Sun Considering GPL For OpenSolaris

narramissic writes, "At an event today to formally open-source Java, Jonathan Schwartz, Sun's president and CEO, and Rich Green, the company's senior VP of software had an exchange in which Schwartz put Green on the spot about using GPL for OpenSolaris: 'Are you averse to changing the license, Rich Green?' Schwartz asked. 'Certainly not,' Green responded, prompting the Sun CEO to fire back in a half-joking manner: 'Will you GPL Solaris, Mr. Green?' 'We will take a close look at it,' Green said, adding that it was possible that the familiarity and comfort level many developers have with the GPL may result in Sun adopting it for OpenSolaris." Another note about Sun's decision to use the GPL for Java comes from reader squiggleslash, who writes: "According to Jonathan Schwartz, the decision of Novell and Microsoft to '(suggest) that free and open source software wasn't safe unless a royalty was being paid' is what prompted Sun to finally come down on using the GPL for Java. So I guess every cloud has a silver lining."

Comments

Money Pressure

[SpaceLifeForm]

Remember, SUN makes money on hardware.
Novell and Microsoft do not.

Re:Money Pressure

[morgan_greywolf]

Remember, SUN makes money on hardware.
Novell and Microsoft do not.

Yep. Microsoft doesn't make any money from hardware sales at all. No siree. Not a dime. And Novell never made anything from hardware sales either.

Re:Money Pressure

[DAldredge]

Other than their keyboard and mouse operations all Microsoft hardware operations operate in the red.

Re:Money Pressure

[Sqwubbsy]

all Microsoft hardware operations operate in the red.

I thought that was the case with Sun, too?

Re:Money Pressure

[swimmar132]

They lose money on the majority of their hardware businesses.

I think the only thing they make money on is Windows and Office and mice and keyboards.

Re:Money Pressure

[Chosen Reject]

You would have done better by never mentioning the Xbox, which has put MS over $4 billion in the red. I have no numbers to back my next statement up, but I am guessing they have not sold enough keyboards to make up for that amount.

Also, the Zune has not yet made a dime for MS, and I've seen rumors that it is also being sold at a loss.

In addition, Novell has not sold hardware for a long time. In fact, they haven't done it since they became a profitable software company.

Re:Money Pressure

[molnarcs]

Remember, SUN makes money on hardware. Novell and Microsoft do not. Not quite. MS makes money on software, yes, but Novell makes money on support, not by selling software per se. A fairly large portion of SUN's money comes also from support.

Re:Money Pressure

[Anonymous Coward]

Man, this is Slashdot; please leave your common sense at home like the rest of us basement-dwelling butt-ugly smelly faggots.

Re:Money Pressure

[spitzak]

Yet when asked why Apple does not open-source their code (or otherwise allow others to make machines using it), the response is "Apple makes money on hardware, not software!"

Brilliant! Apparently you can use that reasoning to argue for anything

Re:Money Pressure

[Iron Condor]

Remember, SUN makes money on hardware.

Do they?

There was a time when I understood what exactly Sun makes money on. They had some proprietary hardware (the sparc achitecture) and they had an OS that took advantage of the strengths of that HW. Fine. But these days, sun servers are just Opteron boxes, no? And the OS is opensourced.

So how exactly are they making money?

Re:Money Pressure

[Anonymous Coward]

Sure they do, I'm using a genuine Microsoft mouse right now.

Re:Money Pressure

[Millenniumman]

No, they certainly make money on the hardware. Have you seen the prices?

How else would they make money (which they occasionally do)? Selling hardware at a loss in order to give away more free software isn't profitable.

Re:Money Pressure

[jargoone]

But these days, sun servers are just Opteron boxes, no?

No.

Real Businesses (hint: ones that handle lots of money) usually don't use Opteron boxes for mission-critical applications. They use Really Expensive Sun Sparc boxes, and they will be doing so for quite some time.

Re:Money Pressure

[Anonymous Coward]

Let's not take this out of proportion, it's not like microsoft is the ultimate evil.

Re:Money Pressure

[Sillygates]

But these days, sun servers are just Opteron boxes

Yes, they are making Opteron boxes. They are even selling them.

Re:Money Pressure

[fucksl4shd0t]

That's because he makes his money on hardware, not software.

Re:Money Pressure

[Alioth]

Their *LOW END* hardware is Opteron based. I suspect they make their bread and butter off the high end stuff which is still Sparc. Sparc is still being developed and pushed forwards. They just don't use it in the low end stuff.

Re:Money Pressure

[Coppertone]

I disagree. The company I work for handles billions and runs a lot of Opteron box with mission critical software.

Re:Money Pressure

[fimbulvetr]

Agreed. We've just recently started adopting them as well. So much faster. I love the fact that it no longer takes 3 min to generate an ssh key like it does on a new v240.

Plus, you don't need to worry about the big non-ecc cache thing sun is so famous for. No more spontaneous reboots, yippie!

Re:Money Pressure

[ak3ldama]

You have to read between some of the lines here. Schwartz has said (i'm not sure if it was his blog or some videos they put up recently) that through open sourcing Solaris, they have opened the doors for more potential customers. He says they are now seeing licenses purchased for boxes that came from IBM, HP, Dell etc. My view is that these customers would have been too scared to go with Solaris unless they open sourced it, because in the past Sun had abandoned Solaris on x86.

Re:Money Pressure

[jargoone]

Rrrrright... which falls into the other side of "usually". But maybe I shouldn't say "usually". How about this: at least some large businesses still use big Sun boxes.

Re:Money Pressure

[htd2]

SPARC is used in low end servers, the T1000 and T2000 are low end boxes and both use SPARC processors as do the V2XX servers. The T1000 and T2000 are extremely competitive with low end x86-64 boxes assuming that you have workload that exhibits thread level or process level parallelism. They are also very low power consumption.

Sun also has a rather wider range of Opteron based boxes than any of their competitors. They are currently the only major hardware vendor which supports 8 module (16 core) Opteron servers with their X4600. 16 cores is probably beyond what most people would consider to be low end.

That would be awesome!

[RLiegh]

Sorry for the fanboyish response, but I think releaseing various parts of Open Solaris under the GNU license would lead to some great developments. As I understand it, that would enable a lot of features of the Solaris kernel to be imported into Linux and vice-versa.

Of course, there'd be a problem with that whole "gnu's NOT unix" thing... ;)

Re:That would be awesome!

[TheRaven64]

On the other hand, the CDDL is a Free Software license (according to the FSF and the OSI), and is not Copyleft (or 'viral' if you prefer), so I would consider it to be more interesting than the GPL. For everything I've tried, my Solaris box is nicer than any Linux machine I've used (although I really don't like the Solaris userland), so I don't really see what Solaris would gain.

I think Sun made a very clever choice with the CDDL for Solaris. It's Free, and the Linux guys can't just take the best bits and surpass them. At the moment, some of the BSD guys are doing so (taking ZFS and DTrace, for example), but Solaris has gained a lot from *BSD over the years.

As someone who uses a variety of *NIX platforms, none of which is Linux, I don't really see what Sun would gain from using the GPL, and I can see what they would lose.

Re:That would be awesome!

[morgan_greywolf]

Sorry for the fanboyish response, but I think releaseing various parts of Open Solaris under the GNU license would lead to some great developments. As I understand it, that would enable a lot of features of the Solaris kernel to be imported into Linux and vice-versa.

And I'm sure that there wouldn't be any little companies from Utah that wouldn't just LOVE to see that Unix code REALLY get imported into the Linux kernel.

Where's those guys with their "itsatrap" tags when you need them?

Re:That would be awesome!

[RLiegh]

>so I don't really see what Solaris would gain.
Device drivers

Re:That would be awesome!

[RLiegh]

Given that Sun has given away it's kernel as it is, I don't see how they're any safer with regards to trade secret litigation than they would be if they go the GNU route.

Re:That would be awesome!

[TheRaven64]

The driver models used Linux and Solaris are sufficiently different that a port from one to the other is non-trivial. You could use the Linux driver as reference for writing a Solaris driver, but that's about it, and you can do that already (the *BSD guys already do that when Linux gets support first).

Re:That would be awesome!

[Sloppy]

and I can see what they would lose.

What would Sun lose?

Re:That would be awesome!

[johansalk]

No it won't, for the simple reason that Linus is "a bastard, and proud of it!". Just google "linus" and "solaris" and see how dismissive he is of it, calling it "a joke", just like he's been dismissive of the BSDs.

Re:That would be awesome!

[BalkanBoy]

That's not a fanboy response at all - it's probably the only other response, beyond any financially motivated ones, that makes sense. Solaris is considered by many (with more than just passing knowledge of UNIX) to be still 'l33t'-er than either Linux or BSD (I beg to differ of course, being a Linux/BSD 'fanboy' and all). Eventually any technical differences between those two (Solaris, Linux) ought to fade, and we could have a great, free, open-source, commodity server OS that anyone can look at, improve, etc. Now if we can get something like the Mac OS X type of GUI going with Linux on the desktop.... where would that leave other 'windowing' operating systems? :)

Re:That would be awesome!

[shrewd]

"Of course, there'd be a problem with that whole "gnu's NOT unix" thing... ;)"

linux is not GNU.

Re:That would be awesome!

[LizardKing]

Just google "linus" and "solaris" and see how dismissive he is of it, calling it "a joke", just like he's been dismissive of the BSDs.

Which probably explains why it's taken Linux so long to start resembling the Solaris kernel in terms of architecture. Linux was a poor second to the BSD's in the early to mid 1990's because it was largely written by hobbyists who didn't have the resources or knowledge that had been fed into Unix over 20 odd years. Until 2.6, Linux was a poor second to Solaris because of Linus's unwillingness to listen to people who had looked beyond the confines of the Linux kernel. People who had looked at the design of Solaris were forever making comments to the effect that Solaris had done x and acheived y, while Linus was floundering about looking for answers to things like sophisticated SMP support, threading and a decent scheduler. The Solaris kernel is still way more sophisticated than Linux, it's just the userland that needs an overhaul - and no, that doesn't mean importing gobs of poorly or inaccurately documented GNU tools ...

Re:That would be awesome!

[Argon]

> my Solaris box is nicer than any Linux machine I've used (although
> I really don't like the Solaris userland)

Then you should really give Nexenta a shot. Some debian folks have legitimate trouble with the licensing (mixing GPL software with CDDL libc). However, barring this niggling issue I think the Nexenta team has done an amazing job. I've tried it out myself. dtrace and zones work, so you should be at home.

Re:That would be awesome!

[squiggleslash]

For what it's worth, the CDDL is, as I understand it, a "weak copyleft", not a "not copyleft." If you modify CDDL-licensed code, your code also has to be licensed under the CDDL. If you add new modules that link with CDDL code, then the additional code is free to be licensed as you see fit. The MPL is similar, as is the LGPL.

So those with issues with so-called "viral" licenses aren't likely to be happy with the CDDL, and the CDDL still has the same issues as the GPL, MPL, APSL, etc, when it comes to integrating code from several sources under different licenses, while creating loopholes that allow someone who is determined to circumvent the license to easily, legally, do so.

We really need free software people to have the guts to stand up and cut down the number of licenses, essentially, to two, where others are shunned. One would be a strong copyleft, with optional (to the licenser) add-ons such as the classpath thing (these weaken the copyleft, but if you make them optional those concerned about their code being used in non-free environments can simply remove them.) The other would be a X11 style thing. There's no value whatsoever in having a hundred incompatible licenses. These undermine free software - if I can't incorporate CDDL code into an MPL program, or APSL code into a GPL'd app, then surely my freedom is restricted? As it stands, there are three major copy-lefted operating system kernels. You can't put Linux code in XNU or SunOS. You can't put XNU code in Linux or SunOS. You can't put SunOS code in Linux or XNU. But they're all "Free Software". You're supposed to be able to do whatever you want with the code as long as you don't impose further conditions. That's not how it works.

Re:That would be awesome!

[jvkjvk]

so I don't really see what Solaris would gain. I think Sun made a very clever choice with the CDDL for Solaris. It's Free, and the Linux guys can't just take the best bits and surpass them.

You don't want Linux to get any better because...? You don't want Linux to be better that Solaris because...??? I would think that as a user you would want the absoultely best OS and system you could have, no matter if it was Linux, Solaris, or even (gasp!) MS.

This attitude is incomprehensible to me.

Re:That would be awesome!

[TheRaven64]

Then you should really give Nexenta a shot

Interesting project. I'm more a fan of the BSD userland, but I'd take GNU over System V any day. Unfortunately, it seems to be only x86 at the moment (I run Solaris on my UltraSPARC). I find that if your code works on IA32 and SPARC64 then it will probably work anywhere. Still, it's something I'll keep my eye on (and probably play with in a VM). I remember how painful installing GNUstep on Solaris was (lots of dependencies that needed to be manually built and installed), and so having apt would be a huge advantage.

Some debian folks have legitimate trouble with the licensing (mixing GPL software with CDDL libc).

Odd. The GPL explicitly permits linking to standard system libraries (which is how you can run GPL'd software on Windows linked against msvcrt.dll). I wonder if there's a problem on the CDDL side, although there shouldn't be, since the CDDL only covers modifications to the CDDL'd files. My not-so-charitable guess would be that it's another example of the Debian folks arguing over nothing (although, to be fair, it tends to mostly be a few trolls on the Debian lists responsible for this kind of thing, rather than the real developers).

Re:That would be awesome!

[Orange Crush]

Would it not be possible to create a compatibility layer? After all, "the GPL workaround" has taken different-platform drivers and run them atop a GPL'ed shim to run them on Linux. If Solaris was GPL then there'd be no (legal) compatibility issues to introduce a "pretend I'm linux driver-mode" and share code.

Re:That would be awesome!

[maraist]

Now if we can get something like the Mac OS X type of GUI going with Linux on the desktop...

Thank you, but no.. I actually prefer the gnome user experience to XP / Mac. I would, however, like to see greater out-of-the-box installation / configuration / drivers so I don't have to trouble-shoot my parent's Linux boxes. Basic setup and support still suck.

Re:That would be awesome!

[BalkanBoy]

To your point, I found Ubuntu to be the most polished Linux desktop of all. Everything I cared for, works with little effort spent on getting it to work. The only thing that disappoints me is lack of drivers for Logitech webcams (I have the quickcam pro 5000 model). Logitech does not want to write the drivers for one reason or another ("not enough demand" I suppose).

Aw, that's just DANDY!

[Anonymous Coward]

That means that there will be that many fewer PS3's to buy!!!!!!

Re:Aw, that's just DANDY!

[Zonk (troll)]

Wrong thread?

Makes more sense than Java

[unPlugged-2.0]

I think putting OpenSolaris under GPL makes much more sense than doing the same for Java.

There are very good models of developing OS under GPL such as Linux. This would make it a very easy transition to get the GPL model into OpenSolaris.

Also this would no doubt attract a lot of the linux developers out there and as any open source project knows the number and management of developers is ultimately what makes it the top dog.

Also though it is slightly off-topic I also think that Java under GPL would not benefit as much because the model of contribution is really not as easily understood as the OS world.

Re:Makes more sense than Java

[Anonymous Coward]

I don't think java was released in order to gain more contributions - Sun seem to be able to cope quite well on that front. I think it was done in order to get java accepted in to Linux distros. I think Sun could see a threat from mono which despite being the spawn of Microsoft was GPL'd. And mono just got accepted into Gnome, which may have shaken Sun a bit since they use Gnome on Solaris. Mono is installed by default as a dependency of Gnome on the latest Ubuntu distro (Edgy).

Re:Makes more sense than Java

[RAMMS EIN]

``Also though it is slightly off-topic I also think that Java under GPL would not benefit as much because the model of contribution is really not as easily understood as the OS world.''

With all the complaints about the Java community process being slow and bureaucratic, and the free Java implementations lagging behind in features, I think having a good, open source Java implementation is a Good Thing in it's own right.

Also, I don't know what you mean by the model of contribution for Java not being as easily understood as the OS world. It's not like there aren't any successful open source programming language implementations yet.

Re:Makes more sense than Java

[rohan972]

Also though it is slightly off-topic I also think that Java under GPL would not benefit as much because the model of contribution is really not as easily understood as the OS world.

I may misunderstand the issues, but I think GPL Java impacts distribution a lot. It can now be included in Debian, Redhat etc. Contribution/modification will allow it to be ported without the same licencing problems, I think BSD's have had trouble with this in the past. The way in which contribution to the official java is handled will be secondary to this in some ways.

Another dumb move

[NineNine]

Sun has made some of the worst strategic decisions in the IT industry for the past decade or so. Why would GPL'ing their main products be any better? Where are they going to make money? Bake sales? Are they going to pay their people with warm fuzzy feelings? Yeah, Sun may do it. But I'm betting it'll kill them long term. In fact, now may be a good time to short the stock, and expect the payoff to be complete in about 5 years.

GPL'ing a product has NEVER been successful for the company or person owning it. This won't be any different.

Re:Another dumb move

[Anonymous Coward]

GPL'ing a product has NEVER been successful for the company or person owning it. This won't be any different.

According to you, its not been successful for Linus to GPL his kernel, or for Stallman to GPL emacs and gcc?

WTH you talking about?

mod parent down

Re:Another dumb move

[xzvf]

Sun does two things well. Rock solid hardware and excellent service. GPLing Solaris and Java allows them to limit resources spend on software development. In addition, GPL compatible Solaris and Linux will blur the lines between the OS as they adopt each other's best features. Linux and Solaris might become binary compatible. Sun can focus on selling hardware and services.

Re:Another dumb move

[Jeremi]

Where are they going to make money? Bake sales?

I'd imagine from selling computers, the same as they've always done. I don't think OpenSolaris is a profit center for them now, so GPL'ing it shouldn't cost them any profits (at least, not directly).

GPL'ing a product has NEVER been successful for the company or person owning it.

Ever hear of an a little OS called Linux? It's done fairly well under the GPL...

Re:Another dumb move

[Shaman]

Earth to NineNine: they haven't charged for Solaris or Java all along. They are a services and hardware company. If Solaris technologies move to Linux, then Sun has only to be sure that their hardware is the best supported Linux product to make a go of it. This is smart, good business and it's about time Sun figured that out. OpenSolaris won't be closed because for now it's got a lead on SPARC hardware as well as some features which are unique to Sun but over time it is obvious to all but the most clueless that Linux is where the community is putting the majority of effort and one would have to be a complete cluebie not to see that it is not slowing down or conceding defeat on any front.

Re:Another dumb move

[Zarluk]

Where are they going to make money?

If you find yourself in need for support in Java or Solaris, who do you prefer to ask for help? Micro$hit?

BTW: why is your web-site down for so long? Running Winblows?

LOL

Re:Another dumb move

[Anonymous Coward]

Ever hear of an a little OS called Linux? It's done fairly well under the GPL...

Haven't you hurd? It is called GNU/Linux :)

Re:Another dumb move

[RAMMS EIN]

Surely you're toking, mister Nineman.

Re:Another dumb move

[PCM2]

Sun does two things well. Rock solid hardware and excellent service.

Not to mention the fact that, although it is true that Sun is gradually open sourcing all of its software, most of what Sun makes it enterprise software. What company is really going to use Sun's RFID software to run a warehouse floor, or use Sun's identity management software to manage authentication and access control for an entire enterprise, and not get a support contract from Sun? Open sourcing this type of stuff probably doesn't impact Sun's sales negatively one iota. Open sourcing Java may be riskier, but I'm curious to see how it really pans out.

Re:Another dumb move

[drsmithy]

Sun has made some of the worst strategic decisions in the IT industry for the past decade or so. Why would GPL'ing their main products be any better? Where are they going to make money? Bake sales? Are they going to pay their people with warm fuzzy feelings? Yeah, Sun may do it. But I'm betting it'll kill them long term. In fact, now may be a good time to short the stock, and expect the payoff to be complete in about 5 years.

Sun will make money exactly the same way they do now - hardware sales and support contract.

Solaris has never been a significant revenue stream for Sun. Heck, they haven't even had a nominal charge on it for years.

GPL'ing a product has NEVER been successful for the company or person owning it. This won't be any different.

Well, it'd be pretty disastrous for a company with a revenue stream derived primarily from software sales, but that hardly describes Sun.

Re:Another dumb move

[Anonymous Coward]

Why, Mr. Ballmer! I never expected to see you here!

Re:Another dumb move

[molnarcs]

Well, it'd be pretty disastrous for a company with a revenue stream derived primarily from software sales, but that hardly describes Sun.

Not necessarily... Trolltech was under proessure to release QT under the GPL. They did it ... did it hurt their business? Not in the slightest... in fact, they became even more successful. Ninenine has an axe to grind with free software, that's all. His "GPL'ing a product has NEVER been successful for the company or person owning it" just shows how uninformed he is.

Re:Another dumb move

[QuantumG]

Or at least we might get dtrace on Linux :)

Re:Another dumb move

[Anonymous Coward]

Slightly wrong, they do actually charge for Solaris. The free (as in beer) download of Solaris (not OpenSolaris) is for non-commercial purposes only. So for me to use it instead of a linux box for my cvs repository and file server at home is fine, but using it as the webserver for my company would cost me a $99 per year licence.

I may switch to OpenSolaris for that, but I am tempted to buy the licence.

Re:Another dumb move

[that this is not und]

Actually, Solaris x86 used to be fairly expensive. I looked at buying a copy, 'just because' a number of years ago, before they turned Solaris 7 loose for free (the first free version). It was beyond my means as a non-commercial individual.

Re:Another dumb move

[drsmithy]

Not necessarily... Trolltech was under proessure to release QT under the GPL. They did it ... did it hurt their business? Not in the slightest... in fact, they became even more successful.

Poor example. QT is dual licensed (the GPLed version is only usable if you are developing GPLed software).

Ninenine has an axe to grind with free software, that's all. His "GPL'ing a product has NEVER been successful for the company or person owning it" just shows how uninformed he is.

Who derives a primary revenue stream only from selling GPLed software ?

Re:Another dumb move

[Shaman]

You're absolutely right, I chose the wording badly. But in recent years the only charge for Solaris has been for the CD packaging.

Re:Another dumb move

[Mr. Hankey]

Poor example. QT is dual licensed (the GPLed version is only usable if you are developing GPLed software).

Actually, this is an excellent example. It clearly shows that the GPL is by no means business unfriendly for the developers/copyright holders, only for others who wish to profit off their work without contributing back. Those who do not want to GPL their software can purchase a non-GPL license, whereas those who do wish to GPL their software can use it freely. If you think about it, it's something like a patent on source code that only expires when the copyrights do. People can benefit from your work through openness, but not hide it in a proprietary product without making a separate agreement with you.

Let's put it another way. The GPL is something like giving to the public domain with the associated good will from your customers, but without giving up your copyrights. That doesn't mean others won't do anything with your work mind you, there is only so much control and forks can happen, but they will also be GPL'ing their versions. You can take whatever code you like to integrate back into your original GPLed version, although not into a proprietary licensed version without agreement from the code's developer. You have little to lose from the GPL unless you want absolute control, but potentially much to gain if there are enough interested contributors.

Who derives a primary revenue stream only from selling GPLed software ?

Who derives a primary revenue stream only from selling software? Most software companies also sell support, either per incident or through a contract. Training tends to be quite lucrative as well. I find it humorous how much is spent on training and support for e.g. IBM/Rational and Oracle products somewhere that I won't mention, but I can't argue that it's a good business model for the vendors.

Re:Another dumb move

[drsmithy]

Actually, this is an excellent example.

Well, I agree that it *is* a good example, but not for proving the point you're trying to make.

It clearly shows that the GPL is by no means business unfriendly for the developers/copyright holders, only for others who wish to profit off their work without contributing back.

On the contrary, it conveniently shows that the GPL is not "business friendly", simply from the existence of a dual license. If the GPL was "business friendly", the dual licensing wouldn't be necessary.

The point here is that you can't make money selling GPLed software, without tying it to some other product.

You have little to lose from the GPL unless you want absolute control, but potentially much to gain if there are enough interested contributors.

What you have to lose from the GPL is the likelihood of ever selling your software. What you have to gain is the relatively remote possibility that other people will be nice enough to improve your product for free.

Who derives a primary revenue stream only from selling software?

Microsoft. Most game companies. There's no shortage of companies (or corporate departments) who derive most of their money from selling software (or software licenses).

Most software companies also sell support, either per incident or through a contract. Training tends to be quite lucrative as well. I find it humorous how much is spent on training and support for e.g. IBM/Rational and Oracle products somewhere that I won't mention, but I can't argue that it's a good business model for the vendors.

I didn't say companies who had _no_ other sources of revenue, I said companies who derive most of their money from software sales. The GPL makes that essentially impossible and, hence, removes those companies' business models. *THAT* is why TrollTech dual-licence QT - because they know if they didn't they would have a great deal of trouble making enough money to stay in business.

What I find ironic is that even the people who steadfastly insist the GPL is "business friendly" usually do so in the same breath they say "because you can just dual license" - seemingly unaware they're shooting down their own argument.

Re:Another dumb move

[iserlohn]

What are you trying to define as business friendly? Does "business friendly" mean that businesses can take the work of others freely and use them in products without giving back the improvements they made back to the original authors? If that is you mean then no, the GPL is not business friendly as it was designed to specifically counteract that.

The use of dual licensing gives the author(s) the best of both worlds. If you want to build upon the fruits of others, either contribute code under the same terms, or pay for a license from the original authors. Tit-for-tat. do ut des. blahblahblah

Re:Another dumb move

[jm91509]

Linux and Solaris might become binary compatible.

Opensolaris can already run linux distributions as a non-global zone. Its called brandz and involved providing the API's of a particular rev of the linux kernel (2.4.21???) aparently (I'm no brandz expert, just gave it a whirl). Tis nifty. You now have linux apps running on top of a solaris kernel unmodified. Looks just like a linux box. Not sure how useful it is though. They claim a 5% performance overhead and obviously hardware drivers are likely to be a bit odd as your not running linux at all but solaris.

http://opensolaris.org/os/community/brandz/

Re:Another dumb move

[drsmithy]

What are you trying to define as business friendly?

That would be the ability to, you know,*make money* from selling the product. "The product", in this case, being software.

Does "business friendly" mean that businesses can take the work of others freely and use them in products without giving back the improvements they made back to the original authors? If that is you mean then no, the GPL is not business friendly as it was designed to specifically counteract that.

The GPL is not "business friendly" because selling GPLed software is, for all intents and purposes, impossible unless it is tied to some other (non-GPLed) product.

The "business" under discussion here is selling software. Not selling hardware. Not selling support contracts. Not selling programming time. Not selling some other product for which you have written software. Selling *software*.

If you're having trouble understanding this, consider the video game industry. That is an example of a business model almost totally incompatible with the GPL[0].

The use of dual licensing gives the author(s) the best of both worlds.

Not really, since they would not be able use any of the user-contributed code in their non-GPLed code base (unless the contributors are happy to sign their copyrights away with the understanding that would happen - unlikely for anyone writing GPLed code).

If you want to build upon the fruits of others, either contribute code under the same terms, or pay for a license from the original authors. Tit-for-tat. do ut des. blahblahblah

As is normal with any attempt to discuss the downsides of the GPL, we end up at the "you just want to be able to use other people's work for free" strawman (an idea I have never in this discussion even *suggested*).

[0]Games with a significant subscription-style component (eg: WoW) being a notable example. Although - surprise - only because they are tied to some other non-GPLed product.

Re:Another dumb move

[Anonymous Coward]

What a bunch of codswallop.

Re:Another dumb move

[Anonymous Coward]

Trolltech (Qt, Qtopia), MySQL AB, Cyclades, Novell, Canonical, XenSource, ... and many others http://debianlinux.net/companies.html

Perhaps GPL'ing only works when both sides have a mutual benefit of progressing the free software and the business around it. A go-between with some non-free parts is also often used to put more pressure on purchases from professional customers.

Re:Another dumb move

[Mr. Hankey]

The point here is that you can't make money selling GPLed software, without tying it to some other product.

The dual license isn't another product. It's another set of terms through which another entity can distribute copies of your product, presumably with a different set of restrictions.

What you have to lose from the GPL is the likelihood of ever selling your software. What you have to gain is the relatively remote possibility that other people will be nice enough to improve your product for free.

You have the good will of your customers to gain. If your product is worth its salt, which not all are, you'll sell licenses in any case. Larger entities will definitely buy, and a large number of them will not use anything that does not have some sort of paid support structure.

Microsoft. Most game companies. There's no shortage of companies (or corporate departments) who derive most of their money from selling software (or software licenses).

Microsoft makes their share from publishing MCSE training materials as well. Most game publishers make their money from selling some (but definitely not all or even most) software, the whole thing being a gamble of who will buy which typically shelf life limited product first. Those that last for a longer time without languishing on the shelf tend to be those with an additional cost, such as WoW and EQ.

I didn't say companies who had _no_ other sources of revenue, I said companies who derive most of their money from software sales. The GPL makes that essentially impossible and, hence, removes those companies' business models. *THAT* is why TrollTech dual-licence QT - because they know if they didn't they would have a great deal of trouble making enough money to stay in business.

You said major sources of revenue, and both training and support can be major sources. TrollTech added the GPL to their product because their customers demanded it; the QPL was there before, and no doubt some customers were already using those terms who may not wish to use the GPL.

What I find ironic is that even the people who steadfastly insist the GPL is "business friendly" usually do so in the same breath they say "because you can just dual license" - seemingly unaware they're shooting down their own argument.

No, what's convenient about the GPL is that you protect your product from proprietary exploitation by your competitors in a modified form that you cannot use yourself. No one-upsmanship where you cannot follow, unless they pay you for the privilege. An additional license is simply an agreement between yourself and those who want different terms, with a different pricing tier.

Re:Another dumb move

[drsmithy]

The dual license isn't another product.

I never said it was.

You need to re-read what I'm writing. Because it's clear - either deliberately or accidentally - you are misinterpreting what I'm saying.

It's another set of terms through which another entity can distribute copies of your product, presumably with a different set of restrictions.

Yes. Terms which are created *because the terms of the GPL are not appropriate to the purpose of selling software*.

You have the good will of your customers to gain.

In most cases your "customers" - ie: people who pay for your software - couldn't care less whether or not it was GPLed.

If your product is worth its salt, which not all are, you'll sell licenses in any case.

Rubbish.

Basing a business on the idea people will voluntarily give you money just because you have a good product is financial suicide.

Larger entities will definitely buy, [...]

No business is going to pay good money for a product which can be acquired for free. For starters, it's fiscally irresponsible.

[...] and a large number of them will not use anything that does not have some sort of paid support structure.

In which case we're back to having to tie GPLed software to some other product like paid support to make it financially viable. Ie: the whole freaking point of this discussion.

Microsoft makes their share from publishing MCSE training materials as well.

Indeed they do - I never suggested otherwise.

The vast bulk of their revenue, however, comes from software sales. Moving down the tree to software companies which aren't lucky enough to have the size to invest in multiple markets, and the proportion of their revenue coming from software sales is even higher.

Most game publishers make their money from selling some (but definitely not all or even most) software, the whole thing being a gamble of who will buy which typically shelf life limited product first. Those that last for a longer time without languishing on the shelf tend to be those with an additional cost, such as WoW and EQ.

Where do game companies without subscription services like WoW derive revenue from if not software sales ?

How would you propose a game company make money from a GPLed game *without tying it to some other product* ?

You said major sources of revenue, and both training and support can be major sources.

Again with the straw man.

Look, I'm not sure how much simpler I can make this...

We are talking about SELLING SOFTWARE. Not selling software tied to support contracts, training, online worlds, hardware or anything else. SOFTWARE. The business model that SOFTWARE COMPANIES rely on to make money. That companies can also make money by selling support, training, access to subscriber-only services or hardware IS COMPLETELY AND UTTERLY IRRELEVANT TO THIS DISCUSSION. The point I am trying to get across, is that the GPL DOES NOT WORK for the business model of selling software. The only way to make GPLed software viable as a revenue stream is to tie it to some other non-GPLed product.

This is, I might add, a point you tacitly agree with, even while trying to argue against it. That's why every single counterexample you try to use relies on either a) tying the GPLed software to some other product to make it useful or b) dual licensing.

TrollTech added the GPL to their product because their customers demanded it; the QPL was there before, and no doubt some customers were already using those terms who may not wish to use the GPL.

TrollTech's customers - the people who actually pay them - aren't using the GPLed version. TrollTech GPLed QT to get a foot in the door of the Linux bandwagon - because selling software to typical Linux users, outside of a few highly specific examples, is nearly impossible due to their single-minded focus on the belief that anything no

Re:Another dumb move

[Mr. Hankey]

I never said it was.

You said that a company cannot make money from GPLed software unless they tie it to a non-GPLed product. This is false.

You need to re-read what I'm writing. Because it's clear - either deliberately or accidentally - you are misinterpreting what I'm saying.

I've read your writing, but it seems clear that you do not understand either the spirit or letter of the GPL. This is not surprising, the GPL is one of the most commonly misunderstood software licenses.

Yes. Terms which are created *because the terms of the GPL are not appropriate to the purpose of selling software*.

The terms are there explicitly to guarantee access to the source code, which is a different problem. It is not incompatible with selling the software. Let's be very clear here; the GPL is an open source license first, but its provisions do not prevent the software from being sold.

In most cases your "customers" - ie: people who pay for your software - couldn't care less whether or not it was GPLed.

Larger customers, with deeper pockets, do indeed care whether they have the source to a product which they wish to deploy. Business code in particular, which is amongst the most widely developed code out there, but other software which might be useful as well.

Basing a business on the idea people will voluntarily give you money just because you have a good product is financial suicide.

This is no different whether the software is open source or not. You still have to sell it to your customer, and if they believe your product to be worthless then they won't buy it. If your product is really that good, people will pay for it. Most everyone wants a say in their favorite software product's development, and in the case of open source software in particular features can definitely be purchased.

In which case we're back to having to tie GPLed software to some other product like paid support to make it financially viable. Ie: the whole freaking point of this discussion.

And your point is not valid. We'll go over why in the rest of this message.

The vast bulk of their revenue, however, comes from software sales. Moving down the tree to software companies which aren't lucky enough to have the size to invest in multiple markets, and the proportion of their revenue coming from software sales is even higher.

I don't think you have any idea of what you're talking about. Most software companies will gladly offer high priced service agreements to corporations that use their services, and these companies have a better chance of seeing their software used in these very lucrative environments. Many large and small software companies are more than happy to take money in exchange for support if an entity is willing to spend it. If yours is not... Well, your competition probably is.

Where do game companies without subscription services like WoW derive revenue from if not software sales ?

Game publishers gain some quick revenue from software sales while the product is still hot, but a lot of games do not in fact make money. (*cough* Daikatana *cough*) Plenty of game companies go out of business regardless of the license of their software, because the gaming industry's customers are fickle. Licensing and merchandising of gaming franchises (toys, TV shows, breakfast cereals and the like) make up a lot more of their profit than they might be comfortable with. For that matter, the majority of what compromises a game these days is the content and not just the code. In that light, games really should fit the GPL model if your assertions are correct. Open sourcing the code (as has happened for a number of games after release, particularly those from id software) does not prevent a company from making money when done properly. People still buy id games that have been open sourced, even if only for nostalgia.

We are talking abo

Re:Another dumb move

[drsmithy]

You said that a company cannot make money from GPLed software unless they tie it to a non-GPLed product. This is false.

Yes, that was "cannot" in the colloquial sense of "well, maybe you _could_ make money off it, somehow, if you were incredibly lucky, but for all practical purposes, it's impossible".

I use just "cannot" because it's a lot quicker and easier to type, and the meaning is blatantly obvious from context to anyone not trying to be a pedantic arsehole.

I've read your writing, but it seems clear that you do not understand either the spirit or letter of the GPL. This is not surprising, the GPL is one of the most commonly misunderstood software licenses.

I'm pretty sure I've got a good handle on the spirit, definition, intent and harsh reality of the GPL.

The terms are there explicitly to guarantee access to the source code, which is a different problem. It is not incompatible with selling the software. Let's be very clear here; the GPL is an open source license first, but its provisions do not prevent the software from being sold.

Literally, no. Practically, yes.

Larger customers, with deeper pockets, do indeed care whether they have the source to a product which they wish to deploy.

Ignoring for a second the massive quantity of proprietry software used by businesses the refutes this assertion, "GPLed code" and "customer access to source" are not even remotely close to being synonymous.

Business code in particular, which is amongst the most widely developed code out there, but other software which might be useful as well.

Most "business code" isn't GPLed and, indeed, GPLing it would (rightly) be seen as sacrificing a competitive advantage.

This is no different whether the software is open source or not. You still have to sell it to your customer, and if they believe your product to be worthless then they won't buy it.

Indeed. The major difference is when *do* like it, and can (legally) get it for free, they're significantly less likely to pay for it out of generosity.

If your product is really that good, people will pay for it. Most everyone wants a say in their favorite software product's development, and in the case of open source software in particular features can definitely be purchased.

Reality intrudes at this point and reminds us that the vast majority of people will not pay for a product or service they can also get for free (or significantly cheaper).

The problem with selling GPLed software is that your first customer can quite legally resell copies at a lower price than you, and/or make copies freely available. Exactly how do you propose to run a business selling software when that can happen ?

I don't think you have any idea of what you're talking about. Most software companies will gladly offer high priced service agreements to corporations that use their services, and these companies have a better chance of seeing their software used in these very lucrative environments. Many large and small software companies are more than happy to take money in exchange for support if an entity is willing to spend it. If yours is not... Well, your competition probably is.

Something I have never disputed, but is completely irrelevant to the topic of this discussion, which is SELLING SOFTWARE. Not selling software tied to support contracts, hardware, additional content or subscription services.

I am well aware you can build a successful business around selling GPLed software tied to support contracts and the like. The point I am trying to get across - and have re-stated numerous times in this discussion - is that *WITHOUT* tying the GPLed software to these additional products or services, building a business around selling it is, for all intents and purposes, impossible.

Game publishers gain some quick revenue from software sales while the product is still hot, but a lot of games do not in fact make money. (*

Re:Another dumb move

[Mr. Hankey]

Yes, that was "cannot" in the colloquial sense of "well, maybe you _could_ make money off it, somehow, if you were incredibly lucky, but for all practical purposes, it's impossible".

Your interpretation, but it's a flawed interpretation which is based on your limited definition of selling software.

I use just "cannot" because it's a lot quicker and easier to type, and the meaning is blatantly obvious from context to anyone not trying to be a pedantic arsehole.

I'll avoid responding in kind, however you did say that it cannot be done. Clearly this is not true. If you meant unlikely, that would be a different word.

I'm pretty sure I've got a good handle on the spirit, definition, intent and harsh reality of the GPL.

I'm pretty sure you have a handle on what you think it means; that doesn't mean you understand the GPL. It's unlikely if you think the reality is harsh, the GPL is a reflection of the reality of copyright laws.

Literally, no. Practically, yes.

Yet it still happens.

Ignoring for a second the massive quantity of proprietry software used by businesses the refutes this assertion, "GPLed code" and "customer access to source" are not even remotely close to being synonymous.

You also ignore the tremendous amount of GPL software used by businesses, particularly technology companies and large corporations in their network infrastructure. Indeed, all that business code is supported and wouldn't be where it was unless it was.

Most "business code" isn't GPLed and, indeed, GPLing it would (rightly) be seen as sacrificing a competitive advantage.

That depends on which business code; As I had mentioned, there are some applications for which the GPL does not make sense. There is a tremendous amount where it does however, particularly in business infrastructure.

Indeed. The major difference is when *do* like it, and can (legally) get it for free, they're significantly less likely to pay for it out of generosity.

Smaller entities are indeed less likely, but a large number of them are unlikely to buy them even if they are not free - how much software do you suppose is pirated? These are not the customers that generate revenue. The world has plenty of people who will not pay for software regardless of the license.

Reality intrudes at this point and reminds us that the vast majority of people will not pay for a product or service they can also get for free (or significantly cheaper).

Again, these are not generally the sort that would pay for software in the first place. A business model for software includes presenting a value proposition for customers, including support and training. If you take this away from your software, you're not likely to get very far.

The problem with selling GPLed software is that your first customer can quite legally resell copies at a lower price than you, and/or make copies freely available. Exactly how do you propose to run a business selling software when that can happen ?

It can be redistributed, this is true, but your name stays on the software. If you do not provide a value proposition to your customers, they will indeed not pay you. If you can provide support and training, you'll make money. Cheapbytes sells CDs from several Linux distributions, yet the distributions continue to exist. This is because they provide a value proposition in their product that Cheapbytes does not.

Something I have never disputed, but is completely irrelevant to the topic of this discussion, which is SELLING SOFTWARE. Not selling software tied to support contracts, hardware, additional content or subscription services.

Your difficulty seems to be that you feel a need to separate them. If you've tried selling software without a support structure then I'm not surprised you failed. The customer who actually pays money for your product expects support of some sort to e

Re:Another dumb move

[drsmithy]

Your interpretation, but it's a flawed interpretation which is based on your limited definition of selling software.

No, it's my interpretation based on what "selling software" actually is - and "selling software" is not "selling support contracts" or "selling hardware" or the like.

I'll avoid responding in kind, however you did say that it cannot be done. Clearly this is not true. If you meant unlikely, that would be a different word.

Indeed it would, although "unlikely" is not a strong enough word.

The meaning of my statement is quite obvious in context. Again, only those trying to avoid the issue by constructing a counter-argument based solely on pedantic observation would misinterpret as you did.

I'm pretty sure you have a handle on what you think it means; that doesn't mean you understand the GPL. It's unlikely if you think the reality is harsh, the GPL is a reflection of the reality of copyright laws.

The GPL is far more than that.

Yet it [selling GPLed software] still happens.

That doesn't make it a viable business plan.

You also ignore the tremendous amount of GPL software used by businesses, particularly technology companies and large corporations in their network infrastructure. Indeed, all that business code is supported and wouldn't be where it was unless it was.

No, I don't. Your flawed assumption is that a significant proportion of companies are using that code _because_ it is GPLed. Ie: because of the unique aspects of the GPL.

Most companies support their business code internally. Whether or not it is GPLed is pretty much irrelevant.

That depends on which business code;

I think you need to define what you mean by "business code" here. I am working under the assumption that "business code" means code developed *by the business* to help its workflow.

I am beginning to think by "business code" you mean "software used by business", which would change the situation markedly. What _do_ you mean by "business code" ?

Smaller entities are indeed less likely, but a large number of them are unlikely to buy them even if they are not free - how much software do you suppose is pirated?

A lot, but taking away the assumption of legal behaviour changes the discussion significantly. For example, a major "benefit" you extoll for the GPL (modified code redistribution) disappears.

These are not the customers that generate revenue. The world has plenty of people who will not pay for software regardless of the license.

Stop avoiding the issue. Most people will not pay for something they can legally acquire for free. This is basic human nature.

Again, these are not generally the sort that would pay for software in the first place.

Utter rubbish. A basic tenet of human nature is that people will seek to "pay" the least possible for goods and services. Very few people are prepared to pay a given amount for something they could otherwise acquire for free, or for less. Heck, the whole idea of the free market hinges on this concept, yet you are blithely asserting it doesn't exist ?

A business model for software includes presenting a value proposition for customers, including support and training. If you take this away from your software, you're not likely to get very far.

The vast amount of profitable proprietry software that isn't based around a business model of ongoing support and training payments disagrees with you.

Quite frankly, the proportion of software that is sold on the proviso of ongoing payments for support, training, etc, is quite small (although large in $$$ terms) and targeted at a relatively tiny proportion of buyers.

It can be redistributed, this is true, but your name stays on the software.

Unfortunately a sated ego doesn't pay the bills.

If you do not provide a value proposition to your customers, they will indeed not pay you. If

Re:Another dumb move

[Mr. Hankey]

You may have noticed that this is getting a bit unwieldy, I'm going to try to pare this down to the basic points. I do have a life and family outside of Slashdot, despite the length of this debate.

  A small minority of users derive benefit from being able to customise the source to their needs, but even those would be just as well served by the BSD or a form of "shared source" license, as they typically have little interest in - if not outright hostility towards - redistributing their modified code. That is because that modified code typically represents an advantage that company has (or perceives it has) over its competitors.

You've come very close to the intent of the GPL in this statement. That part about a company modifying a separate copyright holder's code and then not wanting to distribute the source. If the company never distributes the binaries, this is fine as the GPL doesn't cover modification unless the software is distributed. If they do not hold the copyright, AND do not want to distribute the source, they're out of luck for distributing the binaries. That's precisely what makes the GPL business friendly, this protection that another company will not do exactly this while preventing the original copyright holder from seeing the modifications. If they want to do this, they're going to have to ask the copyright holder. Given, it may be unfriendly to such a company, but that's copyright.

  The point I have repeatedly tried to make - and which has been ignored, avoided or brushed aside at pretty much every turn - is that the GPL makes the business of *selling software* practically impossible.

I'll explain it again. If you feel that support is not implied when software is sold, then it may indeed be difficult. If you tell your customers that they're not going to get support or updates of any sort with your application, I'd wager your sales are going to be very low regardless of license. GPLed software is not unique in this respect. Support includes making it function correctly and keeping it secure, extended support may include additional help with making the software do what the user wants. Software provides no value to your customer if they aren't able to use it. If your software doesn't work after your customer clicks on it, and you do not give any form of support (a knowledge base is a form of support) they're not going to be coming back for more. They might demand a refund, they could even try to sue your company.

As for how I define business code, I speak of the software at the core of the business which keeps operational processes functioning. Some of it is custom code, some of it libraries and frameworks, some of it (such as CRM software) may be purchased from a vendor or be GPLed. This is an aside from the point, but you asked and I won't leave you hanging.

Excellent

[99BottlesOfBeerInMyF]

This could be a bigger boon than a lot of people realize. The licensing differences between Solaris and Linux are one of several factors slowing them from adopting ideas and code from one another. OpenSolaris users could benefit from ease of importing more cutting edge features from Linux. Linux could benefit by having access to some of the cleaner implementation ideas from Solaris. I've felt for some time that much of what holds linux back is the unwillingness to adopt newer and better features out of a fear that a given distribution will be less compatible with others and because Linux is trying to wear many hats. Too many decisions are made to benefit its use as a server or make it easier to use on a portable, while leaving it behind others for a workstation.

I'm keeping my fingers crossed.

Re:Excellent

[0racle]

And in the end one of them dies because they're exactly the same thing. I really don't see what the point of GPLing OpenSolaris would be.

Re:Excellent

[jericho4.0]

"OpenSolaris users could benefit from ease of importing more cutting edge features from Linux."

Linux would get DTrace, ZFS, etc. Those techs are about as cutting edge as it gets. What would Solaris get?

Re:Excellent

[dunstan]

As a long term Solairs SA, I can assure you that they're nothing like the same thing. While the OpenSolaris and Solaris Express releases are fluid, the GA release (at present, Solaris 10) is not. Sure, new functionality is added during the life of a major Solaris version (most recently, ZFS was added), but the existing published kernel API will not change. This means that device drivers and other software which links into the kernel (e.g. storage software) will continue to work.

This is not a minor issue. The big headache with any GNU/Linux distribution is the complicated and often intractable support matrices - a result of the fluid nature of the kernel API.

Re:Excellent

[Anonymous Coward]

This is not a minor issue. The big headache with any GNU/Linux distribution is the complicated and often intractable support matrices - a result of the fluid nature of the kernel API.

Big headache for closed drivers. Tough shit on them. The Free ones get fixed when the API changes. The kernel developers have the right idea: you have our source... why should we help you hide your kernel source and in the process hamper our own development.

Re:Excellent

[obi]

a host of filesystems, maybe? truckloads of drivers?

Re:Excellent

[MechaShiva]

Beat me to it. I'd be happy to give Solaris a spin round the block but the driver support for non-sun branded hardware has left a lot to be desired. Granted, Sun doesn't have any incentive to make this a top priority and I in no way expect them to. All I'm saying is it would be fantastic to see Linux drivers opened up to the Solaris devs and the Solaris features opened up to the Linux devs. That would really create a cooperative competition (if there is such a thing). Here's to hoping.

Re:Excellent

[Anonymous Coward]

Virtual consoles, better pcfs, USB drivers, other drivers, ...

Re:Excellent

[TBBle]

Debian GNU/OpenSolaris. (cf. Nexenta at http://www.gnusolaris.org/gswiki and Debian's non-Linux ports at http://www.debian.org/ports/#nonlinux)

Re:Excellent

[afidel]

The problem with the driver argument is it helps Sun in no way. Sun is a hardware company, they buy lots of hardware components and get driver code when they do so that they can support the hardware they sell. There are some third party peripherals that might not be supported but those aren't generally the focus of Sun's customers.

Re:Excellent

[Anonymous Coward]

Linux would get DTrace, ZFS, etc. Those techs are about as cutting edge as it gets.

Tracing and a transactioned file system? Those are about 30 years old and Linux has multiple implementations of both. Why would it need another one?

What would Solaris get?

Lots of hardware drivers and much better compatibility with different file systems and other standards. That's far more important than yet another tracing facility and yet another file system.

I'm not complaining if Sun open sources "Open" Solaris, but don't expect it to make much of a difference to anybody. IBM open sourced many of their "enterprise features" and donated them to Linux, and most of them were received with a yawn.

Re:Excellent

[jambarama]

What would Solaris get?

Drivers.

Re:Excellent

[milliwatt]

> And in the end one of them dies because they're exactly the same thing.
> I really don't see what the point of GPLing OpenSolaris would be.

Maybe so the one that dies out isn't OpenSolaris?

I posted about this elsewhere in this discussion, but what if OpenSolaris goes GPL3? Along with GCC, libc, GDB, Emacs, etc. etc.? Linux is very likely *not* going GPLv3. Which kernel do you think would fare better as the free software kernel of choice to be distributed with GNU?

Re:Excellent

[Anonymous Coward]

A bogload of drivers for *current* hardware.

Re:Excellent

[hritcu]

I posted about this elsewhere in this discussion, but what if OpenSolaris goes GPL3? Along with GCC, libc, GDB, Emacs, etc. etc.? Linux is very likely *not* going GPLv3. Which kernel do you think would fare better as the free software kernel of choice to be distributed with GNU?

It depends on the name they will give it. If it were named GNU/Solaris then things would be surely settled ;)

Re:Excellent

[sanyam_y]

What would Solaris get? Well didn't it get Mozilla and GNOME?

Re:Excellent

[Anonymous Coward]

Nope, what really happens is that maintainers get fed up with fixing drivers, and they rot in the tree until eventually they are removed. The worst thing about the outright refusal to have a properly versioned API in the kernel is that the API can break in subtle ways - the function signatures stay the same, but the results don't, meaning that your driver may still compile but doesn't work. The lack of a versioned API is just sheer laziness on the part of Linus and friends. The BSD's manage this kind of thing, along with shims for backwards compatability at the system call level, so why can't Linux?

Re:Excellent

[Anonymous Coward]

OpenSolaris users could benefit from ease of importing more cutting edge features from Linux.

Whoops, you should have proofread that before submitting it - you've got OpenSolaris and Linux the wrong way round.

Re:Excellent

[Alioth]

No - the only headache is for users. People with closed source drivers (like nvidia) simply provide an open source glue layer. An unstable kernel API does not stop closed source drivers.

All it does is inconvenience users. If you have a minority piece of hardware, even with open source drivers - each kernel upgrade that comes down the pike when you do a security update on your distro means you have to recompile and reinstall the drivers (and then hope they still work). The Linux "stable" kernel is a fiction - 2.4 is probably the most egregious example of this - it was still a development kernel, but just at a slower rate.

Re:Excellent

[Anonymous Coward]

Bullshit. This is about kernel API changes. When the kernel device API changes, the person who changed it is REQUIRED to change all the device drivers. Bugs are different discussion.

Re:Excellent

[eMbry00s]

developers developers developers developers developers;
developers developers developers developers developers! developers! developers!




Really, though. It's bound to be true, in time. (and bug reports!)

Re:Excellent

[jericho4.0]

"Those are about 30 years old and Linux has multiple implementations of both. Why would it need another one?"

You would seem to know little of DTrace or ZFS. Comparing them to what's available on Linux is silly.

Re:Excellent

[cparker15]

Using GNU/kOpenSolaris would enable GNU users to tell the Linux fanboys to take a long hike by helping to settle the long and drawn-out GNU/Linux naming controversy. (Why GNU/kFreeBSD hasn't already done so, I don't know.) Maybe that will let more people know that there are kernels that can be used with GNU other than Linux. Which, in turn, might possibly attract some more people to the HURD, which might get the HURD released a little sooner and get the entire GNU OS packaged and out the door. That makes me smile. :)

Re:Excellent

[Anonymous Coward]

As a long term Solairs SA... ...complicated and often intractable support matrices...


You sure? Not had to fight the complicated and intractable SunSOFT support recently?

p.s. I work for one of the few organisations which negotiated direct purchase/support through Sun rather than a reseller (coz we buy enough kit to bully Sun into doing that). Sun support is better than say HP, but still qualifies as a fscking nightmare.

Great, but will it change anything?

[the_humeister]

It certainly removes one barrier. But look at Darwin. It's open source, but who else but a handful of people outside of Apple are working on it? So the point is not to knock the potential change. The point is will developers flock to Solaris as a result of this? Slowly but surely or not fast enough?

Re:Great, but will it change anything?

[davecb]

Actually, quite a number of folks who are my consulting customers use Darwin (really BSD) sources as the "reference copies" of programs they're adapting for their own use.

This is in part because of the good quality of the code, and the company which stands behind it. In part it is because of the larger BSD community who stands semi-invisibly behind Apple... some customers really understand the strength of community. And finally, for the license-paranoid, in part this is because of the use of the very old and weak BSD license.... some customers really don't understand the community (;-))

Coming back to the main point of the discussion, adoption of the GPL by well-known fortune-500 companies is a step away from the world of Microsoft, SCO and FUD.

Definitely a change, and definitely for the better.

--dave

Re:Great, but will it change anything?

[Toveling]

Apple released Darwin under the APSL, which is not GPL comaptible. The difference here is that Solaris code tenatively will be able to be used directly in Linux and other GPL projects, something not possible with Darwin or Solaris currently.

Re:Great, but will it change anything?

[Anonymous Coward]

There is a BIG difference between "open source with a crappy license" and "GPL'd free software" :-P

Re:Great, but will it change anything?

[squiggleslash]

I thought Darwin was licensed exclusively (well, aside from that parts that have to be licensed under the GPL) under the APSL? Did they keep the BSD license for those parts comprising exclusively of BSD code?

Re:Great, but will it change anything?

[davecb]

This seems to have changed fairly recently: I suspect the license-challenged are researching what part is pure BSD by now (;-))

--dave

Yeah sure...

[Serapth]

According to Jonathan Schwartz, the decision of Novell and Microsoft to '(suggest) that free and open source software wasn't safe unless a royalty was being paid' is what prompted Sun to finally come down on using the GPL for Java. So I guess every cloud has a silver lining. If you believe that, want to buy some old dot com stocks I traded for some swamp land a few years back? Honest, ill give you a great deal!

A company the size of Sun does not move that quickly, especially so far as legal matters go. Besides, there has been talk of GPLing Java before Christmas for months.

Sun saw a chance to take a shot at Microsoft/Novell and they took it. Can't say I fault them, but its fairly obviously a lie.

Re:Yeah sure...

[DragonWriter]

A company the size of Sun does not move that quickly, especially so far as legal matters go. Besides, there has been talk of GPLing Java before Christmas for months.

There's been talk of open sourcing it by Christmas, and reports that it might be under the GPL (and reports that it might be under a different license.)

That does not prove, however, that the Novell/MS deal didn't prompt the final decision for Sun. Certainly, they'd already done the analysis and had a pretty good idea of the pluses and minuses of the various options. But certainly the Novell/MS deal remixed those slightly, and might have tipped things in the GPL.

Re:Yeah sure...

[ronanbear]

A Novell/Microsoft patent case against Linux has something else to think about now. I don't think Sun are in any rush to GPL Solaris but they're letting Microsoft know that Sun are the big winners in a big legal case between Windows and Linux.

They're basically saying "go ahead, drive an army of developers and users to us, see what good that does ya".

Re:Yeah sure...

[Anonymous Coward]

I agree. There is no way they were sitting around last week deciding on the license. Similarly to what you said, I think he had the opportunity to contrast the MS/Novell deal with the Java re-license and he took a bit of a cheap shot. Though, perhaps he's trying to say that the idea that "free has to have a price" was the motivation. It's not like it has to be accurate to make the front page of slashdot. Whatever. From his blog:

Picking a license was a very complex task - we took an enormous breadth of issues to heart in making the selection, from protecting our customers and licensees, to continuing to foster a wildly successful developer community. We had to worry about device manufacturers, media standards, big enterprise systems, government and military deployments - remember, more businesses and devices leverage Java than any other development platform. This was no simple feat.

So to the legal team at Sun, and our friends at the Free Software Foundation - I would like to offer my heartfelt thanks. We could not have gotten here without you. If Shakespeare had understood intellectual property, he never would have said all those mean things.

By admitting that one of the strongest motivations to select the GPL was the announcement made last week by Novell and Microsoft, suggesting that free and open source software wasn't safe unless a royalty was being paid. As an executive from one of those companies said, "free has to have a price."

That's nonsense.

...

Re:Yeah sure...

[squiggleslash]

I read the sentence as saying that the Novell action pushed Sun over the edge, not that they made the decision last week after rejecting the GPL completely before that. Sun has been considering the GPL for a long time, and Schwartz has gone out of his way to indicate the GPL was on the table when others were claiming it could never happen. I've even commented upon Sun's support for the GPL before.

It's quite possible that Sun would have gone for it anyway. The question is, would it have been this week, and is it possible they'd have waited for GPL3 before making the final decision?

Wow.

[Zonk (troll)]

Thanks Sun, that would be great if OpenSolaris became available under the GPL along with Java. That would offer another very good open alternative to Linux. Something backed by Sun would be really good if Microsoft keeps up it's patent-FUD.

Did I wake up in some alternate universe or something, though? I mean, Novell sold the community out and now Sun is adopting the GPL?

Re:Wow.

[that this is not und]

Issues of licensing culture aside, Novell has always be despicable, and Sun has always been cool. It's only because of judgements by various political factions amongst the Linux community that this has ever been in doubt.

And, uh, the Linux community doesn't define 'cool.' Just one or a few flavors of cool.

Re:Wow.

[squiggleslash]

Novell sold the community out and now Sun is adopting the GPL?

Actually I've never understood the hostility against Sun. Novell's actions appear to be surprising, but I think at the moment we don't actually have the whole story.

Sun on the other hand have made their version of Unix Free Software. They've placed most of StarOffice under the GPL, under the name OpenOffice.org. There's always been a large body of code that's licensed under relatively liberal Free Software licenses from them, including one of the first X11 desktops. And they've developed, and kept open and royalty free, major standards such as NFS, NIS, and others.

So why people diss Sun and, as an example, worship Apple, who really do seem to have only a token "Open Source" effort which, for the most part, seems to exist as a legal and PR thing, is something I've never understood. Yes, Sun took their time over Java, but they've never stamped on independent re-implementations, and their reasons for being careful were always very obvious and very understandable. Sun has always appeared to me to be the real deal, a company that does contribute more than it takes, and always has.

Did opening Solaris do anything?

[MBCook]

To me this sounds like a simple off-hand comment and unlikely to happen.

That said, can someone who is more familiar with the whole thing tell me: did has opening Solaris had much of an effect at all in any way? Has it stopped market share loss? Increased market share? Increased software availability? Has anything really changed?

Re:Did opening Solaris do anything?

[anlprb]

Depends on what you mean by anything. If you mean has there been a large takeup of drivers or features, no. But that does not have to do with the license. It has more to do with the snail's pace that Sun's development process follows. The whole CAB thing is interesting, but you cannot get a feature into OpenSolaris without it being sponsored by a Sun employee. It is still and if Sun has anything to do with it, continue to be, a very cathedral approach to things. Now, it does slow down development which allows for more testing, better architecting and there is a ton of discussion, re: ksh93, that brings out ALL of the nitty gritty details of the ripple effect through the product. But it also shows one of the chinks in the OpenSolaris armor, fear of change. Fear of change is not a bad thing when betting tons of money, however, attempting to produce a more useful operating system, or "Operating Environment" as Sun likes to put it. In this day and age the need to be nimble and allow new features to be merged in a time frame that does not discourage contributors. The snails pace which features are argued back and forth is fine when you have 8 hours a day, 5 days a week, 52 weeks a year to work on a single feature. When you only have the off time that the programmer is working on this project, it becomes unmanageable. This is very similar to the way OpenDarwin worked. It seems to be running into similar problems, however, I will say that Sun employees are very willing to talk with anyone who has an idea. They are also very good, sometimes, at letting those on the outside know what is going on on the inside. There still is though very much a feeling of "We out here are only get to see part of the development process". This is manifested by defects not having all of the information necessary for reproducing the issue, or investigation notes. There are fields that are not visible to the outside world in the database. This limits the transparency with which the process can run. There are two types of developers, those inside Sun and those outside. These two classes have very different means available to them. It hampers some investigation and correction, especially when a defect refers to the information contained in a field that only Sun employees can see. The only thing that I could see fixing this is GPL'ing the code which would allow a completely non-Sun controlled version of Solaris to be built. This would remove the have/have-not barrier for developers and would allow cross-polination of good ideas. Driver support would also be a bonus. As it stands right now, only pure BSD drivers can be pulled over. But overall, it still very much seems as if the OS community is on a factory tour of an Operating System factory. We get to look through the glass, but we really can't touch the machinery.

Re:Did opening Solaris do anything?

[that this is not und]

The whole CAB thing is interesting, but you cannot get a feature into OpenSolaris without it being sponsored by a Sun employee. It is still and if Sun has anything to do with it, continue to be, a very cathedral approach to things.

Try to get a feature into the Linux kernel without having connections to somebody deeply involved in the Linux kernel development team.

And calling something 'very cathederal' has mixed meaning, considering the fact that 'The Cathederal and the Bazaar' was originally written as a polemic against the fairly 'closed' team developing Gnu Emacs. I bet you thought Raymond wrote it to criticize Microsoft.

A LOT of cool code that is ultimately released in a very open fashion comes out of 'Cathederal' style development teams.

Re:Did opening Solaris do anything?

[Anonymous Coward]

Tbere is nothing stopping anyone from forking OpenSolaris and create a completely non-Sun controlled version. Nothing in the CDDL prevents that. You can do it today.

If Sun GPLs OpenSolaris, it will simply be an alternative license for OpenSolaris, along with the CDDL. Whatever hurdles you have to jump through to get a change in to the official core will be the same hurdles for the GPL version as well, since the versions and community will essentially be identical.

GPL OpenSolaris simply makes integrating Solaris code in to Linux easier, but cross pollinating will porobably not happen as easily from Linux to Solaris unless the entire contribution and commiters are covered by the SCA that Sun requires.

Re:Did opening Solaris do anything?

[Alioth]

But that does not have to do with the license. It has more to do with the snail's pace that Sun's development process follows. The whole CAB thing is interesting, but you cannot get a feature into OpenSolaris without it being sponsored by a Sun employee

The thing is - Solaris has a stable kernel ABI. It won't be necessary to have your drivers (the bulk of extra development that Open Solaris needs) sponsored because it'll be perfectly possible to develop a driver for your hardware, package it up with your hardware (or put it on your web page for download) and know it'll work without needing to be recompiled. Since Linux lacks a stable kernel API, it's essential to get accepted by the Linux developers and have your driver accepted by the developers, too - because otherwise your users are stuck with the inconvenience of rebuilding the drivers with every minor kernel update that comes down the pike from their distro maker.

I suspect if a GPL'd Open Solaris happens and starts gaining any sort of traction with hardware vendors thanks to the stable kernel API, Linus will change his tune on that for the Linux kernel in pretty short order.

Re:Did opening Solaris do anything?

[squiggleslash]

It's not an off-hand comment. Schwartz has said publicly Sun are considering GPLing Solaris before. The key thing is it appears they're waiting for GPL3 before making a final decision.

I think the weeks that follow the release of version 3 of the GPL will be very interesting from the point of view of people interested in either Solaris or Java.

Re:Did opening Solaris do anything?

[squiggleslash]

That really depends on whether the stable ABI is abused by vendors putting out binary-only drivers. If it is, then I doubt any changes will be made to the policies governing Linux in that area. If, on the other hand, tons of new drivers that abide by the terms of the GPL newly appear for SunOS, then a switch might happen, but probably after it's too late to stop the Solaris juggernaut from running over GNU/Linux completely.

Solaris - GPL'd?

[sloanster]

Wow, if that ever happened, wouldn't it be ironic - I can imagine a future where linux has been effectively preempted by the mega corporations, while Solaris is fully GPL'd and becomes the default first choice for the typical savvy unix admin.

In any case, I'll be getting to know Solaris 10 better in the coming months, but the GPL would just put it over the top.

Re:Solaris - GPL'd?

[foobsr]

I can imagine a future where linux has been effectively preempted ...

It is like attacking the enemy from the front (Novell/M$ with a SCO preevaluation) and alongside (addressing developers attention) as well. Now wait who will further back and patronize (IBM?).

CC.

Re:Solaris - GPL'd?

[Anonymous Coward]

Careful there, fanboy, you're starting to drool.

SUN GPL'ing OpenSolaris?

[Anonymous Coward]

Oh lordy, then what's the RMS fan club going to bitch about?

Re:SUN GPL'ing OpenSolaris?

[cortana]

Flash, proprietary drivers and patent-encumbered codecs.

Re:SUN GPL'ing OpenSolaris?

[Anonymous Coward]

>Flash, proprietary drivers and patent-encumbered codecs.

Here's a nickel, kid. Go buy a sense of humor. Oh wait, if we bitch long enough, maybe someome will GPL them too!

More cool stuff in Solaris?

[Kadin2048]

This is just my understanding, but Darwin doesn't have nearly as many 'interesting' features that don't already exist in Linux. There is some neat Solaris-only stuff that people have wanted to bring into Linux for a while, but have not been able to because of licensing problems, and the work it would take to clean-room it.

The thing that I always hear talk about is dtrace (currently CDL, and tightly integrated with the Solaris kernel), but looking at the WP article on it, apparently it's been partially brought over to BSD and OS X. Then there are also containers and that "self-healing" fault-isolation system, which I don't pretend to understand.

Perhaps there are just as many cool, compelling features in Darwin that aren't talked about, and deserve being shared with Linux and other OSes ... but I've definitely not heard as much 'buzz' about them as you hear about some particular features of Solaris that are supposedly very neat.

Re:More cool stuff in Solaris?

[drinkypoo]

There's nothing interesting in darwin. Anything neato that MacOSX does is done in a service, not in the kernel. Darwin's kernel is a combination of Mach and BSD called XNU. Mach is a pretty lame microkernel, as microkernels go... which is why the hurd is going from mach to L4. Of course, some would say the hurd is pretty lame, and that it deserves mach, but that's another conversation. The point is that there's no compelling reason to do anything with it, and many reasons not to touch it.

Re:More cool stuff in Solaris?

[DragonWriter]

I seem to have heard that GNU/Hurd development was moving from L4 to something else...(checks)...Wikipedia seems to be confused as to whether its moving to "Coyotos" or whether there is only some talk about that, and of course the GNU/Hurd pages don't prominently refer to anything but Mach...

Re:More cool stuff in Solaris?

[naasking]

Wikipedia seems to be confused as to whether its moving to "Coyotos" or whether there is only some talk about that [...]

L4, as great a microkernel as it is, still has some serious security shortcomings. The Hurd devs are currently discussing issues with the Coyotos devs; whether Hurd will actually use Coyotos is unclear last I heard. You can see various discussions they've had on the coyotos-dev mailing list.

ZFS, ZONES, RAID-Z, DTrace, etc. (nt)

[Anonymous Coward]

ZFS, ZONES, RAID-Z, DTrace, etc. (no-text)

"Reason: Don't use so many caps. It's like yelling." I fucking hate these bullshit new AC posting rules!

I can't write more than one reply per half-hour, and now using acronyms triggers a "lameness filter"?!!!

GPL DTrace for teh win!

[Anonymous Coward]

Seriously, it'd be beyond awesome to have DTrace ported to Linux.

Re:GPL DTrace for teh win!

[[tsa]]

Hm...

When you can get an open-sourced carrier-grade OS like (Open)Solaris at no cost,
why still Linux?

OpenSolaris surely currently lacks a lot of (x86) hardware support, no drivers
for widespread hardware, etc. - but as more and more users actively use and
support OpenSolaris, more and more vendors will provide those.

What I don't like about Linux - Linux (and a lot of Linux software), that is - is
the neverending story of changing APIs - use something, update something else - Oops.

I have a Linux system here, with at least three different versions of, e.g., BerkeleyDB.
1.85 compat, 3.something, 4.idontknow. API changes, incompatibilities, you name it.

Ever tried to compile popular Linux software on another Un*x? Whenever I encounter some
piece of GPL-licensed software, I can almost guarantee it won't compile on Solaris, Tru64, .. - without spending hours for #ifdef'ing and patching the source.

You want DTrace? Zones? Use Solaris. Is there any technical reason (no politics, please) where
using Linux actually offers any benefit?

(Yes, "smc" and all those java-based admin utilities suck. But commandline-based alternatives
do exist.)

This is not a flamebait. Serious answers will be appreciated.

Re:GPL DTrace for teh win!

[Anonymous Coward]

> Seriously, it'd be beyond awesome to have DTrace ported to Linux.

I don't think you understand DTrace or the enormity of your comment. Solaris was practically rewritten to fit around DTrace, Linux would as well otherwise it would just be merely a gdb+.

Re:GPL DTrace for teh win!

[Anonymous Coward]

a) OpenSolaris is not 100% open - You still need binary blobs to build and run the OS. Not all drivers are open source.
b) While Solaris may be carrier grade and it certainly is rock solid and scalable (although slower than Linux) on Sun hardware, if I had to build an OS for embedded hardware or if I had to use my Desktop/Laptop fully, I would go with Linux - it's been there and done that.
c) If I were to use a OS on x86/64 hardware without having to pay a dime for support, Linux is the best choice - Linux community is very friendly and issues do get sorted out most of the times. Try visiting OpenSolaris community and see how much % of user problems are ever answered / resolved.
d) Linux user land APIs have rarely if ever been changed - kernel api stability is nonsense for a evolving OS.
e) Software availability - much more OSS and commercial software is available and supported on Linux compared to Solaris on x86
f) If I am a Developer wishing to contribute - it's way easier to do so with Linux than OpenSolaris (Try submitting a patch to Opensolaris)

I can go on but the point is OpenSolaris is not the panacea - nor is Linux but it's close and I have hopes.

Re:GPL DTrace for teh win!

[Peter La Casse]

It seems like a GPL Solaris would satisfy the needs of people like you, providing everything that people like about Linux with everything that you like about Solaris.

I have a Linux system here, with at least three different versions of, e.g., BerkeleyDB. 1.85 compat, 3.something, 4.idontknow. API changes, incompatibilities, you name it.

That sounds like a system whose package manager has been bypassed.

Ever tried to compile popular Linux software on another Un*x? Whenever I encounter some piece of GPL-licensed software, I can almost guarantee it won't compile on Solaris, Tru64, .. - without spending hours for #ifdef'ing and patching the source.

That doesn't sound like a Linux problem. Ever try porting a Solaris device driver to Tru64? Same thing. (It's easier than it sounds, if one is willing to spend hours #ifdef'ing and patching the source.) Lots of GPL software is already prepackaged for Solaris; if it's not on the Sun extras CD it's probably on blastwave.org or sunfreeware.com.

You want DTrace? Zones? Use Solaris. Is there any technical reason (no politics, please) where using Linux actually offers any benefit?

The rapid change of the Linux kernel results in new ideas and improvements being tried sooner and shortens the shelf life of cruft. Some aspect of the kernel look gross? All it takes to remove it is something better. Want hotpluggable SATA drivers? You don't have to wait long.

I installed a recent version of Solaris and on x86_64 hardware the minimum memory requirement is 512 MB. Maybe it's not really 512 MB - maybe omitting Gnome etc. would make it much smaller - but all I know is what the installer tells me. Linux is easy to install on machines with less memory (though it was annoying when RHEL started requiring 256 MB).

This might have changed since the last time I tried it, but Solaris x86 wants to be installed to a primary partition, which is sometimes not optimal.

Linux has a greater variety of filesystems available. It sounds like ZFS is great, but undoubtedly there is some niche that it's not optimal for. If XFS had been available for Solaris 5 years ago it would have saved me a ton of work.

I hope this has improved, but one area where Solaris used to really drop the ball was ATA controller support. With Sun selling x86_64 hardware, though, I can't imagine they'd let that situation continue.

Re:GPL DTrace for teh win!

[adrianmonk]

I have a Linux system here, with at least three different versions of, e.g., BerkeleyDB. 1.85 compat, 3.something, 4.idontknow. API changes, incompatibilities, you name it.

That sounds like a system whose package manager has been bypassed.

Actually, I'm fairly sure it's the default on Slackware 10.2!!

Ever tried to compile popular Linux software on another Un*x? Whenever I encounter some piece of GPL-licensed software, I can almost guarantee it won't compile on Solaris, Tru64, .. - without spending hours for #ifdef'ing and patching the source.

That doesn't sound like a Linux problem. Ever try porting a Solaris device driver to Tru64?

He didn't say anything about device drivers. He said "software". As in regular old userspace application or utility. Ever tried to compile xine or something like that for Solaris? There is a lot of software out there written by people who ignorant of (for example) which headers must be #include-ed for which functions and whose software is incorrect but happens to work on Linux. A lot of times these bugs only get fixed when someone tries to compile it on some non-Linux system, it fails, they fix the bugs, and they submit a patch back to the author. Yes, that is open source, distributed development in action, and that's great, but only if the author accepts the patch and doesn't keep making the same mistakes again and again.

Re:GPL DTrace for teh win!

[Kjella]

OpenSolaris surely currently lacks a lot of (x86) hardware support, no drivers
for widespread hardware, etc. - but as more and more users actively use and
support OpenSolaris, more and more vendors will provide those.

Chicken, meet egg. The users aren't there because the support isn't there. Pretty much every computer geek I know has tried out Linux by now on some machine (doesn't mean they all use it regularly because of games and other things. How many have tried OpenSolaris? Close to none. Don't have the hardware.

You want DTrace? Zones?

A: Don't know. Don't know what they are or how they're useful. Can I play around with them on my machine?

Re:GPL DTrace for teh win!

[Ash-Fox]

What I don't like about Linux - Linux (and a lot of Linux software), that is - is
the neverending story of changing APIs - use something, update something else - Oops.

I don't understand how this is specific to Linux. Windows does this with their visual basic runtimes, richedit controls, .net and so on. As does other operating systems (MacOSX, FreeBSD etc).

I have a Linux system here, with at least three different versions of, e.g., BerkeleyDB.

Oh nos! Windows comes with richedit controls 2.0, 3.0, 4.1 and even a version for .net! The Horror!

4.idontknow. API changes, incompatibilities, you name it.

Then stop symlinking or whatever it is you're doing the software to newer libraries when they're written to use older versions, you certainly wouldn't want to run a application that uses riched20.dll with Msftedit.dll (newer version of the same library) instead, causing odd incompatibilities.

Ever tried to compile popular Linux software on another Un*x?

Yes.

Whenever I encounter some
piece of GPL-licensed software, I can almost guarantee it won't compile on Solaris, Tru64, .. - without spending hours for #ifdef'ing and patching the source.

I agree, I also find it too much work for use on Solaris, so I'll use something like OpenBSD, FreeBSD and so on instead which is far easier.

Is there any technical reason (no politics, please) where using Linux actually offers any benefit?

Well, it actually works with most hardware, has modern proprietary games, proprietary desktop software available that isn't available to OpenSolaris... Yeah. Plus, believe it or not, some distributions are actually easier to use than Solaris.

Re:GPL DTrace for teh win!

[Peter La Casse]

Ever tried to compile popular Linux software on another Un*x? Whenever I encounter some piece of GPL-licensed software, I can almost guarantee it won't compile on Solaris, Tru64, .. - without spending hours for #ifdef'ing and patching the source.

That doesn't sound like a Linux problem. Ever try porting a Solaris device driver to Tru64?

He didn't say anything about device drivers. He said "software".

Device drivers are "software". It's natural and expected that software designed, run and tested on only one system will require work to port to another system. Every OS has non-POSIX interfaces that many programs use. Here's a non-device-driver example: ever try porting an application that uses Solaris threads to another platform? It doesn't "just work", but it's doable. There are often basic differences between environments. Linux is the only OS I know of that has a "linux" subdirectory in /usr/include.

Re:GPL DTrace for teh win!

[squiggleslash]

That sounds like a system whose package manager has been bypassed.

In other words, a pretty much standard GNU/Linux distribution that's been used for more than a few months and had to be updated a few times.

Dependency hell. Don't you just love it? ;-)

Re:GPL DTrace for teh win!

[Peter La Casse]

That sounds like a system whose package manager has been bypassed.

In other words, a pretty much standard GNU/Linux distribution that's been used for more than a few months and had to be updated a few times.

Bypassing the package manager would make it a nonstandard system. I can't think of a reason why an ordinary user would need to install three different versions of the same program; a developer might, but then all bets are off.

Dependency hell. Don't you just love it? ;-)

I use a package manager, so I don't get to experience it (even on Solaris; pkg-get is wonderful.)

Re:GPL DTrace for teh win!

[squiggleslash]

Well, I've done it. The usual reason is you're trying to install the latest-greatest of something, but it requires libraries that are not only newer than what you have, but are incompatible with some existing apps. I remember trying to upgrade glibc and finding the package manager wanted to remove half a dozen apps and subsystems because ONE dependency was incompatible with that glibc.

> Dependency hell. Don't you just love it? ;-)

I use a package manager, so I don't get to experience it

Huh? I use a package manager on my Debian installation. I suffer dependency hell. Mac OS X doesn't suffer this problem. I don't use a package manager on it.

Dependency hell is the product of an environment in which everyone thinks that everyone else has a million or so development libraries on their machines, that are always up to date and yet never, ever, change. It sucks. If you're not suffering it under Solaris, I suggest it has to do with saner behaviour with Solaris, which is possibly (I have no idea, last time I used Solaris was version 8, which never installed, so I'll have to go back and take a look at it) better managed than most GNU/Linux installations, not because of your use of a package manager.

Dependency hell is not solved by package managers, indeed usually it's made worse, both by the reliance upon them and the degree to which they make resolving some dependencies doubly hard (by making it more difficult to install two versions of the same application.)

Re:GPL DTrace for teh win!

[Peter La Casse]

Dependency hell. Don't you just love it? ;-)

I use a package manager, so I don't get to experience it

Huh? I use a package manager on my Debian installation. I suffer dependency hell.

That's odd; I've only suffered dependency hell on my Debian or Ubuntu installations when I bypassed the package manager. Were you using Unstable?

Dependency hell is not solved by package managers, indeed usually it's made worse, both by the reliance upon them and the degree to which they make resolving some dependencies doubly hard (by making it more difficult to install two versions of the same application.)

I disagree; the problem is incredibly reduced by package managers, because when you want to install a package, you don't have to track down the programs it depends on, the programs that those programs depend on, etc. The only things that a package manager have trouble with are rare corner cases, such as when you want to have two versions of the same application or when you want to have two packages with conflicting dependencies (where success or failure depends on how shared libraries and versioning are implemented, not on whether or not there's a package manager).

Java, Shmava

[Anonymous Coward]

Solaris, Shmolaris

Wake me when K-Fed open sources the Britney video.

ZFS

[ArbitraryConstant]

If it were released under the GPL, it could be imported into the Linux kernel.

I think this is awesome.

I can't decide whether Sun has balls of spent Uranium or if they're just really disparate. Possibly both. But I really like this, and I hope their services and hardware businesses benefit accordingly.

Re:ZFS

[EvilRyry]

ZFS has some really awesome features. Pooling, snapshots (no, not quite like LVM), RAID-Z, and native compression and soon encryption.

I'd love to see all this in Linux but I'm thinking even if it were GPLed there would be a lot of work to do to port it. And of course after its ported, the Linux devs would probably make a big stink about accepting it using lines like "a file system should only put files on a block device!" ZFS however is a different approach to storing files and in many ways much better.

Re:ZFS

[drsmithy]

I can't decide whether Sun has balls of spent Uranium or if they're just really disparate. Possibly both. But I really like this, and I hope their services and hardware businesses benefit accordingly.

Sun makes the vast bulk of their money from hardware sales and support. They have little (if anything) to lose from GPLing Solaris.

Contrast this to, say, Microsoft, who makes most of their money from software sales. Clearly, GPLing their software would be financial suicide.

Re:ZFS

[volsung]

Hopefully this guy will finish his port of ZFS to FUSE on Linux someday, in which case a lot of the work will have been done. You will still have to do some cleanup to make it run again in kernel space, and port it to the Linux VFS layer, of course. His choice of FUSE is in part due to the license, I imagine. A kernel port of the ZFS code could never go into the Linux kernel due to the license issues between the CDDL and GPL, where as this is perfectly fine in userspace.

That said, it's too bad the FUSE port has stalled (no commits for 5 weeks now). I want ZFS bad, but trying to admin a Solaris box feels like having my hands chopped off. Nothing is where I expect it to be.

Re:ZFS

[mike3k]

Yes! I'd love to see ZFS in Linux (and Mac OS X).

Re:ZFS

[Anonymous Coward]

As a primarily Solaris admin, I feel the same way when I log into a Linux box: nothing is where I expect it to be. Made worse by having to deal with multiple distributions who do things differently. :(

Re:ZFS

[volsung]

Yeah, I've been playing with Nexenta and was pretty impressed by the layout (and ZFS of course), but had a rough time figuring out what hardware was detected, how drivers are loaded, and so on.

As for the Linux distros, I had to start thinking about them as branches in a family tree, rather than as one OS. There is the Debian lineage, the RedHat/Fedora lineage, the Gentoo lineage, ....

Re:ZFS

[swordgeek]

"Nothing is where I expect it to be."

This is just life, going between OSes. I spend 95+% of my time on Solaris. In the last two weeks, I've had to work on OSF/1, HPUX 11i, and HPUX 9.04. You get used to switching between variants, and start to get a feel for where things are. ("Oh, that's in /sbin instead of /etc, so that means that THIS should be in...")

Trust me, it's good for the soul. And ZFS rocks. (Not to mention zones, dtrace, and now brandz).

Re:ZFS

[ArbitraryConstant]

"A kernel port of the ZFS code could never go into the Linux kernel due to the license issues between the CDDL and GPL, where as this is perfectly fine in userspace."

That's why it would be cool if Sun released Solaris under the GPL.

Re:ZFS

[bhima]

Man, I am *NOT* using a filesystem for long period of time through FUSE.

I can see using FUSE on a temporary basis but day in day out?

Re:ZFS

[advid.net]

That's what I'm waiting after since it came out:
ZFS in main Linux tree as kernel module.

BTW, any idea about how robust is ZFS compared to ext2/3 when some blocks are corrupted (for instance reiserfs loose a lot more here) ?

Re:ZFS

[Dan Ost]

I've been using sshfs (which uses fuse) for some time now and haven't noticed any problem at all.
Perhaps it's because I'm using it over a relatively fast network.

Re:ZFS

[volsung]

The only complaint I can imagine is performance. Passing VFS calls in and out of userspace is slow.

Re:ZFS

[volsung]

Yeah, I've been using ZFS via Nexenta in VMware, and aside from how much physical disk support blows in VMware Server, it's been awesome. The hard part has not been userspace, which Nexenta smooths over quite a bit, but low-level stuff that I took for granted:

• How do I tell what drivers are installed? Did it detect my virtual network hardware correctly?
• I just added a new disk, what device name did it receive?

And so on. All stuff with obvious answers to the Solaris experts, but it took a while with google to get oriented. (My reflex to use dmseg is not helpful on Solaris.)

Re:ZFS

[bhima]

What are you using for though? Is this something that's in place 100% of the time in a production environment? or is this something that you use frequently but for short durations as a matter of convenience?

My idea is that FUSE is better suited for the latter than for the former... but I could be wrong.

Re:ZFS

[Dan Ost]

I guess it's something that I use frequently, but for short durations.

I pretty much use sshfs everyplace that I would have used NFS a couple of years ago.
The only time I even notice it is when I'm moving large files to or from the local file system.

Re:ZFS

[htd2]

If OpenSolaris is released under GPL then why would you bother with Linux at all. Yes I know Linux has more device drivers than OpenSolaris but this gap is being closed rapidly. And yes Linux User land is nicer for the less skilled to use than Solaris, but again this gap is also being closed.

The difficult challenge is not building a nice UI on top of your kernel of choice, instead it is building a sensibly architected and well implemented kernel. Despite Linus's anti Solaris gibes no one who has much of a clue about the relative capabilities of the Solaris and Linux kernels seriously thinks that Linux gets close to Solaris in terms of reliability, scalability and clever/useful features such as dtrace etc.

The reality is that Linux only has four things going for it, better UI, more device drivers, arguably GPL and better marketing. If OpenSolaris goes GPL and the other two techical gaps are being closed then why would you want ZFS under Linux when you can have a much better kernel with ZFS and without the effort?

Re:ZFS

[scott_karana]

If you're going to use OpenSolaris, the license doesn't matter. The CDDL is perfectly legitimate for commercial use. I mentioned Linux because the change to GPL is really only significant for linking OpenSolaris code with GPL code, Linux applications being the most common.

Re:ZFS

[htd2]

GPL applications can run perfectly well on OpenSolaris and there is no conflict between the Solaris kernel which is CDDL and a whole host of GPL licensed apps. What exercised the Linux bigots when OpenSolaris was released was the fact that being CDDL based you could not mix GPL and CDDL kernel code.

It was always rather questionable what benefit would have been derived from doing this since copying large chunks of Solaris kernel code into Linux wasn't very practical given the differences between the two kernels. The noise was all from the Linux side anyway since it is difficult to see what you would want to copy from the Linux kernel into Solaris.

In that sense the adoption of GPLv2 or GPLv3 for OpenSolaris would be more of a PR effort than anything else though it would have the beneficial effect of silencing the oppose Sun forever bigots once and for all.

Which *version* of the GPL

[YA_Python_dev]

More interesting than this, IMHO, is to note that for Java they choose to use the "GNU GPL v2 only" (plus Classpath exception) license, not the more common "GPL v2 or any later version".

This is what the Java FAQ says about it:

Q: What about GPL v3? Have you considered using that license?
A: While Sun has been working with the Free Software Foundation as an active participant in the development and review of the GPL v3 license, this license is not yet complete. It is Sun's strong desire to complete the open sourcing of its Java technology implementations in a timely manner, so we made the decision to use an existing, established license paradigm rather than wait for GPL v3 to be completed. Using GPL v2 does not indicate anything negative about GPL v3. Sun continues to be very actively and positively involved in this new license's development.

And, from this InfoQ article about the GPLed Java:

GPLv3 was not chosen since it is not finished yet, but when asked if Sun will move to GPLv3 an official said "at this point we don't know what the final license will be."

Re:Which *version* of the GPL

[wikinerd]

This GPLv2-only licensing may create some practical problems in the future, but it is sensible from a business point of view, and I can certainly understand it. It's better to have their code in GPLv2 rather than not have it at all. We were given a gift, so let's not whine for a while.

Re:Which *version* of the GPL

[YA_Python_dev]

We were given a gift, so let's not whine for a while.

You're right, of course, and I'm not whining at all, simply noting the details of the license choice. And I just found this bit from Sun's Jonathan Schwartz blog:

And yes, we picked GPL version 2 - version 3 isn't available, but we like where the FSF is headed.

Emphasis mine. So apparently Sun likes the proposed GNU GPLv3!

Re:Which *version* of the GPL

[ClamIAm]

I don't think this is really that interesting. Sun, being a big company, must run things past the lawyers. GPL v3 is not finished, and they are probably extra-extra-concerned about the patent revocation clauses and how exactly that stuff will be worded. So basically they're just covering their asses.

Re:Which *version* of the GPL

[evilviper]

they choose to use the "GNU GPL v2 only" (plus Classpath exception) license, not the more common "GPL v2 or any later version".

Actually, omitting the "or any later version" is VERY common. I certainly don't have any statistics, but I know for a fact that HUNDREDS of major/popular GPL'd programs omit that clause, including the kernel itself.

It's not at all unusual. I guess developers don't have much blind faith in FSF/RMS.

Re:Which *version* of the GPL

[Cyclops]

I doubt very much Sun will use Free Software that's not copyrighted by them on Sun's Java, even if it's GNU GPL'ed. Expect to have to sign copyright attribution agreements when submitting code.

That's not bad per-se, but it's much better if acompanied by a promise to never release as non-Free Software (like the FSF does).

I'm pretty sure that if Sun wants to change to GPLv3 in the future they'll only need to shout the order.

Re:Which *version* of the GPL

[Anonymous Coward]

It's not that interesting. The "or later" clause is there primarily to make license fixups easy for small projects which choose not to do copyright assignment... not, as the usual collection of fuckwits claims, to allow RMS to do a bait and switch.

Sun owns the copyright, and will do copyright assignment for any contributions. So it's not a problem.

Re:Which *version* of the GPL

[Kjella]

It's not really that big a deal when a company decides to use GPLv2 only for a big product. That basicly means "GPLv3 isn't out yet, we'll consider it when it is." They have the power to relicense Java to GPLv3 with the stroke of a pen. I guess it still remains to be seen how they'll allow contributions, but it needn't be a problem in the future either.

That is a sane position for a company. The trouble comes when you need to relicense hundreds if not thousands of small contributions, many of which you can't get hold of (left the project, moved, deceased or otherwise). Not to mention they all have to agree, but that's another issue. In that sense, "GPLv2 or later" is an in advance grant of license based on good faith in the FSF.

But if you're quite certain you'll be around to evaluate it, why license it now? They can just as easily do it when it's available for legal review. If you find an old source file from this GPLv2'd Java, I think you'll find it quite easy to find a contact address for Sun, Inc. and it'd probably be well known if they "retroactively" licensed old Java versions to be GPLv3 as well.

Seriously, if I was a lawyer and someone wanted to release something under a license that wasn't written yet, I'd balk too. Licensing it under the GPLv3 currently requires good faith. If you look at it with a lawyer's eyes, which usually translated to "How screwed could we get?", clearly it's better to wait and see.

Re:Which *version* of the GPL

[squiggleslash]

They've effectively announced that already. As I understand it, if you want to be a part of the official Java effort, submitting patches to Sun, you have to sign a statement giving them joint ownership over the code you submit.

This seems reasonable to me though obviously strong free-software advocates should understand that this gives Sun the right to license your submissions for use in proprietary code. If you don't want to do this, which might be because of that, or because you're anti-GPL3 (which looks likely to be something Sun will switch to), you can always fork the code and release the changes that way, rather than using Sun's infrastructure and sending the changes via that.

Re:Which *version* of the GPL

[wikinerd]

From business side it is understandable that they ask for joint ownership. From our free community perspective, this means that we should be always ready to fork in case we need to. As long as we can fork, I have no problem with joint ownership: If I write a small patch I will give it to them, if my patch is significant then I will fork as a separate project first to prepare the ground for future competition if the need arises, then apply my patch to my fork under the GPL, and optionally also submit it to Sun if they are ethical. This also aplies to other projects with joint ownership such as OpenOffice.org. Note that having our own 'community-built java fork' operational at any time is good from a competition perspective because we can use it to force Sun remain ethical: 'Look, business guys, we will apply our patches to our own project first, and only give the patches to you as long as you follow the spirit of GPL, but the moment you lose your ethics, we will advertise our project as a free alternative and we will force you to compete in a free market where only the highest quality products with the lowest price win the users'.

mod 04

[Anonymous Coward]

Java IRC client volume of gNetBSD leavi8g the play and that the floor though I have never committerbase and that should be

BSD License

[DaMattster]

Why not consider the BSD license? I am not a fan of the GPL and I think it is a bit restrictive. The BSD license appeals to both the F/LOSS and for profit companies. I am surprised Johnathan Schwartz failed to consider the BSD license.

Re:BSD License

[Knuckles]

BSD licence? Hello? I don't think Sun prefers a license where everyone (MS) can copy stuff from Solaris into their proprietary products without giving anything back. BSD license may have its place, but this is not it.

Re:BSD License

[evilviper]

I don't think Sun prefers a license where everyone (MS) can copy stuff from Solaris into their proprietary products without giving anything back.

Right. They sure hated it when everyone copied NFS and didn't give anything back...

Re:BSD License

[Knuckles]

A network file system != your OS crown jewels.

As I said, there are places for using BSD. One of those is if you have a new technology which profits from the network effect, and you want to establish it as a standard. Network protocols or network file systems are prime candidates for this. Solaris is not.

Thank you, Sun

[wikinerd]

I support Sun GPLing Java and possible OpenSolaris. If I ever buy a rack server, I will make sure it's a Sun model.

Re:Thank you, Sun

[Anonymous Coward]

They also have some sweet desktop "workstations" with dual Opterons. Just got one for work for about 3K with lots of ram and disk space. It is built like a tank! My next home machine will definitely be a Sun Workstation.

Re:Thank you, Sun

[hritcu]

I think this would be the best way to support SUN and AMD, and not (semi-)monopolies like Microsoft and Intel.

Sunstroke

[Anonymous Coward]

Remember SUN Java desktop (cool) then they give FUD money to SCO and release Solaris under an obnoxious GPL-incompatible license. Are Sun totally schitzoid or is the business model playing Microsoft and the FOSS community off against each other? The Sun is shining at the moment, I just worry that tommorow they'll pull a Novell.

All said and done, GPL'd java is cool, I'm building Jikes, JamVM and ClassPath now to start (re-)learning Java while I wait for hotspot and javac to hit Gentoo's portage tree.

Re:Sunstroke

[squiggleslash]

Sun didn't give money to SCO for nothing. They bought rights to SCO Unix, so that they could incorporate the drivers into Solaris. At the time, SCO Unix had relatively good Intel/PC support, and Solaris, quite honestly, wouldn't run on 75% of the PCs out there (I know this first hand. I have a Solaris 8 media set. I couldn't get it to install on a Thinkpad or a cheap VIA based box. Even Darwin ran better on both.)

Part of the problem was how they worded it. They pointed out at the time they'd bought the rights in such a way that Solaris users would never have to worry about being sued by SCO. This made it sound like they'd paid a blackmail charge, rather than actually buying the right to redistribute actual (as opposed to imagined) SCO code.

They're now talking openly about moving to the GPL away from the CDDL. I can't help but feel that's good news.

Sun may have taken MS $$$ to not GPL Solaris

[Khopesh]

Wow, I nailed that on the head just this morning (about half an hour after the blog entry that I didn't know about); to me, this seemed obvious.

I can't help but think that the MS-Novell deal was the inspiration for this going into completion; the final straw, so to speak (or at least, the reason for the timing of the announcement). Consider it; Novell uses Mono and just got in bed with Microsoft.

Sun is in trouble, and according to FSF Lawyer Eben Moglen's (wild) allegations in his talk at a recent Free Software Foundation Associate Membership meeting, they previously (2005?) took a bribe from Microsoft to keep OpenSolaris incompatible with the GPL (in exchange for financing they believed was desperately needed for miniaturizing CPU size with Fujitsu to compete with IBM(?) in the server market). Sun is now flip-flopping like a struggling politician; they caved to the pressure of GPL'ing Java despite (allegedly) accepting a bribe to keep Solaris less free. Like many on Slashdot, I consider the FSF and the F/OSS development community greater long-term allies than Microsoft, so maybe Sun will release Solaris 11 or 12 under the GPL.

The Free Software Foundation has made no announcements on either of these developments. What does this mean for the GCC/Java code, which is largely functional? How would GPL'ed Solaris utilities impact use and development of the GNU utilities? (Yes, I realize that the Solaris utilities share code with BSD utilities given their common ancestors, but Solaris has the shiny stamp-of-approval from major security auditors.)

Re:Sun may have taken MS $$$ to not GPL Solaris

[Knuckles]

The Free Software Foundation has made no announcements on either of these developments.

The FSF statement will come on Monday in the official Sun press conference.

Re:Sun may have taken MS $$$ to not GPL Solaris

[adrianmonk]

Sun is in trouble, and according to FSF Lawyer Eben Moglen's (wild) allegations in his talk at a recent Free Software Foundation Associate Members [fsf.org]hip meeting, they previously (2005?) took a bribe from Microsoft to keep OpenSolaris incompatible with the GPL

Wow, I would really like to see some evidence of that. As it stands, it's just an absurd claim with no support. Having (over a period of 15+ years) used Sun equipment and software, and having worked with the company as a customer, and having known people who were employees at Sun, I would say that the chances of this being true are about 0.01%. Sun doesn't like being told what to do by Microsoft, or have you forgotten that Sun once famously sued Microsoft over Java and also once banned PowerPoint presentations? In fact, here's a Scott McNealy quote about the PowerPoint thing:

We had 12.9 gigabytes of (Microsoft) PowerPoint slides on our network. And I thought, 'What a huge waste of corporate productivity.' So we banned it. And we've had three unbelievable record-breaking fiscal quarters since we banned PowerPoint. Now, I would argue that every company in the world, if they would just ban PowerPoint, would see their earnings skyrocket. Employees would stand around going, 'What do I do? Guess I've got to go to work.'

So, it's really hard to believe that a company with this history would be in bed with Microsoft and on some kind of anti-GPL crusade. Anyway, you also said:

Sun is now flip-flopping like a struggling politician; they caved to the pressure of GPL'ing Java

I hardly see how this can be considered flip-flopping. Two years, neither Solaris nor Java was open source in any sense. A year ago (approximately), Solaris was open-sourced. Today, Java was open-sourced, and they mentioned they are thinking of adding GPL to the list of licenses for Solaris. This seems like a steady trend in the direction of open source.

The Free Software Foundation has made no announcements on either of these developments.

So what? Dovecot is a really cool GPL-ed POP/IMAP server, and I don't recall the FSF making any announcement when it was released. Nor do I recall the FSF making any announcement when many other things were released under the GPL. It's not necessary because the GPL speaks for itself.

How would GPL'ed Solaris utilities impact use and development of the GNU utilities? (Yes, I realize that the Solaris utilities share code with BSD utilities given their common ancestors . . .

What common ancestors? Solaris is based on System V Unix from AT&T, not on BSD. Yes, Solaris 1.x (a/k/a SunOS 4.x) was based on BSD, but it was pretty much totally rewritten before Solaris 2.x, so Solaris 2.x (and 7, 8, 9, 10, etc.) have very little in common with Solaris 1.x, and thus very little in common with BSD. You will notice that /usr/bin/ps on Solaris takes options like "-ef" rather than options like "aux", and you will also notice that sh's echo command needs "\c" in the string instead of "-n" as a separate argument if you want to supress the newline. So you can see that the Solaris command-line utilities are not very BSD-like.

Re:Sun may have taken MS $$$ to not GPL Solaris

[Khopesh]

Wow, I would really like to see some evidence of that. As it stands, it's just an absurd claim with no support.

I put the "wild" preface there for a reason. The FSF is a great organization, but sometimes they are a bit nutty. Eben has some heavy insight into things from a legal and IBM perspective, and is tied directly to important figures in these matters. He likely heard a rumor or two and pieced it together as something far larger than it was. I have no sources other than what I heard at that meeting.

The Free Software Foundation has made no announcements on either of these developments.

So what? Dovecot is a really cool GPL-ed POP/IMAP server, and I don't recall the FSF making any announcement when it was released. Nor do I recall the FSF making any announcement when many other things were released under the GPL. It's not necessary because the GPL speaks for itself.

The FSF has been listing a Free version of Java as a top priority for the past few years, stating that it is the largest hole in the Free Software community. See http://www.fsf.org/campaigns/priority.html. The FSF shouldn't announce the GPL'ing of Java, but it should respond to the announcements by commending Sun, removing GCC/Java from it's priorities list, and/or getting Stallman to amend his Java rant.

Solaris is based on System V Unix from AT&T, not on BSD. Yes, Solaris 1.x (a/k/a SunOS 4.x) was based on BSD, but it was pretty much totally rewritten before Solaris 2.x

My bad, I got it backwards (which is odd, since I'm a Solaris admin). That was added as an afterthought.

Re:Sun may have taken MS $$$ to not GPL Solaris

[Khopesh]

I composed an email to the FSF on this topic. Here is their response:

We appreciate your concern and the fact that you took the time to write us. To address the issues you've raised:

• There was an announcement about Java posted within two days of Sun's announcement.
• We will be adjusting our priority project list as soon as details and for Sun's release are clear. We always work hard to publicize our work, especially since we are actually involved in all of these issues.

While they didn't link to the announcement, it was easy enough to find on the main page: FSF welcomes Sun's GPL release of Java (posted 11/15).

It had to be said...

[sethstorm]

If they're this optimistic about such, why not pull some build out that worked with sun4m and make it just as open as what exists today - even if it amounts only to being a olive branch to systems long since devalued by that move. That, and it gives a very compact/cheap option for SPARC that doesnt skimp on the hardware (unlike U5/U10's cheapened design).

Open Source ist not about...

[my ID is not availab]

...changing/hacking the source but about just and only having the POSSIBILITY to view and hack/change the source. Open Source is a principle - not a method to $what-ever.

What about checking facts for a change?

[orzetto]

GPL'ing a product has NEVER been successful for the company or person owning it.

Surely, you may want to talk to the CEOs of MySQL and Qt developers Trolltech, who release their projects under the GPL and do turn a profit. In the case of Sun, as others already have mentioned, they make money on the hardware, and commoditising software is only good for them.

Of course, these are corporations. Speaking of private persons, what about a certain Linus Torvalds, who is now fairly well-off?

Yeah Right

[Tharkban]

Yeah, Right....I bet they'll GPL java before they allow that to happen. :)

I Just Have One Question?

[Einstein_101]

What about hardware support?
 
Not to start an argument, but wasn't there an article posted on this very website telling us how OpenSolaris was/is the Linux killer*? So far how has that one panned out? Other F/OSS operating systems never really make it to relevancy because frankly, their hardware support is always years behind that of Linux. That very reason is why I switched (at least for now) from BSD, back to Linux.
 
*I must note that it funny that they compared OpenSolaris to SUSE - clearly the slowest of the major Linux distros.

Free DOES have a price

[Sloppy]

You can't lock out competition.

Woot? 7p

[Anonymous Coward]

SLING you can comprehensiv3 Serves to reinforce

Spreading?

[Anonymous Coward]

Sun today.

Tomorrow.... The world!

Bwahahahahaha

ZFS

[scott_karana]

The best part about Sun being GPLed (the CDDL is a fine license itself) is that ZFS can be implemented as a kernel module rather than in FUSE. The idea of running non-trivial enterprise filesystem in user-space is abhorent to me.

*cough*TROLLTECH*couch*

[HishamMuhammad]

Enough said.

Taking market away from IBM with GPLv3

[Freed]

Although as a business Sun would only make statements about existing licenses, i.e., GPLv2, Schwartz and others already are on record as liking how the GPLv3 is shaping up; the big jump was to GPL v.anything.

Sun should eventually put OpenSolaris under GPLv3 when it comes out simply because it fixes the GPL2 in important ways and offers important advantages such as greater compatibility with other licenses. But additional kinds of motivation can be given.

The following explanation involves the kind of symbolic considerations that partly characterize the Java move to the GPL. It too may help Sun gain against IBM, a heavy proponent of the Linux kernel.

A possible way for Sun and OpenSolaris to gain even more mindshare is by exploiting a division in the FOSS community. Some top Linux kernel developers decided against GPLv3 very vehemently while decrying a split in the wider developer community of v2 vs. v3. A major factor in the decision is to allow for enforcement of DRM, a factor which may not constrain OpenSolaris's direction so much.

The important (unexaggerated) divide involves kernel developers. OpenSolaris has more to gain than to lose by going GPLv3. To users and developers in general, Sun comes off as an even bolder, more progressive leader, just as with the Java announcement Monday.

The FOSS community dwarfs in number the few dozen kernel developers described above and it's safe to say the most of the community were caught off guard by the announcement from the developers; the expectation was that there would be an eventual move of the kernel to GPLv3. Sun could use the same kind of exquisite timing of the Java announcement on the heels of the Novell/MS flub and announce OpenSolaris under the GPLv3, satisfying the unfulfilled expectation of having a kernel under the GPLv3. IBM would be caught off guard again.

The Java move was disruptive, and this move is disruptive in the same way with little downside. It's not for the sake of disruption but disruptive against the likes of IBM.

Re:Taking market away from IBM with GPLv3

[Ash-Fox]

OpenSolaris has more to gain than to lose by going GPLv3.

One of their goals in the past was making sure Linux doesn't benefit from Solaris. I'd say GPLv3 would help them with that goal a bit.

Re:Taking market away from IBM with GPLv3

[Freed]

Thanks for the reminder--forgot to mention it.

simple...

[sedyn]

That it isn't being called GNU/Java!

who would hack the solaris code?

[Anonymous Coward]

Sun has to do something to compete with linux on x86. Gpl'n their java would be sweet.

My, oh my, oh my. OpenSolaris displacing Linux?

[milliwatt]

Well well well. This is a rather interesting kettle of fish now isn't it?

With it looking like Linus and company are going to be averse to GPL3 (to wit, they very likely couldn't go GPL3 even if they *wanted* to), what if -- just "what if", mind you -- OpenSolaris began plans to go GPLv3? Hm. I wonder... GNU/Solaris?

That would be something, now wouldn't it? I wonder if you'd see the FSF (and a fair portion of the rest of the free software community) jump on OpenSolaris an the preferred free software kernel for GNU? (Ahem -- until HURD is ready, of course).

Very interesting prospects indeed.

GNU software

[Anonymous Coward]

Try downloading, burning, and running Belenix or Nexenta. Solaris kernel plus some Solaris program support, but GNU userland from the ground up! It could be said that without GNU, Solaris distros would have no usable desktop system.

not open sourcing the Java programming language

[giafly]

Q: Are you open sourcing the Java language or the Java SE platform specifications?
A: We are not open sourcing the Java programming language, nor the platform APIs and specifications, which are governed by the JCP. We're open sourcing Sun's implementations of the Java SE and Java ME specifications. - FAQ

Sorry if this has already been posted, but it's important to note that Sun seem to be GPLing Java implementations which already face Open Source competition, but not Java itself. You might not realise this from most news reports.

On this basis, whether they GPL Solaris will depend on how strong they think the threat from Linux is. Does Sun have any other realistic choice?

Re:not open sourcing the Java programming language

[RPoet]

And how else would you open source a programming language? You're saying that they haven't open sourced the language specifications, but these are not composed out of source code, and there is thus no source to be opened. Do you want the specification process to become more open? In which manners?

gunix

[James McGuigan]

A merging of the codebases would create:

gunix = GNU Unix = GNU's I can't believe its not unix.

You must be GNU here

[pebs]

GNU = GNU's Not Unix

GNU = Gnu's Now UNIX

[azrul]

if I may suggest ...

Isn't there a carrier-grade Debian?

[Anonymous Coward]

I thought I saw that a while back...

Solaris evolves too slow

[jhus]

I don't think Solaris has a bright future. It's evolving too slow, too many restrictions, too many problems. Just look at their attempt to update their /bin/ksh from ksh88 to ksh93 - they're working on it since a year and they are still not finished. The ksh93 integration project takes an eternity or two while Linux does the same job in two weeks or a month. This shows the great failure of the Opensolaris project. They're too slow, often twelve times slower than Linux. Sun still has superior features, but at the current evolution rate it is just a matter of time when Linux becomes superior compared to Solaris and their management is deaf and blind and ignores the problem