The Register has reviewed OpenSolaris, and concludes: “Sun has made good on its promise to deliver OpenSolaris, the company’s Unix-based answer to Linux, with a company-supported, commercial update arriving in mid-May. Although far from a complete product, the latest OpenSolaris is impressive and in the long run could prove a viable alternative to Linux.”
So, we take Solaris, we plop Gnome on it, and make it look and feel just like Linux. Um, what’s the point? Is the Open Source movement going to move beyond copying to actually creating a new way to use our computers?
The point could be ZFS. Even if Sun changed the licensing so that it could be used in the Linux kernel, it really doesn’t fit in Linux. It is sort of a mix of LVM and a normal filesystem, so I doubt the Linux kernel developers would include it in standard Linux.
Another point would be that competition gives both the Solaris and Linux teams an incentive to continue to improve their work.
I doubt your doubt. ZFS is better than anything the Linux kernel has (and probably better than anything the opensolaris kernel has, too), and the Linux kernel people know it.
Not to suck on ZFS here, it isn’t nearly perfect, but it is pretty good.
Not to mention that the kernel developers also have a serious case of ANIHS (Acquired Not-Invented-Here Syndrome).
And..? – Substitute Solaris with BSD. Is it good, it is bad..?!
I personally would go with BSD at the moment. The hardware support is a lot better. I ran opensolaris on several systems and it didn’t have drivers for my network card and sound cards. I didn’t have this problem with FreeBSD 6.3 and 7.0.
When you’re using 30 or nearly 40 year old computing principles, there’s always copying to be done. Linux, Solaris and BSD are all derivative of Unix, which began in 1969. Before we can come up with new ways to use our computers, I think we need to come up with more modern ideas for how the OS uses the hardware. Something entirely new, and made from scratch. Actually, I wonder how much now-useless code is in the Linux kernel.
YES! lets invent brand new shit, to replace what we currently have developed over the past 40 years, which works pretty damn good.
I KNOW! lets all have a big f–king teletubbie doll with a touchscreen in its belly, then we can have various hotkeys located around its body. OH! i got one more stroke of brilliance going here, maybe instead of having a reset button, we just have a G-meter in its head, and if we punch it really hard in the head, it means we want a cold reboot!
YES! this will be a worthy replacement of my 40 year old /sbin/reboot.
Haha, best comment I read in weeks
Un*x is good… but not THAT good
and you propose we need to replace it with…?
I disagree. I think it actually IS that good.
All I meant was that there are 40 year old flaws clogging up what otherwise works well. 40 years ago, they probably weren’t flaws, they probably made a lot of sense. I just find it ridiculous that in an industry were hardware can change radically over the course of a year or two, we’re still using 40 year old software.
To paraphrase Linus Torvalds (and many others) – Linux IS NOT Unix. Nor is it a derivative of UNIX. Linux implements POSIX compliance, but it is NOT Unix. Furthermore, none of the Linux source code dates back to 1969. The fact that the environment is similar is more related to the fact that the environment is provided by Bash and other shells that also run under Unix.
Linux is merely Unix compatible. It is more closely related to Minix (also not Unix) than any other OS. It’s implementations are its own.
Please, don’t put fodder out for SCOG, Microsoft, and other pesky varmints.
How is the hardware support? Would I have to put a system together piece-by-piece AND troubleshoot it just to use this OS?
I’m running OpenSolaris 2008.05 flawlessly on a X48 based mainboard, C2Q, NVidia 8800GT, Intel PRO1000 and a Xonar. Mainstream stuff, more or less.
Well, the Xonar requires OSS, but still works fine.
But seriously, hardware support isn’t as good as in Linux, but unless you use archaic or exotic hardware, you shouldn’t run into much issues. The LiveCD comes with the DDU tool to check what hardware is supported or not.
There’s one huge IMO f–k up though, no Marvell Yukon chip driver. A not so uncommon onboard network chip.
IIRC Marvell is supported via a (currently in alpha) drive through the Marvell website. With that being said, always get an ‘all intel’ motherboard. If you go ‘all intel’ you’ll haveless problems.
I have it running on two machines here, Lenovo t61p and Dell Dimension 8400 (with some upgrades from when I was originally bought) – both run their beautifully. As of B90, it has been possible to boot from ZFS root – and let me tell you, the Dell has 2.5GB RAM, and the Lenovo has 4GB, and they’re screaming fast.
Edited 2008-06-14 04:14 UTC
The Marvel Yukon is cheap junk though. It’s only now that we’ve got drivers that work reasonably well in Linux with various iterations of Sky, and if you’re looking for a Gigabit ethernet device to work well under Windows, look elsewhere. The sooner the device just dies off the better really.
Not really. The secret to my success in using Solaris x86, including Solaris 10, Solaris Express and OpenSolaris is using good hardware such as Intel NIC’s, mid range ATI or nVidia video cards and a Proxim Orinoco Gold PCMCIA wireless card.
With this same hardware I can install Solaris, Windows, Linux and BSD and all of the devices will work without hunting for drivers. I ran the Live CD on my daughter’s HP Pavilion dv9000 laptop and almost all of the hardware was recognized.
Hardware, it sucks that opensolaris has really bad sparc support. and alot of the new featuers works best on x86. have a couple of old sparc but they will haveto run linux
until opensolaris is possible
I’ve never tried Linux on Sparc. I run NetBSD on my Ultra 10 and I really like it. I used to run Solaris, but got rid of it, ZFS is cool, etc., but Solaris was just too much of a pain in the ass.
So if any of you have old Sparc hardware, give NetBSD a try. It has a small memory foot print, it’s simple to use and it’s pretty fast.
NetBSD is awesome. I just wish that it had NVIDIA drivers.
Opensolaris is more of a differentiating feature for Sun to sell with their products, rather than a direct linux competitor. And this isn’t a coincidence, because this is exactly what Sun is trying to do with it.
You could say that the world is developing linux distros, but only Sun is developing Opensolaris. (regardless of how they spin the open source angle.) Sun doesn’t have the resources to compete with all the people developing linux, and even though they’re trying to share in those resources by going open source, they don’t necessarily need them. I expect that Opensolaris is already selling Sun products and gaining them positive mindshare…
It also allows third parties to point out flaws which make supporting Solaris difficult; take the latest instance where by IOCPARM_MASK has been widened which hopefully should allow OpenSound support Solaris without needing to use a custom work around for a limitation that only exists on Solaris.
OpenSource isn’t just about source availability, it is also the egalitarian nature of where by third parties can have a direct input into the operating system they support through their products. Where limitations can be pointed out publicly and fixed promptly rather than in the closed source world of where programmers hide behind a wall of PR and any requests are put on a list along with everything else (and never addressed) – Microsoft bug reports being an example of attempt to be more open and yet ignoring opensource is a key component of that openness.
You get developers you need to have a platform usable to developers. If a developer can’t even install it on their machine, or if they do install it on their machine, the experience is so painful, the coder isn’t going to have that itch scratch. Instead of the itch of curisosity and coding scratch, he (or she) will be irritated.
The Linux community plodded along, but as things gradually improved, more programmers came on board, it gained more traction and more users came as a result of that. OpenSolaris has only been open for something like 2-3 years. Given that Solaris x86 went from nothing to something, then from that something to opensource – I’m not surprised that there will be some catch up to do in terms of building community.
Its the old story that you can break a community within weeks but it takes years to build one back up. I do believe that Sun does have a long term vision for OpenSolaris – although their communication is hazy at times. What is important is that they have a long term vision now, there are people in Sun and the community who are committed to executing this vision.
I’m not trying to start a flamewar, but since you brought linux into it, I feel the need to point out that there is one substantial difference. The linux kernel does not require copyright assignment for code contribution. Although people often gloss over that fact, I firmly believe that is a substantial reason why it has garned the support it has from major contributors like IBM, HP, Oracle, Intel et al.
It’s a great thing that Sun did, opening their source to the community. I don’t want to detract from that. But if they truly want to build a community, there has to be reciprocity.
THAT is what built linux. If Linus had a requirement that said copyright had to be assigned to him while he reserved the right to relicense at will for future releases, he would never have gained the level of support he did for his project.
Making your code open and accessible to the public is a good thing, even Microsoft is starting to get that. But giving up sole ownership and making it a truly community owned project is a significantly different thing. That, I think, is the difference.
Like I said, not looking to start a flame war, and I have always applauded Sun’s new openness because one way or the other it benefits the community.
But let’s not assume that simply opening code is going to automatically lead to the same success that linux has had. Everybody from Microsoft, to Sun, to the community itself, seems to forget that linux is not owned by any one entity, it’s collectively owned by everyone that has contributed a single scrap of code to it. That’s a big difference.
Call me a little daft, but I don’t quite understand what you’re getting at. What is the point you’re making, I’ve looked through the website opensolaris.org and don’t see anything particularly horrifying – you submit your code, you declare that this is either your own code or code which you are allowed to disclose. Also according to this:
http://opensolaris.org/os/community/on/devref_toc/devref_7/#7_2_3_n…
You put both your copyright and a Sun copyright too. Again, what is horrifying about that? you still own your code whilst ‘donating’ the code to the ‘community’. Which sorry to repeat it, again, leaves me confused as to the purpose of the above – because it makes no logical sense whining over something as trivial as it.
As for HP, IBM and so forth; it started from the ground level, and the ground level got into Linux because it was a cheap, good enough UNIX clone – it had nothing to do with any so-called ‘community aspirations’. Heck, I remember back when NZ ISP’s were using it, it had nothing to do with the community – everything to do with price and ‘close enough to being UNIX’-ness.
Again, based on what? it seems to be that your arguing semantics without actually explaining why it is important. Yes, you retain copyright whilst assigning copyright to another person, what is wrong with that? lets say you contribute some code, substantial amount. In 5 years time you’re hit by a steam roller and it just so happens the project wishes to relicence to Widget version 2.0 – the whole project grinds to a halt because obviously you can’t vote one way or another from beyond the grave.
Edited 2008-06-14 08:04 UTC
Yes, well I suppose that some people that actually choose to contribute time and effort for an OSS project may take exception to granting a corporate entity free right to take ownership of their code and re-license it at will. Which is what Sun reserves the right to do, a fact that you’ve glossed over.
You’re the one that dragged the comparison to linux into this, which is why I responded. The linux kernel doesn’t require contributors to assign copyright to Linus or anyone else. It collectively belongs to everyone that contributes code, not a single entity.
Totally side-stepping my point. I don’t disagree with this statement, and nor will I even pretend that companies like HP and IBM don’t have commercial interests in mind with their contributions to linux.
I suspect you’re dismissing “community aspirations” as some sort of warm-and-fuzzy hug-fest, which isn’t at all what I was implying. When I refer to community, I’m referring to the pool of developers, since that is what OSS projects rely on for sustenance.
Linux is driven by commercial development. It’s the IBMs and HPs of the world that have poured resources into it, and helped it attain it’s current level of capability. And one of the primary reasons they do so is because of the balanced “co-opetition” that the linux OSS development model provides.
IBM can donate code and pour development resources into linux knowing full well that if HP decides to build upon that code and improve it, they can still benefit from it since the code will always remain open. No single individual or organization “owns” linux, and therefore no single individual or organization can arbitrarily decide at some point that linux has progressed to the point where they choose to close it off and proprietarize it.
Sun, on the other hand, requires code contributors to hand over copyright in order to reserve the right to license at will. Sure, the contributor still retains their own copyright over that code, but the contributor runs the risk of losing future benefit from improvements made to their own code. Not to say that Sun will do that, but whether they will or not is irrelevant, the point is that they can.
And that will prevent openSolaris from receiving the major third-party commercial contributions that linux has received.
I’m arguing semantics? You’ve implied that openSolaris will somehow catch up to linux now that it has been opened, and that it just needs time. That’s an incredibly simplistic viewpoint that utterly overlooks the various factors that went into linux development over the years.
Sun doesn’t want to lose control of their crown jewels. There’s nothing wrong with that, nor am I advocating that they should. It’s their choice to make.
But as long as they insist on retaining tight ownership of the project, it is not really an “open” project. It falls somewhere in between Microsoft’s “look but don’t touch” shared source model, and the linux de-centralized OSS development model. Nothing wrong with that.
Seriously, just throwing source code on the web for all to see and use doesn’t inherently equate to a successful OSS development model. Like I said, Sun took a big step just by opening the code in the first place, and I applaud them for it, but they won’t gain the type of community that linux has obtained, unless they’re willing to sacrifice control in order to gain developer mind-share.
In other words, the more control they try to exert over an OSS project, the narrower the field of individuals or organizations to draw support from.
It’s Sun’s choice, certainly.
I’m just pointing out that you’re equating openSolaris to linux in terms of building community, when they’re built upon completely different development models. You’re dismissing the requirement to hand over ownership of code, I’m of the opinion that it is a major differentiator.
Only time will tell.
Did you ever buy that guy the Mac you said you’d get him?
I was one of those thinking: “do we really need another distro?” when I first read about opensolaris project indiana. However, last night I actually tried it, and read some more about it, and I simply loved it. If they get the same amount of hardware support as Linux, and actually manage to fix the package management problem with Linux (i.e. no working standard), OpenSolaris can be a great desktop OS.
I will definitely keep my eyes on OpenSolaris.
One thing that’s still getting in the way of OpenSolaris adoption in the nerd crowd is the bad karma McNealy managed to amass.
You can see it on Slashdot and its ilk. In the discussions, the main argument are almost always Sun’s past antics.
That title is the worst bit of flaming I’ve read on OSAlert lately.
I am keeping Slackware on my server and Zenwalk on my laptop. I tried OSolaris and it is just OK for me now.
-2501
It seems difficult to get things like CITRIX to work on it; and hard even to get SUNS own virtual box to run on it. I mean hard compared to Linux not hard compared to Windows. So I agree that there is some more copying to do. Another couple of hairy things are graphical effects; when you turn them on and they don’t work; its hard to turn them back off.
Then there is the fairly random wireless networking.
Its so close though.
There is some good things about opensolaris. The package manager has been designed from scracth ( and his still in dev ) and brings some new ideas. However the package manager is incredibly slow.
I like the fact that audio card have soft mixing enabled with the default driver ( no need of dmix, esd or others ).