Slashdot Mirror
Torvalds on Opening Solaris
An anonymous reader sent in a link to this interview with Linus Torvalds, where the questions center on Sun's movement toward the open source world (and Linus' dismissive view of the threat posed by Solaris), as well as a few questions about 2.7 and the future of Linux.
That's just the perfect reply. If you've accomplished something great, don't be humble - that's fake - but state the facts and stay grounded.
What I don't understand is how this guy keeps himself grounded...
The owls are not what they seem
In my past life I was an SCO engineer (yeah hate me for it, but it was waaaay back)... and the more Linux evolved--disassociated from UNIX--the more I loved it. Posix/SUS was meant to be a basis for a manufacturer-neutral standard system interface. Linux kernel/gblic internals have been extraordinarily creative in working toward that rationale.
Sigs cause cancer.
I think Linus is being a bit too dismissive towards Solaris. Sure it's not going to completely crush Linux like McNealy wants to believe, but if it ends up being good enough it could slow down the growth of Linux and become a major competitor on x86.
I have recently attended a talk at our local NOVA (Northern Virginia) LUG by Harry Foxwell focused on Solaris 10. And while Harry is a respected scientist and a great presenter, I couldn't help noticing some things that were not exactly in the Open Source spirit if you will. The talk was 90% about Solaris Containers (aka Zones or N1 Grid Containers), and being a believer of giving credit where credit is due, I was somewhat disheartened not to hear ony mention of FreeBSD jails and several statements about how Solaris Zones are primarily based not on any OSS work, but rather prior Sun work on Trusted Solaris. While I believe the Trusted Solaris stuff was partly true (in Linux this is called capabilities, BTW (POSIX 1003.1e/1003.2c)), it wouldn't hurt to briefly mention the origins of the concept of separation, FreeBSD jails, and the fact the Linux Vserver provides the same functionality for Linux (Linux Vserver was mentioned, followed by some condescending analogy of Linux and transformer robots and how Linux developers can "transform" Linux into supporting anything.) The truth of the matter is that FreeBSD jails appeared in 1999, Linux Vserver in September of 2001 and Solaris Zones in 2002. The talk could also use less of "Solaris is for real, Linux is not" comments, especially considering this is a talk at a Linux User Group.
The bottom line is - I salute Sun open sourcing Solaris, but they still need to work on improving the attitude towards other open source OS's, particularly Linux and FreeBSD. The strategy of insisting that Solaris is just better, isn't going to get Sun very far, simply because it isn't true in many respects.
Just a paragraph below your quote. And it's not like he dictates every move for Linux. If there are (and on Solaris/SPARC there sure are) better things in Solaris than Linux, I'm sure he'd welcome any improvement suggestions.
That doesn't take away it's downright arrogant to say you don't need anything the competitors might offer, and really, any other commercial-software spokesman would be modded +5 Ignorant.
- Leon Mergen
http://www.solatis.com
Why does Solaris have to be seen as a "threat"? A threat to whom, exactly?
I don't understand. It still baffles me whenever this kind of mentality gurgles up, like when Jimmy Wales said that Britannica would be "crushed out of existence" as if that should be one of Wikipedia's goals. What, is Britannica somehow a net negative on the world?
Come on, get a grip.
Another one bites the dust
So is making stupid judgement calls without reading the article.
From the article
People who know Solaris better than I do will tell me and other people about the great things they offer. To try to figure it out on my own would be a waste of time.
Linus doesn't know the OS or the codebase and plans to leave the analysis in the hands of the experts.
He's busy. There's nothing arrogant about that.
Agile Artisans
That doesn't take away it's downright arrogant to say you don't need anything the competitors might offer, and really, any other commercial-software spokesman would be modded +5 Ignorant.
He doesn't say he doesn't need anything from Sun Solaris. Instead, he says that he doesn't think Solaris has anything left worth taking. He goes on to say, from the article, "But more importantly, if I'm wrong, that's OK. People who know Solaris better than I do will tell me and other people about the great things they offer. To try to figure it out on my own would be a waste of time." Yeah, a lot of arrogance there.
Solaris/x86 is a joke, last I heard. (It has) very little support for any kind of strange hardware. If you thought Linux had issues with driver availability for some things, let's see you try Solaris/x86. (Editors' note: Drivers enable an operating system to communicate with specific hardware such as a video card or network adapter.)
Oh really? I guess Linux was a joke for the longest time, then, considering its lack of hardware support. In fact, I guess it's still a joke compared to windows, if driver support is all that (apparently) matters. Why is Linus ripping on the new kind on the block for the exact problem his OS has had since its inception? This is disgusting hypocrisy.
He should be proud of what he's accomplished, and I'm grateful for his and other's work--but to take this snide attitude when another OS comes along, because it has some of the same problems his OS did originally, is pretty sad.
-Dan
"Solaris/x86 is a joke, last I heard. (It has) very little support for any kind of strange hardware. If you thought Linux had issues with driver availability for some things, let's see you try Solaris/x86."
I'm sure the comment will be taken all out of context...can't wait to see the mud slinging that ensues. Check Schwartz's blog/marketing tool for the comebacks in the next few days. Torvalds comments are mostly true for the non-server market, and Sun first and foremost is going after the server market with Solaris x86 although they are actively porting JDS to Solaris x86 which I think will run into problems that Linus mentions above.
To personally counter some of Mr. Torvalds other claims, historically Solaris x86 was a non-starter, but with a company like Sun pushing it now fully (especially on x86-64), it shouldn't be hard to find proper driver support for the majority of server installs from IBM, HP, Sun (of course), and Dell going forward. Where they're going to have to work hard is getting all the ISV's to port apps to another Unix with very small marketshare. Money always helps in that department and Sun is not shy about the fact that they have billions in the bank...so it could happen.
As for Linus' comments about Linux being superior to Solaris, I'm not so sure about that. Perhaps from a licensing perspective, and some aspect of the kernel might be as tight...but...
Solaris 10 has some neat features that don't fully exist in Linux or lack the polish that is found in Solaris.
Zones, fair share schedulers, zfs looks neat..., dtrace is amazing,
If Linux can polish up some the projects that do similar things as the above mentionned items than I think there isn't much reason to consider Solaris anymore.
I think that a lot of commentators have misunderstood Linus' answers. When Linus blandly said that he's not even going to bother studying Solaris/x86 in detail, he meant that he's not a Solaris expert. The Linux development model, by its very nature, means that any new technology of sufficient value would be easily incorporated in Linux. Linus simply meant that because he has created a dynamic atmosphere that encourages the adoption of ideas, he really doesn't need to inspect a competitive OS with a fine-toothed comb himself. If Solaris 10 does have features that get widely adopted by it's customers and proves itself over time, it would be a trivial issue to incorporate the feature in Linux (if it makes sense).
He's not dismissive of Solaris; he simply has a lot of confidence in his development model.
IMHO, of course.
Only old Koreans don't question digital signatures.
In Soviet Russa, digital signatures question you!
How can I know if those are Natalie Porman's grits, if they aren't digitally signed?
You forgot "imagine a beowulf cluster of signed articles."
Netcraft confirms it, I need to read up on my Slashdot trolling phenomena, you insensitive clod!
Never had a problem with drivers for Solaris/x86. Here's a hint, and you have to use this with *BSD and Linux too, check that your device is supported BEFORE you buy it.
Did Linus say what cereal he eats in the morning, because I'm begining to have an independant thought so I need someone to tell me what to do. What do I care about what Linus things about Solaris, of what benifit is it?
"I use a Mac because I'm just better than you are."
> Solaris/x86 is a joke, last I heard.
Last I heard? Come on Linus, please don't let what could be a healthy discussion degenerate into childish name calling. Such comments are usually the preserve of those that don't have anything constructive to add.
Solaris 10 is not meant to run on grandma's Pii Celeron.
/dale
You're missing the whole point, and showing your overall ingorance of, Solaris 10. Because of what you said, I greatly doubt that you've even tried to use Solaris 10 in its intended environment and are talking just as Linus did - based on anecdotes which are greatly dated and no longer valid.
It's designed to run on modern, high-end SPARC, x86, and now AMD64 platforms. Does it run on a hacked up Intel box where the average age of the components varies between 5-8 years? Hell no, and I'm glad Sun isn't wasting resources trying to make sure it does.