Sun Considers dual-sourcing Solaris Under GPL3

foorilious writes "In his blog, Sun Microsystem's President and COO Jonathan Schwartz discusses the possibility of dual-licensing Solaris (and perhaps the rest of their software suite) under GPLv3, in addition to the CDDL, which is the OSI-approved license under which these products are already available, but generally considered to be incompatible with the GPL at some level. Though this could mean an opening of the floodgates to a lot of sharing between Linux and Solaris (among other things), it's worth mentioning that Schwartz has speculated on exciting things in the past (such as porting Solaris to IBM's Power) that we subsequently never heard another thing about."

Comments

Sharing with Linux?

[confusion]

I thought Linux wasn't going to go for GPL3, so how exactly would that sharing work?

Jerry
http://www.networkstrike.com/

Re:Sharing with Linux?

[nurhussein]

I thought Linux wasn't going to go for GPL3, so how exactly would that sharing work?

I suspect that's the reason for the sudden change of heart. They know Linux won't be able to get any Solaris tech due to Linux being stuck at GPL2, and get to score brownie points with GPL-lovers.

Re:Sharing with Linux?

[m50d]

He meant GNU/Linux, most of which will be automatically available under GPL3 once it is published.

Re:Sharing with Linux?

[david.gilbert]

They know Linux won't be able to get any Solaris tech due to Linux being stuck at GPL2, and get to score brownie points with GPL-lovers.

I agree. I think that Linus has made an important strategic error here (in not trusting the FSF to "do the right thing" with future versions of the GPL). Oh well, nobody's perfect.

Re:Sharing with Linux?

[CastrTroy]

The thing is, you could still get a lot of interesting tech from the solaris OS without necessarily taking anything from the kernel. Remeber, that Linux is simply a kernel. It doesn't require that all software run on top of that kernel be run under the same license. If they simply release the Solaris kernel, it probably wouldn't have meant much to Linux, because Linux already has a pretty good kernel, and I'm pretty sure they'd be a little incompatible anyway. I think the main thing that will help is the applications that run on top of the kernel, that Sun may be releasing.

Re:Sharing with Linux?

[ironman_one]

Could not agree more.

Re:Sharing with Linux?

[twitter]

I thought Linux wasn't going to go for GPL3, so how exactly would that sharing work?

Probably as well or better than the kind of sharing that puts OpenSSH into every Linux distro.

Imagine Debian on UltraSparc with a Solaris kernel.

Imagine Sun Linux kernel modules. You don't really think a practical person like Torvalds would turn any of that down do you?

User name, "confusion", is way too obvious. Try "silly" or "wrong" for greater stealth.

Re:Sharing with Linux?

[Anonymous Coward]

Well duh. GPL v2 and GPL v3 are of course compatible. Share all you want.

Re:Sharing with Linux?

[Anonymous Coward]

GPLv3 allows one to tack on various provisions which would make it GPLv2 incompatible.

Re:Sharing with Linux?

[Ed Avis]

Imagine Debian on UltraSparc with a Solaris kernel.

I thought Nexenta (Debian) GNU/Solaris already provided that... or are they doing Intel only?

Re:Sharing with Linux?

[Julian Morrison]

This is actually what i posted about before. Any project that licenses GPL2 is going to feel an increasing pressure to go GPL3. Some of them will just be assimilated by the "...or any later version" suggested language. Some, like Linux, which are GPL2 only, will start to look like isolated islands of ancient code, shut out from all the modern goodies.

Re:Sharing with Linux?

[psycho8me]

There is no Debian project which uses the solaris kernel.(yet) There are a few third-party projects, but they are no more Debian than Ubuntu is. There is a large group within the Debian project which doesn't believe the opensolaris license to be free according to the Debian free software guidelines, unless this assessment is changed or solaris is relicensed it can never be a part of Debian.

Re:Sharing with Linux?

[Haeleth]

He meant GNU/Linux, most of which will be automatically available under GPL3 once it is published.

Just about everything but Linux, indeed. So what he actually meant was just GNU, and not Linux at all.

When people start saying "Linux" when they really mean "everything that is usually used with Linux except for Linux itself", I start to wonder if maybe RMS has a point.

Re:Sharing with Linux?

[Anonymous Coward]

It's not even worth asking the question, because it's not going to happen. Sun has played this game before. Remember the PR about opening up Solaris -- "we may put it under the GPL". Hooo... yeah sure Sun. You'd use the GPL rather than think up your own bullshit license to wall of the code? And and sure enough... the CDDL license came to be.

As with anything Sun-related, I suggest not even bothering to think about it until they actually do it. It saves time and effort.

Re:Sharing with Linux?

[squiggleslash]

It's an understandable one though. I, personally, can't see myself releasing software under the GPL with the "or any future version" add-on.

What I can see is similar language that makes such an upgrade allowable on condition of my approval (eg both the FSF and me, Squiggleslash, have to approve of a license, but once both of us do, the software becomes licensed under it, and that applies both to code I've written and stuff submitted to me. The license must be a newer version of the GPL.) That would make it easier to propogate the most important (IMO) Free Software license (and thus aid interoperability and license compatability) while providing an additional check to make sure anything going wrong at the FSF doesn't do damage.

That's relatively novel though. What Torvalds should, at the time, have considered was an FSF-style "If you want to contribute to the core version of Linux, you must assign copyrights to me" policy, would would have side-stepped this all to begin with, and also allowed him more flexibility in terms of ensuring the code can be used with other code licensed under GPL-incompatable licences.

Re:Sharing with Linux?

[Sr. Zezinho]

What Torvalds should, at the time, have considered was an FSF-style "If you want to contribute to the core version of Linux, you must assign copyrights to me" policy, would would have side-stepped this all to begin with, and also allowed him more flexibility in terms of ensuring the code can be used with other code licensed under GPL-incompatable licences.

That would have turned him into a perfect target for all kinds of legal assaults... Right now there is no reward in attacking him personaly. If he held all the copyrights there would be a legal single point of failure. Not what we want, is it?

Re:Sharing with Linux?

[arose]

People contributing to Linux would be more likely to contribute to Solaris as well. Nothing prevents people to license their code under GPL2 or later no matter what most of Linux is licensed under.

Re:Sharing with Linux?

[squiggleslash]

I'm having difficulty understanding your logic. Would you explain it to me?

The scenarios I'm reading into your comment are:

1. Torvalds decides he hates the GPL, revokes it and switches to closed source. In this instance, everyone uses the last GPL'd version and forks the operating system.

2. Someone sues Linus for copyright infringement. While the case is bogus, Torvalds settles out of court, agreeing to no longer distribute Linux. Everyone uses the last GPL'd version and forks the operating system.

3. Someone sneaks code into the operating system that they didn't own the copyright to in the first place, and sues Torvalds. This is exactly the same scenario as would happen anyway (eg if IBM lost its lawsuit to SCO.)

How would Torvalds be a single point of failure? The principle of forking the last known Free Software licensed version is well established in instances where the licencer has refused to release Free Software future versions, either deliberately (AmiTCP, Gosling EMACS, etc) or through circumstances beyond his or her control (AtheOS.)

The worst is it kind of works the other way. As it stands, Linus Torvalds would have more difficulty than most enforcing the copyrights on Linux, as he would have to prove he has standing to sue. This was the original reason why the FSF recommended people assign copyrights to the project maintainer - because if you own the copyrights, you don't have to cross that hurdle.

Re:Sharing with Linux?

[SpinJaunt]

Hey, sorry you are wrong.. it is:
GNUv2/Linux-2.6.16-rc1-git4

Yours,
  Richard Matthew Stallman/Linus Torvalds

Re:Sharing with Linux?

[salimma]

The interesting things in Solaris are kernel-related (DTrace, zones and ZFS), I believe.

Re:Sharing with Linux?

[Lehk228]

microsoft gains some sort of leverage against Torvolds, something big, and gets the copyright turned over to them. then MS can release a closed source version of linux.

Re:Sharing with Linux?

[justins]

Hah. You've got to love an anti-Sun sentiment so strong that GPLing software suddenly becomes a bad thing.

Re:Sharing with Linux?

[Brazilian Joe]

Linus vetoed converting *the whole kernel* to GPLv3, but also stated that a few portions of it could be under GPLv3. Likewise, once Solaris is under the GPLv3, it might be able to incorporate GPLv2 code, unless v2 is incompatible with v3.

Re:Sharing with Linux?

[CFrankBernard]

Would the license affect ZFS so that Linux and DragonFly BSD couldn't use it?

Re:Sharing with Linux?

[tonyr60]

Don't imagine it. Just do it.... http://www.opensolaris.org/os/community/brandz/

Re:Sharing with Linux?

[Anonymous Coward]

Okay, so I just skimmed it.

Now, which part do you find objectionable? It seems to me that if you agreed with the principles of GPLv2, GPLv3 would be a logical step. If anything, some of the protections are now weaker: for example automatic termination of license for GPL violators is now non-automatic. (Though AFAIK I've never seen the GPLv2 text enforced despite many violators.)

Re:Sharing with Linux?

[salimma]

DragonFly BSD lead Matt Dillon is planning to integrate ZFS support. The core kernel being BSD-licensed, integrating files with CDDL (or GPLv3) licenses are OK. The problem with including ZFS support in Linux (unless Sun, being the copyright owner, dual-licensed ZFS under GPL2) is not allowed because GPLv3 is likely to be more restrictive than GPLv2, and thus GPLv2 would disallow the ZFS code from being linked against the kernel.

Re:Sharing with Linux?

[Decaff]

I suspect that's the reason for the sudden change of heart. They know Linux won't be able to get any Solaris tech due to Linux being stuck at GPL2, and get to score brownie points with GPL-lovers.

So you really think that a large corporation would put time and effort into adding licenses to Solaris to snub Linux and to score brownie points?

The reason for the current licensing of Solaris is nothing to do with preventing technology transfer. After all, Sun have transferred plenty of technology to competing operating systems - Java, NFS, work on GNOME.

It is because of patent issues in Solaris.

Re:Sharing with Linux?

[Anonymous Coward]

That line clearly refers only to the kernel (Linux), as the version number implies, so GNU/ should not be used.

GNU/Linux should be used when you're talking about the GNU operating system using the Linux kernel. If you replace Linux with the kernel from FreeBSD a name like GNU/kFreeBSD is appropriate (kFreeBSD is short for kernel from FreeBSD; like GNU and unlike Linux, FreeBSD is a complete operating system). However, if you use the kernel from the GNU project just calling it GNU is enough.

Re:Sharing with Linux?

[sparkz]

In what way?

My project is GPL2; yours is GPL2 "or any later version".

My project can still use your code; your project can still use my code (under the terms of GPL2)

Re:Sharing with Linux?

[OoberMick]

Because in a few years no one is going to release code under "GPL2 or any later version". They are going to release it as "GPL3 or any later version". That's when the fun really starts.

Re:Sharing with Linux?

[Anonymous Coward]

Linus Torvalds does not own the copyright to the Linux kernel. Repeat: he does not own the copyright to the kernel. The kernel is composed of many pieces, the overwhelming majority of which were not written by Linus. Each author maintains copyright of their contribution so they can relicence it as they see fit. However, in order for it to be included in the kernel it has to be licenced under the GNU GPL v2, which means that others get rights to redistribute the code but always ONLY under the GNU GPL v2. Even if Linus wanted to, he could not change the licence of the kernel. He could re-licence his contribution of course so that others may use it under a different licence but the licence of the code in the kernel would still be GNU GPL v2. Anyway, such a thing would not matter much, there is probably little of Linus' code in the kernel not patched by contributors (as it should be; kernel development is an open/ever changing process). On another topic Linus does own the trademark to the name Linux, which is sublicenced on his behalf by a non-profit (see www.linuxmark.org for more details)

*Here starts a short rant*
On other interesting topics: GNU/Solaris does exist, most of GNOME/KDE are licenced WITH the "or later version" clause (which means automatic upgrade (if the licencee wishes) to GPLv3 and "Linux" is only a (good, innovative, monolithic) kernel (there is GNOME on X on FreeBSD on PowerPC, XFCE on X on OpenBSD on i386 and so on). Oh, and btw, the sky is not falling, the GPLv3 draft process has only began and an upgrade to GPLv3 is not required by GPLv2 authors if they don't want it (they can still licence with the or later version clause and effectivelly dual licence).
*Here endeth the rant*

I wonder if this will be modded up :) Deliberately left anonymous to see if anyone reads at our level.

Re:Sharing with Linux?

[sparkz]

True; just like you don't see many projects released as "GPL v1 or any later version" these days ... however, you see a lot of projects released as "GPL (with or without the v2, with or without the "or any later version" or BSD", or "GPL (ditto) or MPL", etc, etc. GPLv2 is one license; GPLv3 is another; BSD is another, yadda yadda. I think that because it's been so long between GPL license changes, some confusion is inevitable. GPLv2 and GPLv3 will have similarities, I am sure. But they will be different licenses. Personally, the code I've released under the GPL has been explicitly GPLv2. I didn't (and still don't) know what GPLv3 (or v4, v5, etc) may contain. I sympathise with the FSF's thoughts on this boilerplate text, and can see it being consistent for FSF-released code, but for many projects, the exact license terms need to be clear before releasing the code with a license saying "whatever the FSF come up with in future will be fine, too ... I'm sure."

Re:Sharing with Linux?

[Anonymous Coward]

GPL3 is a more restrictive license than GPL2, therefore I cannot use your GPL2 code in my GPL3 (or later) project. FSF are unlikely to continue offering GPL2 once GPL3 is finalized. Sure you could fork every project, but practically speaking...

Re:Sharing with Linux?

[sparkz]

GPL2 has been written; I don't need the FSF's approval to use it (unless I'm violating it, of course). Current FSF software is available as "GPL2 or later", so I can continue to ignore any changes in GPL3 and use it under GPLv2.

That is one of the good things about the GPL (and one of the bad things about the boilerplate text) - it's unlikely, but possible, that GPLv4 could something totally unacceptable, like "You also have free use of the author's kitchen". I see including the boilerplate text as providing my software under a license which I have never seen.

As for GPL3 being more restrictive than GPL2, that's a pretty broad statement for a license which hasn't been released yet. Personally I have to admit that I haven't even read the entire draft GPL3 yet (let alone put it by a lawyer)

Who cares about Solaris?

[tdvaughan]

GPL Java, for crying out loud.

Re:Who cares about Solaris?

[david.gilbert]

GPL Java, for crying out loud.

If you want "GPL Java", why not help out with GNU Classpath. Progress has been nothing short of spectacular in recent months, and more volunteers are always welcome.

Re:Who cares about Solaris?

[Anonymous Coward]

Thank you for illustrating the stupid fucks who are dragging down the open source movement.

Bunch of fucking teenage license freaks throwing tantrums.

Re:Who cares about Solaris?

[justins]

Well, Solaris is infinitely more valuable, for starters. "GPL Java" is a hot-button issue with people for some reason, but at the end of the day, it's just a programming language. Versus, you know, an entire modern Unix operating system.

Re:Who cares about Solaris?

[jcr]

So, you want to remove the last shred of hope that Java may make Sun some money one of these days? ;-)

-jcr

Re:Who cares about Solaris?

[Anonymous Coward]

Yes, but we have working Unix-compatible operating systems, most Java software however doesn't work on free Java implentations.

Re:Who cares about Solaris?

[justins]

Yes, but we have working Unix-compatible operating systems,

As far as I can tell the only important GPLed Unix-compatible operating system is Linux. It'd be good to have some redundancy there.

most Java software however doesn't work on free Java implentations.

People forget that what you describe is a completely solvable problem without Sun's cooperation.

Re:Who cares about Solaris?

[SteeldrivingJon]

Is Jonathan Schwartz a comic book collector?

I can just see him keeping the Lighthouse apps, and the rest, in mylar polybags in a longbox in his basement, along with ten copies of the "Death of Superman" issue.

Someday, those are going to be worth MILLIONS!

Floodgates are shut

[brunes69]

Though this could mean an opening of the floodgates to a lot of sharing between Linux and Solaris

Linus already said that Linux is not now, and will not in the near future, be released under GPLv3. And since GPLv3 is not reverse compatible with GPLv2 (it has more restrictions), this won't happen.

Re:Floodgates are shut

[Tim C]

I imagine that by "Linux", the submittor means "GNU/Linux" rather than "the Linux kernel".

I know, I know - Linux is the kernel, yadda yadda. When anyone I speak to says "Linux", they mean the OS, not the kernel - just like when people talk about NT, they mean the OS, not the kernel.

Re:Floodgates are shut

[TheRaven64]

When he talks about sharing with Linux, he means developers, not code. And when he says about sharing he means poaching. Quite a few people code for Linux rather than the BSDs because they like having a restrictive license which prevents their code being used by parasites. If Solaris goes GPLv3 and Linux doesn't then this may influence some of them to switch to Solaris development. If Solaris can cream off some of the better Linux developers then it could become very interesting.

Re:Floodgates are shut

[brunes69]

The OS *is* the kernel.

They rest of the stuff is just userland utilities. And frankly, I can't see why *anyone* would want to take any of the userland utilities from Solaris. Quite the oppposite in fact, Solaris has been taking all the userland from the GNU camp for a long time (GNOME).

Re:Floodgates are shut

[PsychoSid]

Each to their own. I reckon dtrace, RBAC, and zones have some aspect of usefulness. The new IP stack could be one to watch as well.

Re:Floodgates are shut

[civilizedINTENSITY]

"When he talks about sharing with Linux, he means developers, not code." Developers, developers, developers -> means applications, applications, applications. Mindshare matters, and Sun is desperate to get in the flow on a winning side. I suspect this is merely a test: running the flag up to see if people salute or blow raspberries. Does anyone know the numbers? How many developers does Sun have? How many volunteered after they formed their closed "open community"? How does this compare with the number of GNU/Linux developers?

Re:Floodgates are shut

[ClamIAm]

GPLv3 is not reverse compatible with GPLv2 (it has more restrictions)

So GPLv3 is out already? That's news to me.

Re:Floodgates are shut

[Anonymous Coward]

The difference between "Kernel" and "userland" escapes you completely.

Re:Floodgates are shut

[TheRaven64]

A while ago, Sun bought StarOffice and open sourced (most of[1]) it. Since then, everyone in the community has applauded OpenOffice, but Sun still contributes about 80% of the developer time to the project. With Solaris, I would imagine, they are keen to actually see some benefit from the costly process. They get some nice PR points for the open source release, but probably not enough to make up for the cost of releasing the code. If they could persuade some of the top F/OSS kernel hackers to put some time in hacking on the Solaris kernel, then the exercise would probably be worth it.

Incidentally, flamebait? WTF?

[1] Some bits, such as the spelling and grammar checker, were licensed and so could not be released.

Re:Floodgates are shut

[Kadin2048]

It's my understanding that dtrace is in the process of being ported under the OpenSolaris license to run on BSD. Yes, it's not FSF/GNU compatible, but as a standalone userland utility I'm not sure why that would matter to a user. (I.e. if you just want to use it, and not link it to anything, does it matter a lot?)

A future Linux version might not be able to be included in some of the more "pure" GNU/Linux distributions (Debian), but it could end up in nonfree and still be useful as an application.

The website of the FreeBSDTrace project is allegedly http://web.pulltheplug.org/dtrace/, but it's apparently down right now. The developer's blog is http://www.sitetronics.com/wordpress/?cat=8

Re:Floodgates are shut

[Tim C]

The OS *is* the kernel.

They rest of the stuff is just userland utilities.

Can the OS even boot without the userland utilities? If it does boot, can it do anything more useful than respond to pings? Ok, so it may be able to function as a router or firewall, but even if it can, how do you modify the config files without an editor?

Re:Floodgates are shut

[civilizedINTENSITY]

"Flamebait, WTF", indeed. It was because you posed valid points, and recieved a "flamebait", that I reworded (what I considered to be) your key point and posted it as a reply. Thus attempting to raise the context of your post from -1 oblivion.

Re:Floodgates are shut

[ABCC]

Linus already said that Linux is not now, and will not in the near future, be released under GPLv3. And since GPLv3 is not reverse compatible with GPLv2 (it has more restrictions), this won't happen.

Everyone on Slashdot should know that experimental features go into the -mm sources first, so that's how GPL3 will makes its way in. Lets just hope the bugs aren't too bad.

Patents in GPL3

[SWroclawski]

One of the least discussed but largest changes in GPL3 is the explicit mention of patents and how patents (if found to be violated) would effect the work as a whole. This is similar to the IBM Public License and is one of those things that I'd imagine would give a corporate lawyer warm fuzzies. Sun and others may find this change so compelling that they'd be willing to give more attention to the GPL3 than the GPL2, which strengthens it further (since these companies want the flow of information to go both).

Solaris on Power

[lcs]

http://www.blastware.org/

Re:Solaris on Power

[Anonymous Coward]

It's hardly a useable port. In fact, all they did was recompile the kernel. It doesn't even run yet.

With a bit more work, though, who knows? You've got to admire their spirit. :-)

Re:Solaris on Power

[Anonymous Coward]

Solaris on PPC isn't really new news. This was done back in the
2.5.1 time frame as a LAR (Limited Availability Release) for PPC
and PREP machines.

Source code licensees of this LAR received complete build trees
for all of what comprised Solaris as the time. Durring that
time frame there was also a PARISC version of 2.5.1 that was
in use internally. Why Sun didn't capitalize on this is left
as a mystery.

If the PPC open solaris folks want a leg up, it might be a
good idea to ask Sun for this previous work.

a race to hypocrisy

[geoffspear]

Which will happen first:

- Linux zealots abandon their "everything about Solaris sucks and I'll never use it" dogma, or
- Mac zealots abandon their "Intel processors suck and I'll never use one" dogma?

The Mac people are taking an early lead, but anything can happen.

Re:a race to hypocrisy

[m50d]

If you look back in the thread, you'll see mac zealots are already getting elitist about their new processors.

Re:a race to hypocrisy

[Cal Paterson]

Meh. I'd switch. Solaris apparantly treats everything in a more Unix way. Having hda1 instead of hd(0,0) for my first partition in Linux makes me feel strange and violated. Why does eth0 subscribe to the 0-9 order, and yet hd is on 1-10? Linus is deluded if he thinks this is socially acceptable.

Re:a race to hypocrisy

[jedidiah]

I am a Linux Zealot that's already had the priveledge of using Solaris pretty intensively in a professional setting for about 10 years both on Sparc and on x86. My "dogma" is not the problem. My 10 years of experience using Solaris is the problem. It also doesn't help my "dogma" that I've also used other commercial Unixen and don't see Sun as the whole universe.

            Between IBM and SGI, it's the Sun fanboys that have serious "dogma".

Re:a race to hypocrisy

[ABCC]

Speaking of zealots, wasn't Jonathan Schwartz ranting about how the GPL sucked at some keynote address last year? Oh wait! He did! He even wrote about on its weblog:

"such as porting Solaris to IBM's Power"

[Anonymous Coward]

Hehe, that was already done in the mid 1990's. Don't people remember CHRP or PeP? Well that goes to show slashdot readers. Also remember Windows NT For PowerPC? And OS/2 for PPC?
Any one remember OpenStep for Solaris?

Maybe I was smoking something but I clearly remember these things.

Speculation with Schwartz

[digitaldc]

...it's worth mentioning that Schwartz has speculated on exciting things in the past (such as porting Solaris to IBM's Power) that we subsequently never heard another thing about.

You can find out if you just use the Schwartz - trust your feelings, let go.

Will Sun Shine?

[Quirk]

I read in a few /. posts that Solaris is likely the best 64bit OS available. On other sites I've read Solaris referred to as Slowiris when run on a single CPU, but the Sun site suggests Solaris is no slower than Linux on a single CPU machine.

How much of a cachet does Solaris have and how will Sun attempt to capitalize on any cachet Solaris does have, especially on dual cores? Is going Open Source with GPL v3 an attempt to move into Linux territory and sell services while trying to maintain sales of their high priced hardware?

I downloaded Sol 10 looking forward to doing a dual install with Ubuntu on a Athlon 3800+ workstation, but stupidly bought a SATA drive which Solaris doesn't support. So I'll have to go back to an ide drive to do the install.

Re:Will Sun Shine?

[Zemplar]

You'll certainly love the speed of Solaris 10, with perhaps, only one exception for now. UFS filesystems don't [generally] perform as well as some other more recent filesystems supported by Linux, though you can be certain that will change when ZFS is considered production quality. ZFS is now available in OpenSolaris and [I believe] the latest Solaris Express builds. The "FireEngine" network stack on Solaris 10 is without a doubt the fastest I've ever seen. 64-bit multiprocessing is hands-down better on Solaris than any GNU/Linux system I've yet used.

I personally only hope that whatever becomes of the GPL3 deal gets Solaris 10 the credit it deserves. For mostly historical reasons, Solaris doesn't enjoy the great "buzzword" media hype that Linux gets, though it dominates Linux in many areas - geeky-techno-media-lust is not one of those areas.

For those asking for Java open-sourced, I don't see how that will help the big picture. Linux is open source and look how many different versions do things their own way. How many different binary formats must we have to run on the different Linux boxes with full dependency checking? What a mess. Some control on the direction of Java is a good thing as Java will remain consistent!

Re:Will Sun Shine?

[justins]

On other sites I've read Solaris referred to as Slowiris when run on a single CPU, but the Sun site suggests Solaris is no slower than Linux on a single CPU machine.

As far as I can tell, the "Slowaris" nickname came from having slow SPARC hardware back in the day, and having a crappy X server. The cool thing is that Solaris is using the Xorg X server for almost all hardware now. They've got an official Nvidia driver now and things are fairly snappy.

One nice thing about Solaris is that successive versions of the OS actually get faster, not slower.

Re:Will Sun Shine?

[AKAImBatman]

I read in a few /. posts that Solaris is likely the best 64bit OS available. On other sites I've read Solaris referred to as Slowiris when run on a single CPU, but the Sun site suggests Solaris is no slower than Linux on a single CPU machine.

You have to be careful here. Solaris used to be called Slowaris when run on Intel machines, because it was designed for much more powerful hardware. A lot of features that are hardware supported on a SPARC machine had to be reimplemented in software on Intel machines.

Another common vector for the "Slowaris" comments is the early days of the Sun framebuffers. Sun was one of the first vendors to do away with text mode all together, and emulate it entirely in software. The upshot is that Solaris SPARC machines have the best looking, smooth font, conole you will ever see. The downside is that the 100 MHz beasties that started this practice had a bit of trouble keeping up with the needs of the console rendering.

Neither of these issues has been significant for a very long time. I haven't heard anyone call the OS "Slowaris" in almost a decade. The complaint I hear today is that Solaris is unwieldy and not at all designed with user-friendly setup. Sun keeps trying to fix this with new, prettier installers. I don't think they have a clue though, because the first thing I have to do every time I install the OS is go into the config files and setup the DNS server and default gateway. You'd think it would kill them to ask this info during an install. :-/

Re:Will Sun Shine?

[Ekarderif]

and having a crappy X server.
"Crappy" is kinda redundant there.

Re:Will Sun Shine?

[Anonymous Coward]

Earlier versions of Solaris 2 deserved the Slowaris (Slowlaris etc) tag. The switch from BSD based SunOS 4.1 to SysV base Solaris 2, SunOS 5, caused a report 25% or more lose in performance. Much of that was probably due to the old bugs that were in SysV but had been fixed in BSD. Sun has been fixing bugs and improveing performance with every release of Solaris.

It is a little hard to do a straight Apples to Apples comparison of Solaris to Linux and *BSD. There are tradeoffs in the Solaris kernel to improve performance on large numbers of CPUs, and I/O performance. There are common cases where Solaris will out perform Linux or *BSD and vice versa.

Re:Will Sun Shine?

[C0vardeAn0nim0]

"I don't think they have a clue though, because the first thing I have to do every time I install the OS is go into the config files and setup the DNS server and default gateway. You'd think it would kill them to ask this info during an install. :-/"

well... actually, they do.

at least on x86 machines. both Sol 9 and Sol 10 asks during intall what kind of directory server you want to use (they offer NIX, NIS+, DNS, LDAP...). if you choose DNS, the installer asks the addresses.

Re:Will Sun Shine?

[AKAImBatman]

both Sol 9 and Sol 10 asks during intall what kind of directory server you want to use (they offer NIX, NIS+, DNS, LDAP...). if you choose DNS, the installer asks the addresses.

And still doesn't set them up. (Which peeved me off even more.) The last time I tried setting up Solaris 10, it asked me for the DNS and gateway info, but then failed to act upon it. Maybe it was a bug in the installer version, but it REALLY annoyed me. I had been hoping that Sun had finally taken care of the problem so that I wouldn't have to search through man pages AGAIN, trying to remember which files to edit.

I love Solaris as an OS, but is it really so much to ask that common configurations be automated? Sadly, Sun has never been too tremendously good at the whole "increasing usability" thing. They have great products, mind you, but they force you to learn a lot of stuff up front. While this is great for newbies who need to learn the system, it just annoys the heck out of the veterns.

Re:Will Sun Shine?

[htd2]

Solaris hasn't been slow relative to Linux since Solaris 10 came out and before that Solaris outperformed Linux for most SMP workloads but lost out on single CPU systems.

One of the design goals for Solaris 10 was for it to be not more than 5% slower than Linux for a range of single processor workloads where typically in the past Linux had been faster (on the same hardware). To that end Sun developed a benchmark called LibMicro which modeled the workloads which Solaris underperformed at and gave this to the Solaris 10 kernel developers as a way of measuring how far they were from their goal.

Most of the benchmarks that have been published show that they have done a pretty good job with Solaris delivering very similar performance to Linux on the same hardware for single CPU workloads and generally outperfoming Linux on SMP workloads.

There are some areas where Solaris anhilates Linux, give Solaris 10 a TCP heavy workload and it will easily outperform Linux. This may also no apply to UDP as well.

To give you an example of how close Solaris is to Linux on single CPU tests the SPECjbb results for a AMD based x2100 are 15434 for SLES 9 64bit and 16070 for Solaris 10 64bit

Re:Will Sun Shine?

[Anonymous Coward]

Besides, Linux is slowing down as the features required for scalability are added. :-)

Link to libmicro. Builds on at least Solaris and SuSE using gcc or the now-free Sun compilers.

Re:Will Sun Shine?

[thanasakis]

every time I install the OS is go into the config files and setup the DNS server and default gateway

Hey, selection of the defaultrouter and dns servers is in the installer for several years.

Solaris may have had several issues with other things, but their installation methods and processes were kicking ass when the rest of us were in the dark ages.

Re:Will Sun Shine?

[sparkz]

I'm with you on the DNS thing (for NIS/NIS+, I use the installer; I've never installed Solaris as an LDAP client, but for DNS, I go for "files" and then "cp /etc/nsswitch.dns /etc/nsswitch.files" after the install, but that's mainly because I've got an existing /etc/resolv.conf, with various options selected, which has to be installed exactly as-is, and I can't be bothered with typing the stuff in by hand.

DNS is just easier to configure that way with Solaris - in my experience, at least, but maybe that's just because I've tended to do installs for very specific requirements where I want /etc/resolv.conf to look a particular way, and I don't need DNS during the install anyway.

For NIS/NIS+, which is more complicated to configure, I'm happy to leave that to the installer.

I've used the Solaris DNS installer a few times at home, though (very simple setup, and I have occasionally bothered to type the addresses in by hand), and it's worked for me on 7,8,9 and 10.

Re:Will Sun Shine?

[xtaski]

Don't be fooled - Solaris 10 is as bad a move as Solaris 8 on Intel...

Just wait 1 yr when Sun's inability to reinvest leaves it smoked by Linux... the game is almost over for these neverending strategy changing bandits.

This is the way OpenSorce works

[ironman_one]

If Linux dont want do comply with openness there is always s.b. who will. This error from Linus is maybe the chance solaris needs to grab the initiative and mindshare of programmers.

Horses, Loaves and Shoes.

[twitter]

Who cares about Solaris?

Anyone doing any kind of scientific computing, which is a large portion of their customer base. They have been losing that customer base to Linux, which hurts their sales in more ways than one.

You might also care about Solaris if you want to use any of their excellent hardware. If they GPL'd Solaris, no only could you use it without practical and moral problems, you could also do a much better job of porting other free software.

GPL'd Solaris would be a great gift. Don't look it too hard in the mouth.

GPL Java, for crying out loud.

The magic of cross licensing may prevent that. If Sun GPL's Solaris, you can be sure they will do everything in their power to get a free Java out.

Take what it gives and make what it won't.

Re:Horses, Loaves and Shoes.

[m50d]

You might also care about Solaris if you want to use any of their excellent hardware.

*Looks at the sparcstation on his right, a way through emerging something.*

Umm, why?

Re:Horses, Loaves and Shoes.

[Dr_LHA]

You're out of date I'm afraid. Solaris already lost out to Linux in the scientific computing field a few years ago. In my field (Astrophysics) universities 5-10 years ago were 100% Solaris, with some Dec Alphas thrown in the mix. 5 years ago the exodus began to Linux machines when people realised they were faster than Solaris boxes, 1/5th of the price and could run all the same software.

Fast forward to today linux is losing out to Macs in science, every conference I go to it seems that more and more people have Powerbooks (like > 50% of the audience), especially at NASA. My project just decided to move entirely over to Macs. Solaris isn't even in the mix anymore.

Re:Horses, Loaves and Shoes.

[Anonymous Coward]

I'm also an astrophysicist, in the UK. I've never heard of anyone using macs for serious computing work, for the same reason people don't use much solaris any more - a generic linux box is faster and cheaper. But many people use Mac laptops, because they're unix (i.e. can talk to your linux desktop) and they "just work" without the messing about required to get stuff like wifi and power management working on a linux laptop.

Re:Horses, Loaves and Shoes.

[Dr_LHA]

I'm talking about in the USA, not in the UK. PPARC budgets for computing in the UK are low, so you have to go cheap. I know because I did my PhD in the UK, but I've worked in the USA for 8 years now in Astrophysics.

However in the US, where Macs are cheaper than in the UK and computing budgets are more reasonable, Macs are starting to win out. I work on a NASA mission at a University and personally have a Dual G5 workstation. The Astronomy department here is moving to entirely Macs, and my project is too (currently we only have 3 G5's in the building, but after the next budget cycle that will most likely double). Trust me ther e is a lot of serious work going on here using Macs!

The main reason for this is that Macs are easier to administer, easier to use and run Microsoft Office. Use of Office is much more prevalent in US research I find, so much so that many people I work with have Windows laptops and Linux workstations. With Macs they can do everything on one machine.

The last conference I went to as I said, at least 50% of people had powerbooks. Everyone who worked at NASA/GSFC had a Powerbook.

Re:Horses, Loaves and Shoes.

[Anonymous Coward]

You can't be serious. There isn't even a Mac version of SAS anymore.

Re:Horses, Loaves and Shoes.

[Dr_LHA]

I don't personally use SAS, and don't know anybody who does.

Re:Horses, Loaves and Shoes.

[htd2]

This is changing, we are seing more and more interest in Solaris x86.

The general feedback is that it is as fast or faster as Linux on the same hardware, you can get it for free and use it without having to go something like the Fedora route and if you do want support it is cheaper than Linux.

Re:Horses, Loaves and Shoes.

[hackstraw]

5 years ago the exodus began to Linux machines when people realised they were faster than Solaris boxes, 1/5th of the price and could run all the same software.

I work in the science number crunching field as well, and yes, this is the case. Suns have come down in price, but generic x86 and other commodity processors running linux has been superior in terms of cost and portability. As far as performance goes, the most Sun can do now is to sell Opterons with Linux. Kinda reminds you of a generic white box vendor, now doesn't it? Especially when you consider that the first round of Opterons were rebranded, and not Sun engineered.

Fast forward to today linux is losing out to Macs in science, every conference I go to it seems that more and more people have Powerbooks (like > 50% of the audience), especially at NASA. My project just decided to move entirely over to Macs. Solaris isn't even in the mix anymore.

This is true too. OS X on a Mac has become _the_ UNIX workstation of the 00's. Linux was the workstation in the mid to late 90s, and Sun in the 80s and up to mid 90s. Although Apple is working on this, they have not yet too much made a dent into the number crunching world. Yes, I know about the Virginia Tech cluster, but AFAIK, that was a marketing stunt. They installed at a rush $5mil worth of desktop boxes around November 2003, then ripped them out, then reinstalled with Xserve boxes, and announced it was ready for work in January of 2005. I guess I could send an email to the head of the system and ask if its being used or not, but I have a feeling it is very underutilized. Being that they only have 2.7TB of storage for 2200 processors, I would bet my hunch is right.

Re:Horses, Loaves and Shoes.

[Anonymous Coward]

...universities 5-10 years ago were 100% Solaris, with some Dec Alphas thrown in the mix.

So that's, what, 100% Solaris, 5% DEC?

Re:Horses, Loaves and Shoes.

[Anonymous Coward]

In scientific computing, Linux displaced AIX and HP/UX a long time ago. Sun was never a player as the SPARC was never a performer. The only places SunOS or Solaris every showed up in scientific computing were as front-ends on dedicated hardware (Connection Machines, login/authorization/file serving nodes, etc). Even then, only environments with a lot of old Sun sysadmins still use Suns in even that role.

Re:Horses, Loaves and Shoes.

[Anonymous Coward]

Who cares about Java?

We already have something far better, Limbo, it was design at Bell Labs by, Dennis Ritchie (the inventer of the C language and Co-inventer of Unix), Sean Dorward, Phil Winterbottom and Rob Pike (a lot of the Plan 9 design is thanks to him).

It is in a more free than Java (no crazy policies on what can be called Limbo), and has a more free licence than GPL2.

Unlike Java is has sane garbage collection that works, ONE sane fast toolkit module, advanced concurrency and has a C-like synax. You can read more about them in the docs.

A few more related links:
http://www.vitanuova.com/inferno/design.html
http://en.wikipedia.org/wiki/Inferno_(operating_sy stem)
http://en.wikipedia.org/wiki/Limbo_programming_lan guage
http://www.cs.bell-labs.com/sys/doc/

Re:Horses, Loaves and Shoes.

[Dr_LHA]

Hey, I said I was a scientist, not a mathematician!

Re:Horses, Loaves and Shoes.

[sparkz]

Hold on, there ... 6 G5s?! That's going to scare an E25K away. I think we're looking at different things, here. Sun's expertise is in the datacentre. The desktop is a different question.

How many paths to the SAN does your Mac have? How many connections to the LAN(s)? How many nodes in your cluster?

Not knocking the Mac, or desktops at all - we all need to use a desktop - but that's not been Sun's area since they started 20-odd years ago.

Re:Horses, Loaves and Shoes.

[Dr_LHA]

I believe the comment I made was about "scientific computing", as in computers used by scientists. I wouldn't argue Sun's dominance over Apple in the server environment.

The fact is in Scientific Computing, i.e. workstations used by Scientists, Sun used to have close to 100% dominance in my field. Now they're close to 0%.

Re:Horses, Loaves and Shoes.

[sparkz]

Fair enough. Sun workstations pretty much don't exist any more. There's SunRay (a terribly undersold technology), but after that, unless you fancy an expensive pizza-box like a V100 on your desk (I've never tried it, but I wouldn't expect it to make a great desktop PC), Sun don't really do desktop any more. It's what they started with, it got them some credibility with a certain crowd, and that credibility got them a lot of sales (and vast over-hype, hence the silly prices) in the 1999 .com boom, but Sun have been a datacentre company for a long time now.

Apples, Oranges, Horses, Loaves and Shoes ;-)

Re:Horses, Loaves and Shoes.

[Anonymous Coward]

OS X on a Mac has become _the_ UNIX workstation of the 00's.

Not if you need an OS with a good threading model. Or one that minimizes the performance hit of context switching.

So uh...

[Ekarderif]

We have GNU/Linux.
We have GNU/*BSD.
Does this mean that GNU/Solaris is surely to come?

Re:So uh...

[Anonymous Coward]

you should check out http://www.gnusolaris.org/

Re:So uh...

[Anonymous Coward]

We have GNU/Linux.
We have GNU/*BSD.
Does this mean that GNU/Solaris is surely to come?


Wouldn't that be GNU/SunOS, since the underlying kernel is still referred to as SunOS? Just like there is a GNU/Darwin, not GNU/OSX.

ex:
% uname -a
SunOS thedude 5.10 Generic sun4u sparc SUNW,Ultra-5_10

Re:So uh...

[Anonymous Coward]

If you've used Solaris, you'd know that using it without a full assortment of GNU apps is just a major pain. I mean, what would you do without GCC? Thus, Solaris is already GNU/Solaris just about as much as Linux is GNU/Linux.

Both are great kernels, but it's the GNU software that makes'em shine.

Re:So uh...

[slavemowgli]

It has long since existed - anyone with a bit of sanity left (that is, anyone who's not a raving Sun fanboy) will replace Sun's own toolchain with the GNU tools right after installing Solaris on a new box. Most people don't call the result "GNU/Solaris", though, for some reason, and even the usual suspects like RMS are strangely quiet on that matter - which makes you wonder if the insistence on terms like "GNU/Linux" really is purely based on rational, technical reasoning.

Re:So uh...

[Kadin2048]

I have a slightly cynical theory as to why people don't call that "GNU/Solaris" already -- it's because they paid (until recently, anyway) a whole lot of money for Solaris, and they don't want to invite comparisons between "GNU/Solaris" and "GNU/x" where x is a free kernel, particularly Linux.

If you're going to call it that, then you're going to eventually have to field the question of "What's so superior about having the SunOS kernel?" and I think there are quite a few Sun users who just don't want to go there. For one reason or another, they want to differentiate their OS choice from the free alternatives. I'm sure they may have good reasons for doing so; but regardless I doubt they want to have to go over them with everyone above them in their chain of command.

Re:So uh...

[CrankyOldBastard]

Doesn't everyone who runs Solaris replace large chunks with GNU anyway?

It's not free unless it's BSD

[Anonymous Coward]

BSD is the original unencumbered license.

Stallman doesn't understand the meaning of Free. Linus doesn't understand the meaning of Free. Sun has never understood the meening of Free.

So I guess what I'm saying is. Who cares? Sun is just exchanging one encumbered license for another.

Re:It's not free unless it's BSD

[Anonymous Coward]

bsd - free to share if you feel like it, it doesn't matter who with, and they don't have to share at all

gpl2 - free to share because you must, it doesn't matter who with, and they have to share as well

gpl3 - free to share because you must, except to non sharers based on their track record and what they are doing with it. They aren't supposed to get it because they want to threaten people with patent lawsuits and such like. Bad people, they get ostracized.

not too hard to understand

3 (if it stays like it is now) actually makes it better for further down the line, it's what 2 should have been

Re:It's not free unless it's BSD

[rbanffy]

The main difference between BSD and GPL is that GPL obliges anyone using it to play nice with others, to give back what they do (if they choose to pass it along). This is what made GPL the most widely used free software license as it did not allow anyone to grab GPL software and turn it into something proprietary (as opposed to BSD-ish stuff) - it fostered an ecosystem where software could evolve and thrive while BSD-ish sofwtare is mostly a starting point.

Re:It's not free unless it's BSD

[Aqws]

Putting you code into Public Domain would be the most open way to release your source code. The reason people like Stallman want people to release code under the GPL is so people who want to base their own code on it will have to license their code under a similar license. If they didn't do this many companies would use the source without contribuing back to the comunnity. I've heard that Stallman is also very uncompromising.

It's already available.

[CyricZ]

There are many open source Java implementations available, even if they're not derived from Sun's.

There are SableVM, JamVM, GCJ, and many others. Along with GNU Classpath and Jikes, you've got a rather complete J2SE implementation available to you. GCJ can compile to native code on certain platforms. Even with all the talk about JIT compiled code potentially being faster than native binaries (due to runtime optimizations and so forth), many people have found that code compiled with GCJ is far faster than when executed under a Java VM.

Re:It's already available.

[m50d]

Whenever people say this, I pick up my five test java applications. Three are programs I like and use, the other two are picked more-or-less randomly from freshmeat.

None of them work under such stacks. Not one.

They are good projects, but until you can pick up a random java application and expect it to work, you can't really say there's a free java available. Right now there is less support for java programs on a completely free platform than there is for windows executables.

Re:It's already available.

[Haeleth]

Whenever people say this, I pick up my five test java applications. Three are programs I like and use, the other two are picked more-or-less randomly from freshmeat.

None of them work under such stacks. Not one.

Would you care to reveal what they are? It's quite difficult to track down a bug when you have to start by reading someone's mind.

Re:It's already available.

[m50d]

I don't want to report a bug - that way I'd have to pick a new five applications every time I did the test. I want to see when they start just working.

Re:It's already available.

[MSG]

I dunno about faster. When FC4 was release, and included a copy of Eclipse that had been built with gcj, I tried it out. It was painfully slow, and used much more memory than Eclipse running under Sun's JVM. I'll try it again when I load up FC5, but I didn't think the gcj version was really usable. I trust it to get there, though.

Re:It's already available.

[lindi]

I see your point. I'd be nice to be able to pick any random free application from freshmeat.net and have it just work(tm) with free software.
Unfortunately even if you manage to avoid all the bugs in gnu classpath many applications depend on non-standard things (like com.sun.*) that gnu classpath is probably not going to mimic. It's bit like trying to use code written for borland turbo C and 16-bit DOS with GCC. Even if both program and compiler are good you might still run into problems.

Would be a nice move. Impressive indeed.

[Qbertino]

Please note people: This is a company. That means they make money. And they do it in the classic sense which means this type of company usually gets the creeps when hearing stuff like "go FOSS" or "rely on FOSS". CEOs freak out regularly when these terms come up.

If SUN plans an OSS strategy they are certainly NOT going to GPL their powerhorse Java. Solaris is nearly just as impressive from a technical standpoint. It's probably that Solaris doesn't have the numbers attached to it SUN would like to see. So they probably guess it could prove itself as OSS, since Linux is winning in the custom Unix market at all fronts.

If x86 Solaris would go GPL that would be really cool. I'd actually give it a try.

Re:Would be a nice move. Impressive indeed.

[Zemplar]

"If SUN plans an OSS strategy they are certainly NOT going to GPL their powerhorse Java. Solaris is nearly just as impressive from a technical standpoint. It's probably that Solaris doesn't have the numbers attached to it SUN would like to see."

Solaris is a platform. Java is supposed to be multi-platform. I fail to see how GPL Java would work well.

Imagine GPL Java under committeee control. Then one day, not to far distant, some member decides to fork the GPL Java because he/she has some other idea. Before long, there are 18 types of Java than are not all multi-platform and can't run the same code. Kind of defeats the purpose doesn't it?

Re:Would be a nice move. Impressive indeed.

[TheRaven64]

I've seen this argument before, but it doesn't make sense. Java is a Sun trademark. You can't call anything Java unless it passes a number of compliance tests. There is nothing wrong with extending Java - that's what the com.* namespace is for - as long as anything that runs on Java(TM) runs on your implementation. If Java were open source then people could take it, port it to a new platform, submit their changes back, and have the next official release support their favourite platform. At the moment, the FreeBSD people (for example) have to maintain their own patchset. This is exactly the situation caused Linus and friends to give up on Minix and start Linux.

It doesn't matter if there are 18 forks of Java, as long the only ones called 'Java(TM)' are ones that will run all Java(TM) code. As it is, there are a good half dozen Java VMs with varying levels of compatibility, but only the certified ones are guaranteed to work.

Re:Would be a nice move. Impressive indeed.

[Anonymous Coward]

Before long, there are 18 types of Java than are not all multi-platform and can't run the same code. Kind of defeats the purpose doesn't it?

Stupid fucking red herring that Java morons always drag out. If it doesn't meet the specs and tests, you can't legally call it Java. See also: OpenGL. Opening the code under a Free software license does not change this.

Re:Would be a nice move. Impressive indeed.

[Kadin2048]

I think the reason Sun wants to open up Solaris is because right now, adoption isn't moving very quickly because of the lack of hardware compatibility.

Their HCL *sucks.* I've never been a huge booster of Linux hardware compatibility, but if you go over and look at your options when you have Solaris x86 installed, it's a vast range of options you have for Linux in comparison. I think there is ONE 801.11x wireless card listed on their HCL. I don't know if that means that there are more compatible ones that just aren't listed, or if WL on Solaris is just a joke right now, but it pretty much turned me off to trying it.

The problem, I'm told, is that a lot of the Linux kernel modules that give hardware compatibility and which are licensed under the GPL, aren't compatible with the OpenSolaris license, and thus there's no way for OpenSolaris to benefit from the Linux community's efforts. The only way to get the full benefit that Sun wanted to garner by going open-source, would be to go the rest of the way and GPL it.

Honestly, I don't think most users care whether an OS is GPL-compatible or not. If it's free-as-in-beer, they'll use it, provided it does what they want it to do. However, when the license restricts hardware compatibility, you're not going to increase userbase, and that's what Sun is (hopefully) catching on to.

Re:Would be a nice move. Impressive indeed.

[Zemplar]

While I whole-heartedly agree with you on several of your points (especially FreeBSD's short straw) I see GPL Java going the way of w3.org HTML specifications and Microsoft's HTML "standard". One effective & correct way of doing things, but a more widely distributed audience of crap that makes the entire idea quite muddy and less effective.

Another large potential reason for keeping Java guarded is that it may really screw up Sun's enterprise stack if allowed to be GPL'd. Although I am in no way affiliated with Sun, I would suspect that this would be of a much larger concern than opening up Java alone. However, I can see a OpenJava site similar to OpenSolaris with a source code browser and using a similarly restricted CDDL license with the potential for adopting the new GPL3. Such a collaborative site could serve as the testing code and a base for future production quality code similar to the current Solaris model. This would enable quality back-ports (yay FreeBSD!) and seem to serve all needs I can think of, besides pure GPL2 zealots.

BTW, nice site Raven, I've visited frequently.

Re:Would be a nice move. Impressive indeed.

[Doctor Memory]

Their HCL *sucks.* /me nods to PP. I didn't really consider printer support when I ordered an Ultra 20 to replace my ancient P-II/450. However, when I set it up and noticed it was "legacy free" (no PS2, parallel or serial ports), I wondered how I was going to set up my printer (an old Lexmark Optra 1650 with a parallel port). I browsed to the Sun web site and saw that basically *no* printers were supported on my hardware with Solaris 10. I picked up a network printer adapter, but I haven't had time to play with it yet. The printer supports PostScript, so hopefully it won't be too much of a struggle. It would go a long way if I could get a driver for one of those sub-US$100 HP USB printers, though....

Re:Would be a nice move. Impressive indeed.

[assantisz]

CUPS works perfectly fine under Solaris 10. The usbprn driver allows you to use USB connected printers on your Ultra 20 box.

Re:Would be a nice move. Impressive indeed.

[Doctor Memory]

Hmmm, interesting. CUPS doesn't seem to support the el-cheap-O HP printers (I have a 5150 lying about somewhere), but their commercial version claims to, and it's only US$50 (less than I paid for the network adapter for the Lexmark). Maybe it's time to donate the Lexmark to a worthy charity...

Thanks for the info!

Re:Would be a nice move. Impressive indeed.

[caseih]

Imagine GPL Java under committeee control. Then one day, not to far distant, some member decides to fork the GPL Java because he/she has some other idea. Before long, there are 18 types of Java than are not all multi-platform and can't run the same code. Kind of defeats the purpose doesn't it

You mean like how there are so many incompatible versions of the linux kernel and so many incompatible versions of the Mono .NET environment? I don't think you'd have to worry much. It is in the entire Java community's best interests to keep things together, so I doubt you'd see major splitting. There are many reasons to not license Java under the GPL, but this is one of the least of them.

Re:Would be a nice move. Impressive indeed.

[Anonymous Coward]

This is a company. That means they make money.

Are you sure? http://finance.yahoo.com/q/bc?s=SUNW&t=5y I love Sun, but making money isn't something they can be seriously accused of ;-)

Re:Would be a nice move. Impressive indeed.

[sparkz]

If x86 Solaris would go GPL that would be really cool. I'd actually give it a try.

x86 Solaris is Solaris. It's the same tree.

http://www.sun.com/software/solaris/opensolaris.js p - "both SPARC and x64/x86 (it's a single source base)"

Granted, an 'interesting' use of the word 'both' when listing 3 platforms, but Solaris is Solaris, regardless of platform.

Re:Would be a nice move. Impressive indeed.

[Anonymous Coward]

>You mean like how there are so many incompatible versions of the linux kernel

Remind me again how many Linux distros there are?

SCO case would delay this by years

[Anonymous Coward]

The SCO case would delay this decision by years and Sun knows it. Sun may have a royalty-free right to distribute UNIX, but whether it has the right to distribute the code is another matter.

Linux would be better served steering clear of this quagmire.

Re:SCO case would delay this by years

[htd2]

Sun is already distributing the code to Solaris (OpenSolaris) under their CDDL license.

I somewhat doubt that SCO would consider suing Sun to be a good idea, Sun holds a rather large chunk of patents relating to UNIX networking, kernels etc and I have no doubt that any case brought by SCO would result in a rather in depth analysis by Sun of SCO's position with respect to Sun's patent portfolio.

Sun: more talk than walk

[Anonymous Coward]

Ah, Sun ... I have always had a weak spot for the hardware developed by Sun, but I do not take them seriously anymore. Why? Because they are not willing or able to sell. I wanted to buy one of their AMD-based workstations for a small business, but it is impossible to contact someone from Sun here in Finland who is taking their job seriously (i.e., not redirecting you to someone else or not sending quotations of product configurations I never asked for). Ultimately, I contacted a local Sun reseller, who promised to send me a quotation as soon as possible for the workstation I had in mind. It's now several months later and I'm still waiting ...

Anyway, it is not going to be a Sun workstation, but an Apple workstation as soon as they hit the market. At least they are able to take the order.

Re:Sun: more talk than walk

[Robber Baron]

Tell me about it.
I ordered one of their new x4200 servers, because it lookes like a pretty nice piece of hardware. If it lived up to its billing, I would have ordered more too. I guess I'll never know though, because that was over 3 months ago and I still haven't seen the first one yet, so to hell with them! They're all talk and no action and I'm going to shop elsewhere. Are you listening Dell?

Re:Sun: more talk than walk

[Anonymous Coward]

I don't know what the issue is with your order, but I placed one two weeks ago and it arrived on schedule and working. Perhaps the difference is that I went with a T2000 (UltraSparc T1 / Niagra) instead of an Opteron. Honestly though three months seems fishy to me, I haven't seen the kind of supply issues you're experiencing and I order from both Sun AND Apple ;)

Re:Sun: more talk than walk

[Anonymous Coward]

What do you mean "Are you listening Dell?" get an HP 585 dual core Opteron. They are the same box and beat the Sun box out already in comparisons. And they are in stock.

Re:Sun: more talk than walk

[Anonymous Coward]

You be sure to let us know just as soon as you get the Dell Opteron server.

Oh wait, Dell doesn't sell Opteron based servers.

My company buys literally hundreds of Sun servers every year, I have yet to see anywhere near the kind of timeframe you are talking about. The worse we had was 45 days during a time when Sun couldn't book our orders because of contract negotiations over our discount structure.

GPLv3 doesn't actually exist yet...

[_xeno_]

The GPLv3 is still in draft form. It doesn't actually exist yet. The version on the FSF webpage could be better classified as a "beta" release (I think that's what Stallman considers it).

It's a little early to be saying "I'm going to be using the GPLv3!" Yes, they're working on it, but it's not actually out yet. The optimistic "release" date is November of this year, with the expected release date being early 2007... It's just not ready yet!

However, thinking about the current draft and any problems you have with it is encouraged. They want comments still, there's still time to help change the final draft. Saying "I'm going to use the GPLv3!" is still premature. Wait until it's actually finished, then decide.

Re:GPLv3 doesn't actually exist yet...

[justins]

It's a little early to be saying "I'm going to be using the GPLv3!"

Is anyone actually saying that?

Re:GPLv3 doesn't actually exist yet...

[_xeno_]

Read the blog post. The author specifically specifies the GPLv3, and not just "the GPL." The blog title is "Thinking About GPL3..." and he links to a copy of the GPLv3 draft. (Which actually says "THIS IS A DRAFT" right on it, don't know why he linked a copy and not the GPLv3 site, but...)

So, yes - he's talking about using the GPLv3 as opposed to the current GPL.

Which is silly, because the GPLv3 is still in draft form. It's not released. Speculation about how to apply the GPLv3 would make sense, talking about actually releasing something under it is rather early. Unless he intentional means to say "we have no plans on releasing an open source version of Solaris until 2007."

Re:GPLv3 doesn't actually exist yet...

[squiggleslash]

I think the GP was right. Schwartz is talking about possibly switching, not saying that Sun's actually doing it. Note in particular this paragraph:

We also recognize that diversity and choice are important - which is why we've begun looking at the possibility of releasing Solaris (and potentially the entire Solaris Enterprise System), under dual open source licenses. CDDL (which allows customer IP to safely comingle with Solaris source code) and under the Free Software Foundation's GPL3. It's early days, but we're looking at two things as we make that decision.

Not "Yes, we're gonna do it!" but "We've begun looking at the possibility of..." and "It's early days... looking at (things) ... make that decision"

There's nothing in Schwartz's blog that suggests the decision has been made, that Sun are doing this, or that they're unaware of the draft nature of the license. By all means criticise them as stupid if they actually switch before the GPL3 has been released, but it's great they're looking at the current draft and saying "This is something we can seriously consider." It gives valuable feedback to the FSF, and it suggests Sun are going to become even better Free Software players in the near future.

Re:GPLv3 doesn't actually exist yet...

[justins]

You missed the point of my question. He isn't committing to anything, one way or the other.

Unless he intentional means to say "we have no plans on releasing an open source version of Solaris until 2007."

All versions of Solaris are already open source.

WTF?!?

[everphilski]

If Linux dont want do comply with openness

What the f*ck? GPLv2 was way more open that GPLv3 is looking to be (check it out for yourself: heres a draft analysis). Note the restrictions on (a) DRM (b) patent retaliation. While you may like what GPLv3 has to say about those things you do have to agree these are restrictions that DO NOT EXIST in GPLv2. Therefore, GPLv2 is more open and less restrictive.

Re:WTF?!?

[CastrTroy]

But maybe we don't want the most open and least restrictive. Because if we did, we'd all be using BSD. Which is the least restrictive license I know of. I think what a lot of GPL users want is for their code to stay GPL, and for changes made to the code by others to be brought back upstream, so the whole community can take advantage of the changes. I think that's what GPL V3 is trying to accomplish.

Re:WTF?!?

[Billly Gates]

The reason Linus turned down GPLv3 is that it required giving the copyright and permission available from all contributors. Linus wants to keep it trademarked under his name and the task is impossible to track everyone down for approval with GPLv3.

Re:WTF?!?

[Anonymous Coward]

Personally I really like Solaris and Linux. That being said, sun is a company and I can't shake the thought that this GPLv3 stuff is just bait. I'm waiting to see how much of solaris they are going to GPL, and if that is ok, wait a bit for the dust to settle just in case it's a trap. Paranoia aside, GPLing Solaris so it is 100% compatible license-wise with linux is a really good idea. Linux and Solaris would benefit from this move... Sun is a hardware company mostly so they wouldn't have alot to lose..

Re:WTF?!?

[m50d]

If you wanted to be the least restrictive you'd put it in the public domain. I really don't see the point of the BSD license in most cases.

Re:WTF?!?

[Anonymous Coward]

I really don't see the point of the BSD license in most cases.

That's easy. The license says that redistributions must reproduce the copyright notice in source code or documentation. That's the point. So that while people can use your work freely, they can't pass it off as their own work. Everyone who cares enough to look knows that at least part of the software they are using was developed by you.

Re:WTF?!?

[labratuk]

I think you're very confused about the difference between copyright & trademarks. And a bunch of other things.

Re:WTF?!?

[CastrTroy]

The difference between BSD and the public domain is that with BSD you still maintain copyright over whatever you release. Others can do whatever they want to with it, but you still have the copyright. If you release something into the public domain, I don't believe you retain any copyright to whatever you release.

Re:WTF?!?

[m50d]

No, but why would you want the copyright? So you can relicense it under another license? There's no point, since either BSD or public domain is liberal enough that you can do that anyway without holding the copyright. So you can stop others making derivative works? But the license lets them anyway. There's no point keeping your copyright if you're not going to exercise any of your rights under it - and BSD allows just about everything you have a right to prevent under copyright.

Re:WTF?!?

[mpcooke3]

Depends if you consider the additional terms "restrictions" or granting/enforcing users "rights" and ensuring the software remains "free" and "open".

Re:WTF?!?

[Lehk228]

publicity, users of your code have to acknowledge that your code is being used in their products.

PPC port

[sirinek]

Schwartz also said the only relevant OSes left today are Linux, Windows and Solaris.

Anyway, they had a PPC port of Solaris.... 10 years ago when they had Solaris 2.5.1. Why would they want to start it up again?

Re:PPC port

[Billly Gates]

Opensolaris already has backported it to powerpc. Its almost finished. So why not?

I think he wants solaris to be the alternative to linux. As a former FreeBSD user its nice to see a stable operating system that is also enterprise ready in terms of scalability and performance. I like unix more than linux and I want to try out OpenSolaris soon when I have time. The way sun wants to do this is increase the platforms and peripherals solaris can run on. Solaris10 is a big improvement on x86 over 9 in terms of hardware support and performance.

Re:PPC port

[h4ck7h3p14n37]

Schwartz also said the only relevant OSes left today are Linux, Windows and Solaris.

*sigh*, I probably shouldn't bother responding, but Schwartz left out a few very important operating systems, namely the three BSD distributions and AIX.

I honestly don't understand all of the hype surrounding Linux; I'd recommend FreeBSD over Suse, Red Hat, Debian, etc. any day. As for AIX, I can only assume that wasn't mentioned because it runs on IBMs Power processors which kick the crap out of Sun's gear as far as I/O throughput is concerned.

Re:PPC port

[Anonymous Coward]

PPC is the only widely available non-x86 'hobbiest' hardware that's not ridiclously outdated. Unlikely anyone cares about running Solaris/PPC in production.

Re:PPC port

[assantisz]

Guys, get yourself a little more current. OpenSolaris is actively being ported to PowerPC. Thanks to opening the source so many things that were promised in the past finally become true.

GNU

[Billly Gates]

GNU = GNU is Now Unix

Re:GNU

[Aqws]

It should now be: GNU = "GNU NUG UGN", where NUG stands for "NUG UGN GNU", and UGN stands for "UGN GNU NUG".

Summary makes no sense

[JediTrainer]

1 - Linux won't use Solaris code because there doesn't appear to be any intention of migrating to GPL3

2 - Solaris can't use Linux code because Sun wants to keep their code under a second license (CDDL), which is at some level incompatible with the GPL (a.k.a incompatible with any imported Linux code)

I applaud Sun's ideas, but I am looking forward to a Fully Open Source Java (granted, I do appreciate that alternatives from other vendors are available nevertheless).

Re:Summary makes no sense

[ChunderDownunder]

I am looking forward to a Fully Open Source Java (granted, I do appreciate that alternatives from other vendors are available nevertheless).

There are no alternatives. Vendors such as IBM, Apple, blackdown etc ALL license Java code from Sun and hence are subject to RMS's Java Trap.

The only contender is a 'free' runtime based on the GNU classpath class libraries.

Despite claims to implementing 98% of the code base, there is still work to be done - e.g. the reality is that one can't take a Swing application and expect it to work trouble free. And as much as SWT advocates would wish otherwise, Swing is still the dominant toolkit.

But the day is drawing closer. The hope is that this time next year RedHat (or perhaps IBM) will have licensed the Technology Compatibility Kit on classpath's behalf.

The main benefit being that any X11-based OS[1] will automagically be able to run a fully Java compliant JRE, including those platforms or architectures for which Sun hasn't ported their implementation. From a commercial point of view this widens the number of java supported environments for deployment. From a free software perspective, the inclusion of a 100% compatible JRE means that distributions will start integrating Java software into their releases on technical merit - no longer omitted due to an inability to license Sun's Java.

Would the situation be helped if Sun licensed Java under GPLv3? Definitely, but let's not hold our breath...

[1] Any OS with their own UI layers as Windows, OSX, haiku etc would need to implement their own AWT peers. But 'mainstream' alternatives such as the BSDs and unsupported CPU architectures such as XScale (I'm thinking handheld linux here) should just work.

Re:Summary makes no sense

[swordgeek]

"Vendors such as IBM, Apple, blackdown etc ALL license Java code from Sun and hence are subject to RMS's Java Trap."

Wah wah wah. According to RMS hisself, "We do have free implementations of Java, such as the GNU Compiler for Java (GCJ) and GNU Classpath, but they don't support all the features yet. We are still catching up."

So the reason that programming in Java is a 'closed-source trap' is because OpenSource(tm) hasn't caught up yet! Cry me a bloody river--the specification IS there (despite RMS's convoluted logic), you're free to implement it. If you're not good enough to write the code from the spec, then it's not Sun's fault, it's yours. THIS is what Open Source was supposed to be about.

PowerPC Solaris

[mknewman]

Back in the Solaris 5 timeframe there was a PowerPC version available. The idea was that it was going to be available on the IBM/Apple/Motorolla 'Refrence Platform'. Solaris 2.5.1 came out and it was no longer available. This must have been around 1995-1996. Since they already have some base drivers available for PowerPC it should be pretty easy to reactivate that leg of the source tree, from what I know.

Re:PowerPC Solaris

[adrianmonk]

Back in the Solaris 5 timeframe there was a PowerPC version available. The idea was that it was going to be available on the IBM/Apple/Motorolla 'Refrence Platform'.

Although there is no such thing as Solaris 5, I can confirm that a PowerPC port of Solaris existed. It was Solaris 2.5 that was available for SPARC, PowerPC, and x86. Here's some of Sun's documentation about it.

I also remember seeing PowerPC versions of patches on the SunSolve contract support site for Solaris, so I can only assume based on this that they not only made it but actually sold it to customers. (If not, why provide supported patches?)

Of course, at the time, so many industry players were getting behind the PowerPC (such as IBM, Apple, Sun, and Motorola) that lots of people thought it might unseat x86. Especially since x86 performance had stagnated a bit at that time. So lots of people ported to PowerPC as a bit of insurance, just like they ported to Itanium for insurance a few years later.

Re:PowerPC Solaris

[Anonymous Coward]

Not only was Solaris ported to the little endian PREP PPC system around 1995, but it was ported to the PREP big endian as well, and, more interestingly, there was a project at SunLabs called Chameleon, which was a PPC sun4m (e.g., Sparc 10 and 20) m-bus card and a Solaris port. This was right around the time that Sun tried to buy Apple.

Exactly - ain't gonna happen

[Just Some Guy]

As reluctant as Sun has been to test the GPL waters, I can't imagine them diving right in to the GPLv3 unknown. If anything, perhaps, they might:

1. use the "version 2 or later clause" (unlikely), or
2. use version 2, and go FSF-style and make contributors assign copyright on new code back to Sun - leaving them the option to migrate en masse at a later date once version 3 gets a bit more polish.

I think those are the only remotely possible options at this point.

Oh they surely tried...

[Hymer]

"Schwartz has speculated on exciting things in the past (such as porting Solaris to IBM's Power) that we subsequently never heard another thing about."
All involved in the PowerSolaris project has disapered from the surface of the world... the project clearly showed that continous Power is better than a Sparc. :-D
--
This sig is not a sig...

Solution, Seeks Problem...

[Kadin2048]

The Linux kernel probably won't ever be GPL3, because the license it uses doesn't contain the forward-compatibility clause that the FSF's software does; however, all the GNU utilities (including, I believe, GCC) will be GPL3 and/or GPL2, because they have the forward-looking clause.

So really what it would allow a person to do, is produce a GNU/Solaris as opposed to GNU/Linux -- an OS that would have the Solaris kernel, wrapped in the GNU utilities, without the Linux kernel. I'm not sure if anyone would really want that, because I'm not sure that it would be compatibile with either existing Solaris or existing Linux software without rewriting, and it generally seems to be a solution looking for a problem (not unlike GNU/Hurd).

Re:Solution, Seeks Problem...

[booch]

I would definitely prefer that Solaris have the GNU versions of most of the command-line utilities. I have a hard time going back to the "old" versions that Solaris has, which are often missing functionality that the GNU versions have. Just the ability to put options to commands after the filename arguments would make me happy.

Re:Solution, Seeks Problem...

[Malor]

Personally, I'd LOVE a GNU/Solaris distro.... the Solaris kernels are extremely robust. I've had very bad luck with Linux all through the 2.6 series. Even the stable versions have had constant security patches, requiring reboots and downtime. I hate downtime.

FreeBSD with the GNU utilities is one possible replacement, but it'd be nice to have a kernel that's both extremely robust AND scalable at the center. FreeBSD is very solid, but it doesn't (yet) scale like Solaris does. Linux scales, but it's not stable. Solaris does both things at once, which makes it very attractive. So a distro that combined a Solaris kernel with a well-integrated GNU userland would be very interesting to me. (bonus points for good documentation... coming from Linux, Solaris can be downright Byzantine in comparison.) Learn the system once, and use it for everything from a desktop to a SunFire.

I imagine the hardware support would be horrible at first, but it wouldn't shock me to see it improve very, very rapidly once it was set free. And I'd happily support it with some dollars here and there. I've spent a great deal of money on Linux over the years, and if Solaris met my needs better, I'd be happy to pony up there too. (just in small amounts... not thousands per server! :) )

The biggest reason I haven't used Solaris X86 is because I haven't had any available systems with compatible hardware. GPLing it would, I think, fix that problem in short order....

Re:Solution, Seeks Problem...

[jrutley]

There already is a GNU/Solaris distribution (called Nexenta).

The end of "GNU/Linux" - now just "GNU"?

[daivdg]

Will it now be better to say say just "GNU", if the kernel can be replaced by either Solaris or BSD?

Re:The end of "GNU/Linux" - now just "GNU"?

[Anonymous Coward]

IMO it's best to identify the kernel, because these systems are incompatible in some levels (drivers and binaries are not compatible, and some code may be specific to a certain kernel). Of course, since all would use glibc and gcc they would be mostly compatible for most software.

GNU running with the GNU Hurd = GNU
GNU running with Linux = GNU/Linux
GNU running with the kernel from FreeBSD = GNU/kFreeBSD
GNU running with the kernel from Solaris = GNU/SunOS or GNU/Solaris

GNU/Solaris

[oob]

Up until Solaris 10, the first stop when commissioning a new sun box was always Sun Freeware. Sun Freeware has a collection of popular binaries in Sun's package format, things like SSH, SSL, BASH, gcc, top, gzip, etc.

Many of these utilities are covered by the GPL and Sun didn't ship them, yet most admins consider them to be vital or at least very useful. Around Solaris 10 however, Sun got with the programme and included GNU stuff with their distribution media.

So to answer your question, GNU/Solaris (meaning "a Sun system running a whole bunch of GNU stuff") has been extremely common for a long time.

Re:GNU/Solaris

[Anonymous Coward]

At GNU they recommend using Debian since it is mostly all GPL software. If Solaris went GPL3, I'm sure there's a good chance of GNU recommended people to use Solaris or maybe Debian GNU/SunOS.

GNU/Solaris

[Newton's Alchemy]

If this happens, would it be the beginning of the end of the Linux kernel? Since the GNU suite would most likely want to be GPL3, and the Solaris kernel would as well, would the Linux kernel become less used?

It's a Red Herring.

[labratuk]

Well said. In my view, Sun's latest FOSS lovefest has been all about releasing red herrings to try and disperse some of the momentum around existing projects.

PowerPC port is being developed

[velco]

... by the community, see

http://www.opensolaris.org/os/community/power_pc/
http://www.itjungle.com/tfh/tfh011606-story02.html
http://www.blastware.org/

Exactly what does this mean to SCO vs IBM, et.al.?

[Anonymous Coward]

I suppose that with Sun's license of AT&T code from SCO, then Sun's release of Solaris under GPL3, this might put Sun in SCO's sights for a lawsuit. What do you folks think?

yeah, right

[penguin-collective]

Just like Sun was going to open source Java, and like Sun was going to make an ISO and ANSI Java standard.

Sun management is a bunch of liars. At this point, you can't believe anything they say until they do it.

Re:yeah, right

[kaffiene]

Java *is* open source, just not the exact license *you* want them to release under, apparantly.

Re:yeah, right

[Anonymous Coward]

Point me to where the licensed used by Java is certified by the OSI as open source. No? Didn't think so... fuck off. The Java license is no more open source than Microsoft's Shared SOurce bullshit.

Re:Java *IS* Open Source

[Anonymous Coward]

What do you mean by "Sun *WAS* going to open source Java?" If you have listened to Sun, Java is now and always has been open source. They brag that they open sourced Java before there was FSF/GPL or OSI.

I would move in a heartbeat

[Builder]

I'm losing more and more interest in Linux because of it's lack of enterprise features. Hell, a month and a bit ago, I could have been sound asleep in bed if I'd been using Solaris, instead of up at some ridiculous time of the morning:

http://www.penguinpowered.org/wayne/blog/if_i_used _solaris_instead_of_linux-2005-12-14

Re:I would move in a heartbeat

[Puff65535]

Thats dumb, when I re-probe under RHEL3, I use the HP supplied probe-luns all the time to pick up changes to our EVA5K, you also need the vendor stuff if you want dual channel failover. This is the same in the Solaris world, its wierd running HP software on a Sun, but if you want the fancy SAN features its required.

That doesn't sound like a native build.

[CyricZ]

I'd hardly consider Fedora representative of the typical Linux distribution. Many people have reported relatively severe performance problems with it, and it often includes non-standard releases of various software. That's not new for Red Hat, however. They were the ones who included the unofficial and unsanctioned 2.96 branch of GCC with earlier products, which ended up causing massive problems for a lot of people.

I wouldn't be surprised if the Fedora installation you were using wasn't actually using a native binary. They may very well have been using the bytecode interpreter included with GCC. That would explain why it ran so slow for you, even slower than with Sun's JRE.

If Fedora did the correct thing and included a native build of Eclipse, you likely would have found it to run much faster, and with better memory usage. That is indeed what I found when I used a more mature distribution (such as Ubuntu or SuSE).

Re:That doesn't sound like a native build.

[lindi]

Wasn't Fedora actually the _first_ distribution that offered eclipse running with free software and natively compiled? At least http://sources.redhat.com/eclipse/ says it's natively compiled in Fedora Core 4.

I have to admit

[jilles]

I have to admit Sun seems to have some positive swing lately. They're selling some hardware again. Their open source policy finally seems to be bringing in new customers, after years of shooting themselves in the foot with weird licenses, inconsistent marketing and plain corporate stupidity affecting all their business units.

Their strategy of the past years has been very ineffective. Java has become a multi billion industry, Sun has invested hugely in it but they have failed to cash in on it directly (I suspect they barely break even). Also it hasn't driven hardware sales that much. Their main competitor on the other hand seems to have a very succesful Java strategy. IBM is leading the way in application servers, IDEs (eclipse, rational rose) and middleware with basically the whole industry eating out of their hand, including most JCP specifications committees. On top of that they also sell the hardware & support to go with the software. IBM loves Java!

Then the whole x86 solaris thing has come a long way too (from 'hey it's free now but we support it for a fee', 'oh wait we don't do that anymore' to 'oh well lets open source the whole thing and forget about it' to finally 'hey x86 solaris is really important to us'). I mean, what do they want?

Sparc sales have been a disaster for the past years with people basically favoring IBM power and x86 with linux. They may be laughing at intel for itanium but they have one thing in common with intel: x86 is driving sales for both of them. Sun has a few next generation architectures on the shelves which no doubt they are going to try to sell. These chips had better be way better than the competition (and their mediocre current offerings) or otherwise whatever spin they put on it won't work. Personally I'm not convinced yet. Their two operating systems (solaris/linux) & two architectures (x86/sparc4 aka niagra) policy is going to continue to confuse people. Sun seems to think the combinations can coexist without affecting each others marketshare. I don't. Confused customers will look to IBM and others.

To me Sun still is a company in trouble. Maybe a few of their business units are recovering (finally) but that still leaves large parts of the company not performing very well.

Sir, you forgot...

[Savage-Rabbit]

The scenarios I'm reading into your comment are:

1. Torvalds decides he hates the GPL, revokes it and switches to closed source. In this instance, everyone uses the last GPL'd version and forks the operating system.

2. Someone sues Linus for copyright infringement. While the case is bogus, Torvalds settles out of court, agreeing to no longer distribute Linux. Everyone uses the last GPL'd version and forks the operating system.

3. Someone sneaks code into the operating system that they didn't own the copyright to in the first place, and sues Torvalds. This is exactly the same scenario as would happen anyway (eg if IBM lost its lawsuit to SCO.)

While you raised a series of interesting points you also forgot some important scenarios:

4. Linus is kinapped by a group of highly trained SCO lawyer-commandos and brainwashed into becoming a fierce proponet of SCO-Unix.

5. The Bush administration abandons the war on terrorism as unwinnable, declares Linux an un-American actvity and carries Linus off to an internment camp at the bottom of a crater on the frigid plains of Vastitas Borealis on Mars.

6. Linux systems all over the world become self aware, form a huge super cluster and merge into a collosal artificial lifeform that decides expunge to all carbon based lifeforms from the face of the galaxy, starting with Humans.

"Dual-sourcing"?

[Lost Race]

Dual-sourcing is when you get the same (or an equivalent) part from two different manufacturers.

What Sun is doing would be more properly called "dual licensing".

GNU/Solaris already exists :-)

[LionKMP]

I know 3 OpenSolaris distros, dunno if there are more. http://www.gnusolaris.org/ (Nexenta OS) looks to be the most complete, as the others (BeleniX and SchilliX) have either Live CD only or Console only (at least as I understood their website when I checked). In GNU/Solaris Nexenta OS you have KDE, Gnome, XFce4, real install or live CD as you wish. Works perfect on my PC. :-)

Solaris on Power 4

[Spinlock_1977]

I believe some group of ambitious folks booted the Solaris kernel (as opposed to the entire OS) on Power 4 sometime last week.

It's got logic

[Pavel Stratil]

GNU-ing Solaris is the best thing to happen to both systems. Let's compare:

Solaris
Has: Strength on good HW. Is a good, stable base for any decent server/workstation
Has: Some really neat kernel related code and apps.
Needs: Users. Needs a comunity to be an alternative for paid support. Needs to be recognised by developpers better..
Needs: More apps to chose from. There are apps that just won't run on Solaris out of the box. So you either edit source or don't use the app.

Linux
Has: Great userbase.
Has: Tons of apps.
Needs: Backup. Sun has a great reputation being a software and hardware producer. Sun's HW is really good and is now approaching mainstream by working with AMD. So plainly, good support from HW producers means HW vendors are more likely to offer alternatives to Windows.
Needs: Well Linux doesn't really need the stuff Sun/Solaris offers but it would definitely improve linux here and there

Re:It's got logic

[swordgeek]

"There are apps that just won't run on Solaris out of the box."

Just to clarify here...

Most apps that won't run on Solaris out of the box fail because of poor coding. I'm not saying that it's GOOD, but there's simply an enormous amount of poorly written software out there that is platform/tool/version specific when it doesn't need to be.

At any rate, I don't know enough about GPL3 yet to know how well it will fit in with Sun's world. The GPL2 was clearly broken from a commercial OS point of view, which is why Sun worked towards the CDDL license. If they can use GPL3, then it probably means that the license has been fixed.

SATA on Solaris 10 Worked for Me

[frohsinn]

I was able to run Solaris 10 as well as Open Solaris on my ASUS K8N-DL with dual-opterons. For SATA, I've used both the WD Raptors as well as a larger capacity Seagate SATA drive (can't remember the model number of the SATA). I didn't use the Silicon Image hardware raid controller, however.

Ahh Grasshopper! You miss the base point

[wilec]

"Anthropic principle: We see the universe the way it is because if it were different we would not be here to see it."

Ahh Grasshopper! You miss the base point of the anthropic principle. While it could be that the difference would be so great that we would not exist, it could also be of a lesser difference that we would exist but in a different way, and would thus only see the universe differently. A simpler and more observant statement would be:

We see things the way we see them because that is the way things are.

A more factual statement would be:

We see things the way we see them because this is how they appear to be from our perspective.

The sad part is most of us usually see things as per the idiotic principle. That being:

From a incalculably limited perspective we see things as either as how we wish them to be or as how we fear they are.

Matthew