Andy Tucker (who OSNews interviewed almost a year ago to the day) has posted some thoughts on his reasons for wanting to open-source Solaris. Speaking for myself, I share many of Andy’s reasons and I suspect many of our other engineers do as well. As Andy requests in his blog entry, we want to hear from the community specific
requirements and ideas (and not just about the license).
maybe they should have FULLY open sourced solaris and java 6 or 7 years ago. now they are too late.
what makes sun think people are going to stop developing for linux and switch to solaris ?
Because its a superior OS?
If SUN wants to be succesful with this, they will have to choose the licensing terms very carefully.
The license should not be to strict, so as to make GNU/Linux under the GPL not seem as a better deal.
It shouldn’t be to free either (especially in light of GPL compatibility). SUN has not built its reputation as FLOSS Solaris project leader and any missteps in this fase could lead to gutting out the juicy bits and leaving the rest to rot.
It will be interesting to see how open sourcing part of Solaris will pan out.
What’s amazing here is that so many people beat up on Sun for not open-sourcing Solaris, but seemingly no one gripes that IBM is pushing AIX, DB2 and WebSphere, none of which are open-source….
maybe that is because IBM does not promote AIX as an operating system that everyone should be using and that all developers should write for.
sun on the other hand would love all users to use solaris and all developers to write for solaris, and better yet for them all to pay for the privilige.
as they now know, it did not work out like this, so if they open source solaris then they could get developers to write for the platform and hopefully build up a massive user base that they can sell stuff too in the future.
solaris 10 is around the corner and it will be very doubtful if this is released as open source. so even if sun manage to get loads of developers to write for solaris, it will probably all be done on solaris 9.
this should not however be much of a problem for backwards compatability, but how will all the “new” solaris users take up a deal like this, get one older version for free, but pay for updates ? dunno, might work.
Huh?!
We have lots and lots of pSeries hardware here (not to mention a few mainframes), and trust me, IBM would love everyone to be running AIX. They also promote Linux, but they’re still pushing AIX very hard; it just doesn’t get talked about much. And AIX isn’t cheap, either. At least educational and non-profit users can download a free Solaris license from Sun – http://wwws.sun.com/software/solaris/binaries/get.html
Further, what’s wrong with wanting developers to support your platform? Of course Sun wants lots of developers to support Solaris. Apple wants developers on OS X, too. And many, many developers already write software for Solaris. Additionally, there are two projects that port open-source software to Solais – SunFreeware (www.sunfreeware.com) and Blastwave (www.blastwave.org).
I really have to question the logic that “if [Sun] [were to] open source solaris then they could get developers to write for the platform and hopefully build up a massive user base.” First, Sun already has a massive user base – it is one of the most popular Unix flavors. Additionally, it’s not as though open-sourcing software is a panacea. Apple open-sourced a significant portion of OS X (Darwin), but this really hasn’t helped increase their userbase. Instead, I see a lot of Linux zealots whinning that they didn’t GPL the entire codebase. Likewise, the FreeBSD community is a fraction of the size of Linux. The same goes for PostgreSQL -vs- MySQL. So just having code available isn’t guaranteed to bring you new users.
I’m sure IBM would love it if everyone droped everything and started to use AIX. They of course realize that it isn’t going to happen.
If you listen to the more vocal supporters of linux, they are somewhat extreme in their pronouncements, like this dead, that is dead, etc.
linux has been a low-end disruption to a marketplace that has ignored low-end customers… as it moves up market in the enterprise it too will face the same problems and can fall victim to a low-end disruption.
“”Apple open-sourced a significant portion of OS X (Darwin), but this really hasn’t helped increase their userbase”
actually it has ( see the fink project) and apple is really into the hardware business and their hardware is expensive. so poor analogy
“Likewise, the FreeBSD community is a fraction of the size of Linux. The same goes for PostgreSQL -vs- MySQL. So just having code available isn’t guaranteed to bring you new users.”
both are open source so i kinda find that point superflous. besides the amount of code written for linux is higher than freebsd. see the amount of work in gnome and kde or anything else written and sponspored for linux rather than freebsd. not that freebsd is any less good or something like that but the team size and hype does matter.
”
I really have to question the logic that “if [Sun] [were to] open source solaris then they could get developers to write for the platform and hopefully build up a massive user base.” First, Sun already has a massive user base – it is one of the most popular Unix flavors”
sun believes that getting it open would be an advantage and they are right.
solaris 10 is around the corner and it will be very doubtful if this is released as open source. so even if sun manage to get loads of developers to write for solaris, it will probably all be done on solaris 9.
I don’t know how you came by this impression, but it’s not correct; we want to open it all up, Solaris 10 technology included. (Indeed, we think Solaris 10 technologies like Zones, DTrace and ZFS are exactly what make Solaris compelling to, say, an embedded appliance maker.)
Hi
I believe that opening up java would be have a even bigger impact that solaris. isnt solaris based on sys v. how could you really open it up. is it going to be gpl compatible.
whats the current situation with java. any discussions going on?
just curious
The fink project hasn’t really increased the number of people using Macs. Apple’s market share is going down (even though they make some really nice products — I’m typing this in Safari on a new 15″ pbook). Because OS X is Unix, it’s pretty easy to port Unix apps to. That’s all fink is doing. It’s the same thing that SunFreeware and Blastwave have been doing for years on Solaris.
I don’t understand why you say that “the amount of code written for Linux is higher than FreeBSD.” GNOME and KDE aren’t Linux-specific. Both run on Linux, FreeBSD, OS X and Solaris, among others. And while I don’t think that the FreeBSD -vs- Linux or PostgreSQL -vs- MySQL comparisions were superflous, how about considering projects like SapDB or Firebird? Both were closed codebases which were open sourced years ago. SapDB isn’t very popular, despite much recent hype from MySQL AB, and Firebird still has a very small userbase (it’s developers spent a lot of time converting it from C to C++, and I still can’t find docs for it newer than 1999). It’s likely that Ingres (also recently open sourced) will fare about as well.
As another example, consider OpenAFS: IBM open sourced AFS 3.6 in 2001. OpenAFS development has been seriously lacking – there are some more people using it, but it’s still a niche product, and it isn’t that much better than it was 3 years ago. So merely open sourcing closed, commercially developed code isn’t a cure-all.
I think having a more open development process would be beneficial for Sun, but I don’t think that GPL’ing Solaris would be a panacea. Mostly, I think it would just shut up a lot of clueless Slashdot posters who don’t understand much about IT. I’m interested in the tools that Solaris provides – I really don’t need the source.
“I think having a more open development process would be beneficial for Sun, but I don’t think that GPL’ing Solaris would be a panacea. ”
i dont think anybody sensible would claim that. kde and gnome is sponspored and developed mainly by linux developers and enthusiats. of course there is a very active freebsd base and it can and is being used by other operating systems too.
openafs isnt popular due to various reasons. one of them being that it was developed entirely outside of the community and lacked the vfs calls and linux already has ample filesystems to consider using afs.
“Apple’s market share is going down (even though they make some really nice products — I’m typing this in Safari on a new 15″ pbook)”
i dont agree with this. apple knows what its doing and is happy to be a player in niche market. despite what many people think mac on intel would have a tough time with all the diverse pc hardware and stuff. its much better on apple hardware.
regarding firefox and ingres. what does it do better is not clear. having good doc is very important and promoting it is needed
What’s your evidence that GNOME and KDE “are sponsored and developed mainly by linux developers and enthusiats[sic]” ? Sun is shipping GNOME as the default desktop in Solaris 10, btw. Either way, this isn’t related to the point I was trying to make.
Regarding OpenAFS, it really competes in its own space. It’s only real competitor is DCE DFS, which is closed-source and never really ran well on Linux. There isn’t anything else available that competes with it; NFSv4 tries, but it is still several years behind where AFS was a decade ago.
I’m not sure why you brought up porting OS X to x86 hardware. I never mentioned that. I’m a Mac user, and I really like OS X, but the market share numbers don’t lie – Apple’s market percentages have been going down. That’s not a slight against Apple – I think they’re a very good company, and they’re certainly profitable. But they are a niche player. I’m not so sure that Apple’s hybrid open/closed-source development of OS X attracted users. Personally, I think OS X’s success has more to do with it being an excellent laptop/desktop Unix – its power management and hot-plugging abilities can’t be beat, and its GUI is incredibly slick. Again, it’s the features, not the source code licensing, that’s attracting users, IMHO.
Open-sourcing Solaris would get Sun a lot of good PR, which they need pretty badly right now, but I’m not sure what concrete benefits it would bring. There just aren’t too many examples in recent years of a company opening a closed codebase and having wide success with it (see my above posts). Certainly open source is nice — I’m not against open source development — but I don’t think that it’s essential for Sun. If they want to open source Solaris, great, more power to them. But I can’t say how sick I am of reading “Solaris sucks becuase it isn’t GPL” or “open-sourcing Solaris will fix all of Sun’s problems” in these forums.
“If they want to open source Solaris, great, more power to them. But I can’t say how sick I am of reading “Solaris sucks becuase it isn’t GPL” or “open-sourcing Solaris will fix all of Sun’s problems” in these forums.”
software will become popular for various reasons. not just being open source. open source isnt a silver bullet. i however am doubtful whether sun can really do this considering the litigous sco and sysv base solaris is under.
To attempt to answer some of Andy Tucker’s questions:
Are there any examples that you think work particularly well, or not so well?
This is a good question to ask on the appropriate OpenOffice.org lists. Sun came in for a bit of criticism at one stage for the way it had set something up – I can’t remember quite what it was – but it was sorted out eventually. OpenOffice.org has worked quite well, as it happens.
Mozilla was a project that for a time didn’t appear to live up to expectations. Now it is of course the fire underneath Microsoft’s butt that has forced MS to start developing InternetExplorer again.
Several database projects have taken a while to get up to speed – that’s because the current niche for Open Source databases is the Internet, and that is currently loaded with MySQL and PostgreSQL. I expect the relevant formerly commercial and now Open-Sourced databases to get up to speed once Open Source becomes the standard in back-offices and data centres.
what do you think are the requirements? Well, I personally think the more stuff that gets open-sourced, the better. Do you want Solaris “cherry-picked” for its outstanding features? Release titbits, which are easier to integrate into previously-existing code bases. Do you prefer that it gets taken up on its own merits? Release as much as you can. It’s a lot harder to “cherry-pick” an orchard than it is to “cherry-pick” a single tree.
And as a Linux-user, I have to say I believe Andy Tucker that Sun is doing this for its own purposes, and not because “we’re out to “attack” Linux” or any such thing. As far as I can see – at least as far as my nose, the last time I looked – it’s an “insurance policy”, a “general public escrow”. This way Sun can always be in business – as a services provider at the very least. And this way it can ensure that its innovations can remain credited to Sun, and not snapped up by Our Favourite Monopoly …
Just my 0.02c
Sun really has changed their tune over the last few months. I wonder what made that happen. It must be a passion and love for this technology and certainly not an interest in remaining profitable and competitive. Or, well, that’s what I think.
Just my 0.02c
Multiply that by 100 and you’ll have two whole cents ;^)
I actually would like to see Sun release under a suitable F/LOSS license, most of SunOS 4.3 and previous, to The Unix Heritage Society http://www.tuhs.org/ enough of it to get a reaction from the *BSD side of the F/LOSS family where they got most of the SunOS base from.
And then decide what to do with Solaris, based on the reaction to SunOS.
And yes, 0.02c is probably heavily inflated.
who cares about what license, as long as its open source. Something that sucks because of its license is just political propoganda trying to support their ‘virtual license movement’ Ofcourse some licenses give some businesses more freedom which some can use to argue.. but still its old saying everything nonGPL sucks.
Oh yes.. By the way.. I quoted Johnathon Schwartz on here before.. Sun apparently does.. Tell me one thing that sun has specifically said (high profile exec or public statement) will happen that has not happen.
Sooo many people love IBM, well i hate IBM. Just because they are in a fight with SCO is no reason to like a huge company that is evil.
<sarcasim> well, i think we should all bash sun because </sarcasim> they have done more for open source than any other company other company
” but still its old saying everything nonGPL sucks.
”
old age saying?.
http://www.dwheeler.com/essays/gpl-compatible.html
“Oh yes.. By the way.. I quoted Johnathon Schwartz on here before.. Sun apparently does.. Tell me one thing that sun has specifically said (high profile exec or public statement) will happen that has not happen. ”
well. they say a lot of things. support solaris and drop support thrice. java will or will not be open source – many times. solaris will or will not be open source – many times. redhat linux is proprietary – many times. open source java will encourage forking – many times despite openoffice.
“Sooo many people love IBM, well i hate IBM. Just because they are in a fight with SCO is no reason to like a huge company that is evil.”
maybe someone will join this crusada against IBM for no apparent discussion. but why is IBM even mentioned here.?
“<sarcasim> well, i think we should all bash sun because </sarcasim> they have done more for open source than any other company other company”
back up those assertions which you have been claiming on every thread on SUN.
Gasp!Had I the appropriate computer hardware, I wouldn’t be running Linux.
Nothing wrong with Linux, but I spent four years learning Computer Science on Sun workstations, and had come to love their Operating System (and their C++ compiler before gcc). Unfortunately, when time came to leave school and get a real job, and I decided to wipe Windows from my newest built PC, the freely-downloadable version of Solaris didn’t want to run on my IA32 machine. Linux was the next logical choice.
The bottom line is, by opening up Solaris, there will be more developers running it, and hopefully a few more developers rolling drivers for their hardware. In the end, Solaris could gain some momentum and be a viable option for generic build computer systems, rather than be regulated to the awesome SPARC stations that Sun can build. (Note: SPARCs still pull a pretty sum of money from eBay every once in a while. They are beautiful machines, second only–perhaps–to Apple’s newest and sleekest computers…)
And the most important comment of all: IIRC, Sun has recently made new roads in Windows-compatibility via vendor agreements. Whether you like Microsoft or not, it is very common on the corporate Desktop, and this new deal will only bolster Sun in the server marketplace.
-Karrick
well. they say a lot of things. support solaris and drop support thrice. java will or will not be open source – many times. solaris will or will not be open source – many times. redhat linux is proprietary – many times. open source java will encourage forking – many times despite openoffice.
Sun has never dropped support for Solaris. They stopped Solaris for x86 which is not the same thing. Sun’s core business till now has always been Solaris/SPARC, Sun has never dropped support for solaris/SPARC.
OpenOffice is not the same as Java. Microsoft office is a monopoly in Office software, openoffice is the only alternative. No company is crazy enough to spend the money develop an alternative based on OpenOffice source, since StarOffice is cheap for the enterprise (included in JDS for $50/employee) and OpenOffice is free.
The Java specfication is implemented independently by many vendors from the begining, each of them would love to have a bigger pie of the java market, which is huge by the way. IBM, BEA, Oracle have huge stakes in the java server market. Each of them wouldn’t think twice before making incompatible changes to make sure thier competitors are at bay. The only reason that doesn’t happened is because Sun still holds control of Java and the J2EE certification.
There are multiple open source implementations of Java. The open source community loves to have multiple versions of the same software, rather than focusing on one good implementation.
redhat linux is proprietary by definition of the word proprietary.
pro·pri·e·tar·y
…….
Owned by a private individual or corporation under a trademark or patent: a proprietary drug.
proprietary
adj : protected by trademark or patent or copyright; made or produced or distributed by one having exclusive rights; “`Tylenol’ is a proprietary drug of which `acetaminophen’ is the generic form” [ant: nonproprietary] n : an unincorporated business owned by a single person who is responsible for its liabilities and entitled to its profits [syn: proprietorship]
Go with the logic here Redhat linux is a proprietary OS of which GNU/Linux is a generic form. Redhat is responsible for its liabilities and entitled to its profits.
However, Opensource Hacker’s have created thier own definition.
proprietary
1. In marketroid-speak, superior; implies a product imbued
with exclusive magic by the unmatched brilliance of the
company’s own hardware or software designers.
2. In the language of hackers and users, inferior; implies a
product not conforming to open-systems standards, and thus
one that puts the customer at the mercy of a vendor who can
inflate service and upgrade charges after the initial sale has
locked the customer in.
[Jargon File]
Source: The Free On-line Dictionary of Computing, © 1993-2004 Denis Howe
So it depends on who you ask. The answer to the question “Is Redhat linux proprietary?” will differ. And they both may be right.
So Schwartz was not wrong in claiming that “Redhat linux is proprietary”, open source hackers are not wrong in claiming that it is not, either.
Hope that puts the damn debate to rest.
“The fink project hasn’t really increased the number of people using Macs.”
How do you know this for sure in such way that you’re able to use it as a counter-argument against the existence of SunFreeware?
“its power management […] abilities can’t be beat”
Err. Transmeta CPU’s are used in low-power environments, like laptops.
“This way Sun can always be in business – as a services provider at the very least.”
In theory not, because the world ends somewhere. Then there’s no company called Sun anymore. In practice it depends on wether Sun provides a _better_ service than other service providers. And they’ll have competening companies. Big ones. Because of that i’d hate IBM if i were Sun or a Sun advocate. So Sun is better off if they have another insurrance to create money back at hand; for example hardware.
what do you think are the requirements?
Parts on which development is needed and where Sun sees an advantage to open source (OSI) it. I don’t know which ones, because i’ve used Solaris only a few times, but i’ve heard it has problems on hardware compatibility at x86. If Sun were to outsource that part, it would mean that on the longer run less software developers are needed to achieve a better product. Question is also wether SunOS kernel technically can use parts of other FLOSS kernels to achieve this.
It could include a non-commercial GPL* version which also uses the StarOffice/OO.o trick to get the source back in the proprietary version. I don’t think a non-commercial GPL version is open source (OSI) though.
* The same as the current proprietary, binary-only Solaris but then regarding ther source. You’d want the feature of the GPL which implies any deritave has the same license and you’d want the feature the source is available. Sun still is the owner of the original license thus is able to license that as they see it.
“who cares about what license, as long as its open source.”
Your definition of open source however is one which complies with Microsoft Windows and DEC OpenVMS licenses.
Andy Tucker also disagrees with you:
“Clearly we’ll need to release the code under an open source (i.e., OSI approved) license”
With that kind of attitude you blur the whole discussion into one based on your unique definition of a tech term widely acknowledged as it is; OSI approved.
“There are multiple open source implementations of Java. The open source community loves to have multiple versions of the same software, rather than focusing on one good implementation. ”
Dont play the semantic games here. Sun says redhat linux is not open source and is proprietary not in the business sense but as not open source sense. so they are wrong.
i dont want solaris GPL’d i just want it under an OSI license that allows it to be closed if needed. But Remember – A License arguement over various OSI licenses would be nothing but a political arguement!
What I meant OSI open source when I talk about open source Solaris 😉 It’s hard to tell nowadays.
As for the Java bashers I don’t see a reason to open source Java unless to fork it and exploit it and drive it in various ways in which other people want to drive it… There is no chance whatsoever that Java will be under an open source license that is OSI approved at all. Sun is most likely reffuses to just flat out and say it to keep open source people guessing. Theres still research going on about it but who knows!
Dont play the semantic games here. Sun says redhat linux is not open source and is proprietary not in the business sense but as not open source sense. so they are wrong.
I don’t think Raptor is playing semantic games any more than the original eweek article.
Schwartz went on to say that Red Hat’s price increases and proprietary extensions have lead to “CIOs figuring out that open source does not equal open standards. Open standards, which Sun has always supported, are better. Proprietary open source [like RHEL] can come back and bite you.”
http://www.eweek.com/article2/0,1759,1573433,00.asp
Schwartz was quoted directly and from the looks of it he never claimed that Redhat linux was not open source.
Read the article and Raptor’s post. Raptor’s post is right on target.
“”The fink project hasn’t really increased the number of people using Macs.”
How do you know this for sure in such way that you’re able to use it as a counter-argument against the existence of SunFreeware?”
Huh? I don’t even understand what you’re asking.
Every single market share report I have seen for the past 3 years has shown Apple’s total market share declining. But maybe you would provide a link to one that states otherwise?
I don’t even understand what “use it as a counter-argument agsinst the existence of SunFreeWare” means. What I meant was that the fink project serves the same role for OS X that SunFreeware and Blastwave serve for Solaris. While these services are wonderful for OS X and Solaris users, I hardly think that Linux or Windows users are going to jump ship just because of them. They might make people looking to switch OSes more likely to choose OS X or Solaris, but I highly doubt people would switch because of them.
“”its power management […] abilities can’t be beat”
Err. Transmeta CPU’s are used in low-power environments, like laptops.”
And their performance is less than impressive. As a whole, though, power management on PowerPC with OS X is much better than on x86 with ACPI.
Sun has done more for GNU and Linux than a lot of companies. Which is why I think constructive criticism is a good thing. We want them on our side.
Hell, we want Microsoft on our side, too. But don’t tell them that.
There’s nothing wrong with being a technology company that builds technology wisely, to make a profit, as long as they don’t forget they’re a technology company. A company who’s sole purpose is to make money is not a very good business for their customers or employees, unless, of course, they pay their employees and customers well.
But respect is also important. When employees are not being shown any respect what incentive do they have to show their employers any respect? Besides that phat paycheck…
Is it only about money or do we actually have to use these computers and cars and cake? Maybe I’m wrong, maybe it is only about money.