed` or `innovated`. CDE is a mature desktop on 10 but use it only if you must.
I would much rather use a BSD or Linux without a GUI rather than use Slowlaris. Since Oracle is moving toward a more GUI-centric environment I must adapt.
“It’s been widely reported Jon, and that makes it fact-equse”. –Steven Colbert, The Daily Show
I’m not sure where the name Slowlaris came from. I think the perjorative was created when Sun was switching from SunOS (BSD-based) to Solaris (Sys V), and it ate up more memory. More memory on a 8 MB Sparc 1+ was a big difference in performance.
Now adays, Solaris is plenty fast. I did some MySQL benchmarks a while back and found it pretty close to Linux, and trounced the BSDs, largely because Solaris has excellent threading code.
Solaris isn’t slow, and hasn’t been for years, it’s unfortunate the Slowlaris nickname stuck. For all of Sun’s faults, the slowness of Solaris isn’t among them.
Edited 2007-05-01 15:14
No it isn’t and nowadays it could just as well be called Speedlaris because it’s faster than SUSE 10.2, Windows Vista and Mac OS X. A reason Solaris 9 was presumably slow was that IDE DMA was disabled but that isn’t the case in Solaris 10 anymore.
I have only found Slackware and Windows XP to be faster that Solaris 10 and that’s pretty good.
Yes, its highly likely that if Solaris GPLs Solaris elements will find their way into Linux. Is this such a bad thing for Sun?
As a number of people have already pointed out, Solaris lacks momentum. It doesn’t seem to be gaining market share, and its a burden for Sun to maintain it.
So what does Sun get by GPLing Solaris? Free, or cheaper support of their ‘legacy’ OS. A potential upgrade path to a more Solaris-like future Linux. Really if nothing else, GPLing Solaris could boost their OSes marketshare. If anything companies like Redhat have proven that the ownership of the code is less significant than providing good support and service.
I’ll say this, if Sun came out with a line of professional laptops with a GPLd Solaris and Linux available pre-installed I’d seriously consider it.
> As a number of people have already pointed out, Solaris lacks momentum. It doesn’t seem to be gaining market share, and its a burden for Sun to maintain it.
Please tell me what their market share was 2 years ago, and what is it now. I would be really interested in where you get your figures from. Maybe you get them from where the Sun does not shine???
> As a number of people have already pointed out, Solaris lacks momentum. It doesn’t seem to be gaining market share, and its a burden for Sun to maintain it.
Please tell me what their market share was 2 years ago, and what is it now. I would be really interested in where you get your figures from. Maybe you get them from where the Sun does not shine???
Solaris x86 market share has grown. It’s hard not to when you go from “we’re not releasing Solaris on x86 anymore” to “Solaris is free, and open source” within a year or two.
Still, it’s not gone gangbusters. It’s hard to imagine that Sun isn’t privately disappointed with the level of adoption Solaris is seeing. It’s making some inroads, but it’s not in any danger currently of making significant encroachment into the Linux install base.
I don’t know what a lot of people are complaining about. I’m running Solaris Nevada since a week now, it’s pretty fast and stable, and most of the Linux software can compiled on it (I say most, since not all source code is really portable). Right now it runs Gnome 2.18 using hardware accelerated compiz, plays multichannel audio via OSS (which is going to be integrated into it, too, very soon), plays videos using Xvideo smoothly. And everything else I could do with Linux and *BSD.
The thing with Linux is that it has inertia, a damn lot of it, a huge chunk of it thanks to fanboys driven by their Microsoft hate. A lot of people use Linux over other Unices, just because it is Linux, not for any merit the system has gotten. The effort of getting Gentoo running is bigger than getting i.e. Solaris or FreeBSD up and running. But it also means it’s not the non plus ultra everyone makes it out to be. It’s certainly a great system, but other operating systems, namely Solaris and the BSds are great, too.
And as far as I know, it was decided last month to not go GPL.
The thing with Linux is that it has inertia, a damn lot of it, a huge chunk of it thanks to fanboys driven by their Microsoft hate. A lot of people use Linux over other Unices, just because it is Linux, not for any merit the system has gotten. The effort of getting Gentoo running is bigger than getting i.e. Solaris or FreeBSD up and running. But it also means it’s not the non plus ultra everyone makes it out to be. It’s certainly a great system, but other operating systems, namely Solaris and the BSds are great, too.
Solaris has their fanboys, and they’re no less obnoxious than rabid Linux fanboys.
That’s Sun’s biggest problem. Linux, BSDs, and Solaris are all great systems. They all have various advantages and disadvantages, but the delta isn’t big enough to initiate a wholesale change. Linux had a meteoric rise, and now it’s in a comfortable position. Something would have to really be attractive in terms of price/performance in order to start pulling a mass market. Or, like Sun did, Linux would have to blow it, and Sun would have to wait for that opportunity and exploit it.
Sun was the best thing that ever happened to Linux. In 1998, Sun’s slowest SPARC processor was roughly as fast as Intel’s fastest Pentium processor. Quickly, however, x86 caught up in terms of speed. The price/performance was overwhelmingly in favor of Linux/x86 (you’d pay 4 times as much for a system 4 times slower in 2001), so people dumped Sun for the low and medium end gear, while happily keeping Sun for the big-Iron needs. Linux momentum was well deserved in terms of what it offered customers.
Some of the comments here are about how Sun is able to distribute SysV code as part of Opensolaris. The reason they are able to do this is because they actually bought out their Unix license for quite a large sum of money. Likewise I believe HP did this as well with HP-UX.
Neither IBM or SGI bought out their Unix licenses as a result SCO threatned to revoke SGI’s unix license (for Irix) and issued a revokation on IBM’s license. Of course no body paid any attention.
Solaris is a great piece of engineering and has great features for servers (security, scalability, reliability… etc). Somewhat hampered by narrow hardware support but that could change over time.
Making it even more available will lead more folk to try it, and the numbers trying it will be greatly increased if the installation and management isn’t totally alien.
At which point the whole Nexenta project looks good: the features of Solaris under the bonnet and the Ubuntu organisation that so many are comfortable with.
The competition for the server space sure is intense.