Weekly Edition
Return to the Distributions page
| |||||||||||||
Lumsden: OpenSolaris canceled
Alasdair Lumsden has posted a
document claimed to be a leaked Oracle engineering memo describing how
Solaris will be handled within the company. It suggests that OpenSolaris
as a community project (such as it was) is no more. "We will
distribute updates to approved CDDL or other open source-licensed code
following full releases of our enterprise Solaris
operating system. In this manner, new technology innovations will
show up in our releases before anywhere else. We will no longer
distribute source code for the entirety of the Solaris operating
system in real-time while it is developed, on a nightly basis."
(Log in to post comments)
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 20:09 UTC (Fri) by evad (subscriber, #60553) [Link]
So, sources for some (nearly all) of Solaris will still be released, but only after a major Solaris release is available. No community discussion or community development will exist. What then exactly is the point of keeping it CDDL-licensed at all?
What good is having the source code to a release if you can't contribute?
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 20:26 UTC (Fri) by zooko (subscriber, #2589) [Link]
If they do go ahead and open source it after releases, and if the IllumOS project is still alive by then, this would allow the IllumOS project to use their source code in its operating system.
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 20:31 UTC (Fri) by evad (subscriber, #60553) [Link]
Sure, but in the periods between releases the IllumOS project either sits on their thumbs or makes changes, and if they do the latter how to they reconcile those changes and decisions with the changes Oracle then releases much later? It would only work if they took Solaris 11 and then forked - forever.
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 21:26 UTC (Fri) by zooko (subscriber, #2589) [Link]
Yep. I hope they give it a shot. :-)
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 22:40 UTC (Fri) by jmalcolm (guest, #8876) [Link] Well, IllumOS says that they will contribute any changes that they make back to Oracle. So, I guess another possibility is that IllumOS is actively developed in the open, Oracle integrates IllumOS contributions into their code behind the curtain, and when Oracle occasionally does a major release, and drops a bundle of code from the sky, it will contain those changes. Either it would be easy for IllumOS to incorporate the changes or they would simply take the Oracle release as the base for their new release and go from there. I think the best thing that could happen to IllumOS is if they forked forever and only worried about taking in major new features from Solaris. For example, if Oracle introduced something as useful and innovative as ZFS in the future then IllumOS would probably welcome the chance to integrate it even if there was a bit of pain. In a way, I think this could be an opportunity for IllumOS. I think the OS has real potential but tying themselves so closely to the Oracle code base was going to be a real anchor in my opinion. If they can take what they have and build a real community, it could become a real alternative.
Lumsden: OpenSolaris canceled Posted Aug 14, 2010 1:51 UTC (Sat) by Cyberax (✭ supporter ✭, #52523) [Link]
Looks like the most pointless 'free' project of all time.
Most companies building business around OpenSolaris will, probably, migrate to Linux soon. Just give another couple of years for btrfs to stabilize :)
btrfs is doomed Posted Aug 14, 2010 8:20 UTC (Sat) by khim (subscriber, #9252) [Link] Remember: btrfs is Oracle project. It can be switched to OpenSolaris mode (we do all the development inside, code drop happens when we release our new enterprise offering) any moment.
btrfs is doomed Posted Aug 14, 2010 8:31 UTC (Sat) by rahulsundaram (subscriber, #21946) [Link]
Btrfs is explicitly not driven by a single vendor anymore. Red Hat has Btrfs developers as well for example. So no, Oracle can't do with Btfs what it did with OpenSolaris and there isn't a business incentive to do so either.
btrfs is doomed Posted Aug 14, 2010 13:24 UTC (Sat) by Cyberax (✭ supporter ✭, #52523) [Link]
No, btrfs is too big for that. Intel, RedHat, SuSE and other players already have a stake in it.
It'll be quite ironic, though, if btfs is used to kill Oracle's Solaris offerings.
btrfs is doomed Posted Aug 14, 2010 15:45 UTC (Sat) by drag (subscriber, #31333) [Link]
I don't think that Oracle cares.
In Oracle's eyes there is no Linux vs Solaris. They do not make money off of operating systems, they make money off of what runs on top of operating systems. They don't care if it's Linux or Solaris, just as long as it's competitive against Microsoft and other large corporate software vendors.
People always try to make everything out to a A vs B, but I don't think that Oracle really gives a crap one way or the other.
btrfs is doomed Posted Aug 14, 2010 20:10 UTC (Sat) by leoc (subscriber, #39773) [Link] I don't think that Oracle really gives a crap one way or the other.IMHO Oracle's only real goal is to make as much money out of Sun's IP as quickly as possible even if it ends up killing Solaris and/or Java.
btrfs is doomed Posted Aug 14, 2010 20:21 UTC (Sat) by drag (subscriber, #31333) [Link]
Not Java.
Oracle cares deeply about Java and that is probably the #1 reason for purchasing Sun.
Oracle's business depends entirely on Java... it's their vertical application stacks and other big enterprise offerings that are core to their business.
If Java was to fail the next closest thing that would step into the arena would be .NET and that is Microsoft's baby and their direct competitor.
They are killing Java Posted Aug 15, 2010 6:37 UTC (Sun) by khim (subscriber, #9252) [Link] Well, it looks like they really care about Java. But it's obvious that they are ready to kill it in exchange for something. What can they want which is big enough to kill Java? It's hard for me to even imagine... Here is good article on subject.
They are killing Java Posted Aug 19, 2010 15:10 UTC (Thu) by vonbrand (subscriber, #4458) [Link] A succulent piece of the smart phone market (and others to follow)? I just heard that a fourth of new phones sold are Android-powered right now, ahead of Apple and others.
They are not that stupid... Posted Aug 21, 2010 17:35 UTC (Sat) by khim (subscriber, #9252) [Link] Well, any success on this front will be short-lived. Google can not just go and drop Java-based development tools, but it can switch Android to C++ or Go in two or three years. Even if Oracle wins in court, it loses: it gets paid for these two or three years quite handsomely but ultimately it loses both smart phone market (because Java tools are eradicated) and Java market (because it becomes irrelevant).
btrfs is doomed Posted Aug 14, 2010 20:22 UTC (Sat) by Los__D (guest, #15263) [Link]
Hmmm, these changes seems to indicate that they are hell-bent on making more money on Solaris.
btrfs is doomed Posted Aug 14, 2010 22:53 UTC (Sat) by drag (subscriber, #31333) [Link]
Whoever was put in charge of the division in charge of Solaris wants to make money from it. That's his/her's job and they probably take it seriously.
But that's not the same as Oracle caring about Linux vs Solaris and having a pony in the race.
If they do care about A vs B then they are stupid. Linux is more popular then Solaris: it's used by more customers, supported by more vendors, better hardware support, better application support, etc etc. In about every category it's a much more competitive OS then Solaris is.
Oracle has to make Linux better at running their databases and they have to make it better at running Java. That is what helps them competitively in the long run. Anything beyond making Linux run their applications faster and more reliably... I don't think they care. (example: Android or KDE)
It's just self-interest.
I expect you'll find that the 'Solaris 11' being 1337 is more window dressing then anything else in order to keep their existing customers happy. They are not going to be able to manage making the difference between Solaris 10 and Solaris 11 bigger then the difference between Solaris 9 and Solaris 10 when the men/women largely responsible for the improvements in 10 are now gone...
Coming soon... Posted Aug 16, 2010 12:15 UTC (Mon) by pboddie (subscriber, #50784) [Link] I expect you'll find that the 'Solaris 11' being 1337... Given the tendency of Sun and Oracle to invent new and bizarre names for existing products, sprinkling that unicorn dust all over for that extra sparkle, the odds are probably already quite short on the next release being known as "Solaris 1337".
Coming soon... Posted Aug 16, 2010 14:34 UTC (Mon) by SEJeff (subscriber, #51588) [Link]
But it can't be magical... That word is trademarked by Apple
btrfs is doomed Posted Aug 15, 2010 8:51 UTC (Sun) by nim-nim (subscriber, #34454) [Link]
Oracle as a company does not care. Oracle has always consistently been about the money first. If Oracle ever makes mistakes is because they're too greedy, not because they paused to consider other motives.
However, I wonder if they took the precaution to firewall their Solaris and Linux teams. If they were to consolidate them in a single organisational unit, and put ex-SUN people at its head (because their new Solaris team is probably more numerous than the Linux one ever was), I can see relationships with Linux projects go south fast. SUN had a deeply-ingrained "UNIX is best", "we're open, but please sign here first, and we'll be back to you someday", and "everyone is equal, but we're more equal than others" attitude that completely soured their relations with other FLOSS actors.
btrfs is doomed Posted Aug 14, 2010 16:48 UTC (Sat) by niall.noigiallach (guest, #47469) [Link]
Well Oracle could stop contributing to Btrfs, then if they are feeling vindictive they could initiate a patent lawsuit against Redhat/Novell etc using patents they acquired from Sun (ZFS etc.)
All that matters to Oracle is their profit margin.
btrfs is doomed Posted Aug 14, 2010 15:57 UTC (Sat) by zooko (subscriber, #2589) [Link]
Note that what you described is *not* OpenSolaris mode. OpenSolaris did all development in the open. That mode has just been cancelled. What you describe is Solaris >= 11 mode.
btrfs is doomed Posted Aug 14, 2010 19:23 UTC (Sat) by HenrikH (guest, #31152) [Link]
Isn't btrfs a Chris Mason project and he happened to get funding from Oracle, i.e grants from Oracle that he could spend his working hours on the project. That is not the same as it beeing an Oracle project.
btrfs is doomed Posted Aug 14, 2010 21:28 UTC (Sat) by clump (subscriber, #27801) [Link]
Much depends on whether/what Mason signed while working for Oracle. Many American technology employers claim sole ownership of anything developed inside their walls. It behoves the employee to make sure he or she isn't giving away rights.
btrfs is doomed Posted Aug 15, 2010 1:23 UTC (Sun) by drag (subscriber, #31333) [Link]
It's generally assumed that when your working under contract that the work you create is owned by the person your working for.
If you want it to be otherwise you'll have to take extra steps.
btrfs is doomed Posted Aug 15, 2010 3:49 UTC (Sun) by tseaver (subscriber, #1544) [Link]
> It's generally assumed that when your working under contract that the
> work you create is owned by the person your working for.
Not so, at least in the U.S: unless you are an employee whose employment
> If you want it to be otherwise you'll have to take extra steps.
Exactly vice versa (under U.S. law).
IANALAIDPOOTV, of course.
btrfs is doomed Posted Aug 15, 2010 17:02 UTC (Sun) by butlerm (subscriber, #13312) [Link]
> Not so, at least in the U.S: unless you are an employee whose employment
> contract explicitly assigns such rights to your employer
In the U.S., a work "prepared by an employee within the scope of his or her employment" is a "work for hire", the rights to which automatically go to the employer:
"In the case of a work made for hire, the employer or other person for whom the work was prepared is considered the author for purposes of this title, and, unless the parties have expressly agreed otherwise in a written instrument signed by them, owns all of the rights comprised in the copyright." (17 USC 201(b))
The default rules for contractors are different, i.e. the assignment (if any) must be made the other way.
btrfs is doomed Posted Aug 15, 2010 15:05 UTC (Sun) by nicooo (guest, #69134) [Link]
% grep Oracle /usr/src/linux/fs/btrfs/* | wc
44 391 3835
% grep "Chris Mason" /usr/src/linux/fs/btrfs/* | wc
0 0 0
btrfs is doomed Posted Aug 15, 2010 21:59 UTC (Sun) by nix (subscriber, #2304) [Link]
Fundamentally, Chris is the person who designed and wrote a lot of btrfs and who knows how it works. The fictional person called 'Oracle' knows none of that, so if Chris were to leave Oracle would have to scramble, but Chris could keep working on it (it is, after all, GPLv2) and the results would still benefit Oracle. I'm sure Oracle's management know this. Attacking btrfs is as stupid as attacking any other part of an OS critical for their databases. (Note: nobody is running Oracle DBs on Android phones.)
btrfs isn't doomed ;) Posted Aug 16, 2010 15:28 UTC (Mon) by masoncl (subscriber, #47138) [Link]
I think a good metric is to look at non-oracle copyrights in the btrfs sources. Basically btrfs was done from day one as a project derived from the kernel and with significant contributions from companies other than Oracle.
This was done for practical reasons, it's really the easiest way to make a Linux filesystem these days. Also, when you try to make something general purpose, you end up needing a lot of help...Linux has a lot of purposes.
The end result is that Btrfs is a community project where someone at Oracle is the maintainer. I try very hard to be inclusive and while I don't always succeed in keeping up with the contributions, Oracle definitely understands the importance of community in keeping the project alive and useful.
Btrfs isn't unique inside of Oracle, ocfs2 is a good example of a more established project quietly trying to do the right thing.
-chris
btrfs isn't doomed ;) Posted Aug 16, 2010 17:04 UTC (Mon) by nicooo (guest, #69134) [Link]
> Oracle definitely understands the importance of community in keeping the project alive and useful.
I think they should make a public statement if they want to put a stop to the conspiracy theories that are popping up.
Lumsden: OpenSolaris canceled Posted Aug 14, 2010 16:00 UTC (Sat) by zooko (subscriber, #2589) [Link]
> Well, IllumOS says that they will contribute any changes that they make back to Oracle.
They said that before this memo was leaked and I presume that they assumed that the OpenSolaris development mode would continue. Now, I wonder what they will do.
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 22:49 UTC (Fri) by MattPerry (guest, #46341) [Link]
> What good is having the source code to a release if you can't contribute?
I thought it would have been obvious. The answer is so that users can customize it for their own needs. If I had the source to Windows, I can think of some improvements that I would make, or my company would make, regardless of whether we could contribute back or not. At least with Solaris, customers/users won't be stuck without source.
Lumsden: OpenSolaris canceled Posted Aug 14, 2010 9:41 UTC (Sat) by lkundrak (subscriber, #43452) [Link]
You'd break the "warranty seal."
Lumsden: OpenSolaris canceled Posted Aug 14, 2010 19:16 UTC (Sat) by oblio (guest, #33465) [Link]
You can do that if you're big enough. You can see the source. I think you can even modify it :)
But if you customize it, as the previous poster said, you void the warranty (it's logical after all).
Lumsden: OpenSolaris canceled Posted Aug 14, 2010 8:53 UTC (Sat) by lkundrak (subscriber, #43452) [Link]
Well, many people who are not developers, but merely end-users (me including) do read a lot of code, while contributing little. It happens a lot to me that I don't have an idea about some limit, cause of unexpected behavior, a bug in software, or am just curious. A most natural response to it is looking into the code. Those are the times where I don't care much about politics behind the code.
Surely, doing development behind closed door and releasing huge chunks of source code regularly without any community development closes a whole lot of possibities, but saying it's completely useless is not true at all.
Lumsden: OpenSolaris canceled Posted Aug 14, 2010 9:24 UTC (Sat) by evad (subscriber, #60553) [Link]
Well, to be fair, I posed it as a question, not a statement.
Kill Solaris Posted Aug 13, 2010 20:11 UTC (Fri) by midg3t (guest, #30998) [Link] For a moment I thought they were killing off Solaris altogether. I can only hope it's on the roadmap.
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 20:14 UTC (Fri) by ESRI (guest, #52806) [Link] The questions this brings up for me are...
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 20:27 UTC (Fri) by zooko (subscriber, #2589) [Link]
> How will this affect source code accessibility between development cycles?
Sounds pretty clearly that the source code won't even be visible to the public until after the release of the next (proprietary, binary-only) Solaris.
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 22:46 UTC (Fri) by jmalcolm (guest, #8876) [Link]
Right, it sounds like if you want to know how the next version of Solaris is shaping up then you are out of luck.
IlumOS (if it exists) will be developed in the open though so you can get involved in that.
Not a real option if you want to stay in sync with Oracle though.
Lumsden: OpenSolaris canceled Posted Aug 16, 2010 14:22 UTC (Mon) by TRS-80 (subscriber, #1804) [Link] There will "Solaris 11 Express" binary-only releases, which is how things worked before OpenSolaris, so you can get an idea of what features will be there, just you won't get the source until the final release.
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 20:16 UTC (Fri) by juliank (subscriber, #45896) [Link]
Nelson: Haw-Haw!
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 22:19 UTC (Fri) by clump (subscriber, #27801) [Link]
Many former Sun customers are already unhappy enough. Squirreliness like this isn't likely to cheer them up.
Funny Simpsons reference too.
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 23:39 UTC (Fri) by clump (subscriber, #27801) [Link]
Er to clarify, my comment about 'Squirreliness' was directed at the memo from Oracle.
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 20:28 UTC (Fri) by mbaldessari (guest, #36769) [Link]
If this is the approach they have with (Open)Solaris, I really wonder what will happen next to OO.org and java (don't care much about mysql)
Lumsden: OpenSolaris canceled Posted Aug 13, 2010 22:54 UTC (Fri) by jmalcolm (guest, #8876) [Link]
The fact that Oracle is suing Google over Dalvik should tell you what Oracle thinks about a free and open Java.
Keeping OpenOffice.org going might actually be useful for sticking it to Microsoft. Larry probably does not mind that it competes directly with Microsoft's most profitable product line. Another plus is that it has almost no chance of morphing into anything dangerous to Oracle.
It is not impossible that Oracle will invest in OpenOffice.org more aggressively than Sun did.
In addition to Solaris, I am interested in how Oracle stewards MySQL and VirtualBox. Selfishly, VirtualBox is a part of my daily toolbox and I hope it sticks around (and is relevant). I have a feeling that MySQL is more strategic than people realize for Oracle. It will probably see some love.
Status of MySQL.... Posted Aug 13, 2010 23:45 UTC (Fri) by dps (subscriber, #5725) [Link]
Even if MySQL not strategic for Oracle I know at least one $$$$$$ product that cares about MySQL and can definitely live with the GPL. If Oracle did drop MySQL then users like my employers might help maintain it for commercial reasons.
You could get the GPL bits of the $$$$$ product mentioned above but would be better off getting a proper Linux distribution instead. The interesting bits, like a geographically aware DNS server which returns something close to you that is not too busy, are all commercial.
I am aware of at least one of other product which is open source modulo most of the interesting features.
Maintain? Yes. Develop? No. Posted Aug 14, 2010 8:29 UTC (Sat) by khim (subscriber, #9252) [Link] You can maintain MySQL by adding bugfixes to it, but the moment you'll add significant functionality to it will be the moment you'll subpoena. Oracle already shown what it thinks about GPL business: yes, you can use all this code and patents... but only if you don't add new functionality we've patented before. Good luck trying to find some new functionality in databases not patented by Oracle. All free Oracle products are now zombies (neither alive nor dead): you can use them (because license says so), but you can't actually develop them (because license does not cover additions - and Oracle is quite happy to sue you if you add something it does not like). The only exception is probably Linux kernel: so many powerful companies are involved that even Oracle will think twice before suing anyone. But all the projects stewarded by Oracle? They are zombies now and we should think what we can do to replace them...
Lumsden: OpenSolaris canceled Posted Aug 14, 2010 2:00 UTC (Sat) by ewan (subscriber, #5533) [Link] The fact that Oracle is suing Google over Dalvik should tell you what Oracle thinks about a free and open Java.I'm not so sure. We're used to thinking of patents being used to keep things from happening, but I rather suspect that Oracle isn't interested in killing Dalvik, they're interested in a slice of the money.
Lumsden: OpenSolaris canceled Posted Aug 14, 2010 2:31 UTC (Sat) by basic (guest, #12705) [Link]
Well, depending on how they "slice" it, it might not be "free and open" anymore after that. Not to mention that in the public's eye right now, it is already less "free and open".
Lumsden: OpenSolaris canceled Posted Aug 14, 2010 4:16 UTC (Sat) by jmalcolm (guest, #8876) [Link]
Oh, I am sure that Oracle would rather see a healthy Android marketplace that serves as an Oracle revenue stream. But the process of monetizing Dalvik by Oracle will happen through a reduction in the freedom and openness of the Java platform (at least as it is perceived today) for everybody else.
It is obvious that Larry Ellison does not see "developers, developers, developers" as having any great benefit for Oracle. Making sure that big industry players are paying them licensing fees is more important.
I'm not even sure Google can pay for Dalvik Posted Aug 14, 2010 8:41 UTC (Sat) by khim (subscriber, #9252) [Link] Well, Google may agree to pay for Androids sold in US but not for Androids sold abroad. Then eventually just abandon US market altogether. It's just 8% of world telecom market, not a huge deal. The problem with this plan is the fact that Android is more popular is US then abroad. Google will lose momentum if it'll stop sales of Androids in US so it's not a good short-term strategy. We'll see what actually happens soon...
I'm not even sure Google can pay for Dalvik Posted Aug 16, 2010 5:57 UTC (Mon) by klbrun (subscriber, #45083) [Link]
Since Google is fighting this, do you suppose the end result will be to further restrict software patents? I'm not saying Google plans to change software patents, it just might be the result of the legal process.
There are too many unknowns... Posted Aug 16, 2010 7:14 UTC (Mon) by khim (subscriber, #9252) [Link] We'll know it... seven years in the future. The biggest short-term danger for Google is the injunction. If the court will not stop sales of Android phones then Google can do many things. Google is not TomTom, they have money to fight this fight, but if court will actually declare that "harm" is irreparable and stop sales of Android phones then Google will probably fold pretty soon: Android is not popular enough abroad to just go and abandon US market.
There are too many unknowns... Posted Aug 16, 2010 8:32 UTC (Mon) by nim-nim (subscriber, #34454) [Link]
> Android is not popular enough abroad to just go and abandon US market.
Sources?
Mine say Android has reached 17,2% of worldwide smartphone sales lately, which would not be possible with the USA market alone (even with insane 100% local penetration)
This is not just matter of popularity... Posted Aug 16, 2010 8:58 UTC (Mon) by khim (subscriber, #9252) [Link] 17,2% are cool, but they can become 1,72% very fast if handset manufacturers will switch to other platforms. There are only one company which committed most of it's resources to Android and can not switch easily: Motorola - and this is exactly the company which needs US market the most and will probably just die away without it. The fate of AndroidMarket is uncertain without US. And so on. Google can not abandon US market right now. May be in a few years, but not today.
Lumsden: OpenSolaris canceled Posted Aug 14, 2010 4:47 UTC (Sat) by basic (guest, #12705) [Link]
With the latest patent lawsuit, I'm more concern with other patents that SUN might own. Do they have MySQL related patents? I'm sure Oracle have other DB and other patents. Do they have OpenOffice.org/StarOffice related patents?
Does this mean that whenever Oracle find an opportunity to use any of these patents to their advantage (fairly or not), they will not hesitate to go to court? Do we need to come up with some way to protect against Oracle patent lawsuits specifically? Or worst, is this a trend in the making?
OpenOffice.org might lose Oracle's support if Oracle decide that they can make more monies (or lose less monies) if they shut it down and make proprietary. Them investing into the software does not mean that they will always keep it open. The same with MySQL, of course they might get into trouble with antitrust authorities on that one.
I've said this before and will say it again... Posted Aug 14, 2010 8:50 UTC (Sat) by khim (subscriber, #9252) [Link] All free software Oracle projects are now zombies (Java, MySQL, OpenOffice.org, etc). The only possible exception are Linux kernel related projects (btrfs?) because powerful companies are involved in Linux kernel so even Oracle will think twice before suing anyone - and even there they can easily freeze development and only show code drops when new release of Unbreakable Linux will be published. The licenses involved (mostly GPL) are already tested in court (I don't care about OpenSolaris) and this means you can use them freely as long as you are only adding simple bugfixes (implied patent license will probably sav your bacon). If you add significant functionally (go-oo?) - be prepared to see subpoena in your lap (licenses don't cover and could not cover new additions). This means only Oracle can develop this software now and then it'll probably not bother to spend significant resources. So better start perceiving all these programs as "zombie form": they will be available for you in the form they are today, but tomorrow they'll be in the same form as today (may be with some bugfixes).
I've said this before and will say it again... Posted Aug 14, 2010 9:31 UTC (Sat) by nim-nim (subscriber, #34454) [Link]
If you add any significant functionally on any software you make yourself open to Patent attacks, so what's new?
And BTW, what are those GPL projects Oracle subpoenaed? The only ones Oracle is closing are precisely those that didn't use the GPL.
New player - that's what's new... Posted Aug 14, 2010 9:47 UTC (Sat) by khim (subscriber, #9252) [Link] If you add any significant functionally on any software you make yourself open to Patent attacks, so what's new? The actual attack is new. Compare it with IBM's stance: it's software was rejected before (LVM vs EVMS, NPTL vs NGPT, etc). And it's biggest patent holder on Earth. It grumbled for a bit yet never tried to force it's own implementation via court system. Oracle did. That's what's different. And BTW, what are those GPL projects Oracle subpoenaed? The only ones Oracle is closing are precisely those that didn't use the GPL. So not using GPL is somehow a crime? Sorry, but I don't buy this argument. You can relicense Android under GPLv3 (licenses are compatible) and the lawsuit will still not go away. The "crime" is to use non-Oracle-approved implementation - and license is irrelevant.
New player - that's what's new... Posted Aug 14, 2010 10:15 UTC (Sat) by nim-nim (subscriber, #34454) [Link]
> So not using GPL is somehow a crime?
Never wrote that. Just objected to to your "the licenses involved (mostly GPL" when the projects Oracle is actually trying to shut down are all anything but GPL (and not incidentally, they all made this choice deliberately)
Please reread what I wrote... Posted Aug 14, 2010 10:32 UTC (Sat) by khim (subscriber, #9252) [Link] The licenses of the projects Oracle attacks is irrelevant: Oracle can attack any independently-developed software. What does matter are licenses of Oracle's projects. Because GPL is well-established license now Oracle can not ask for royalties from users of Java (Standard Edition, GPLed one), MySQL (GPLed community version) or OpenOffice.org (again: Oracle's version, not Go-OO form). This means it's safe to use these projects but it's not safe to develop these projects. For if Oracle don't like your additions it can force you to drop it so it's stupid to even start to develop anything for these projects. Zombie state. Other projects (like OpenSolaris) use different, less tested licenses (like CDDL). These can be declared dead already - but thankfully there are very few of them (and the discussed article explains that Oracle just killed the biggest one). The most troubling one in the list is Berkeley DB: it's not GPL, but it's widely used and all potential users can be in trouble. We should probably think about replacement fast. But in some sense you are right - this actually shows strength of GPL and gives us time to think about Oracle abandoning strategy, but it does not change the fact that long term all Oracle projects are dead: you can only keep project as Zombie "alive" for so long. Over time it rots and disintegrates anyway... but it takes years (think qmail).
Please reread what I wrote... Posted Aug 14, 2010 20:31 UTC (Sat) by Los__D (guest, #15263) [Link]
In that case, does it really matter if it was a Oracle project or not?
If Oracle has patents on all interesting DB features, and are ready to use them, it does not matter if it is a new MySQL feature or a new PostgreSQL feature that ticks them off.
How does this make Oracle-started projects a bigger target for attack?
Not Oracle-started! Oracle-owned! Posted Aug 14, 2010 21:02 UTC (Sat) by khim (subscriber, #9252) [Link] In that case, does it really matter if it was a Oracle project or not? Yes, it does. Big time. How does this make Oracle-started projects a bigger target for attack? They are not Oracle-started. They are Oracle-owned! Remember? All the copyrights were assigned to Sun so Oracle owns everything. If Oracle has patents on all interesting DB features, and are ready to use them, it does not matter if it is a new MySQL feature or a new PostgreSQL feature that ticks them off. Recall the famous line: "This suit is specifically about Google and that's it." Oracle can say so because it's sole owner of copyrights and patents. It can decide what and when to do with any potential contributor/infringer. And this will be true for anyone who'll be foolish enough to participate in non-Oracle-blessed development of said projects - it can be kicked out anytime and noone will be able to do anything. But it's hard for Oracle to single-out one particular contributor to PostgreSQL. They can do this right after patch is accepted but at this stage it's trivial to just remove it and send Oracle away - but later, when feature is adopted and widely used and intermixed with other features... in essence any attack on one PostgreSQL developer is equal to attack on all PostgreSQL developers and users. This means it's not "Oracle vs someone else", but "Oracle vs everyone else" now. Big difference. Now, I'm not saying it's 100% safe to switch from MySQL to PostgreSQL or MariaDB (in fact I think MariaDB is not safe at all as it stands right now), but to "continue with Oracle platform and hope for the best" is stupidity itself. Of course there are always possibility "to play it safe": to work with Oracle and develop official version by fully accepting Oracle directions and removing everything Oracle does not like. But this will mean you are now unpaid employee of Oracle... why will anyone want this?
Not Oracle-started! Oracle-owned! Posted Aug 14, 2010 22:09 UTC (Sat) by Los__D (guest, #15263) [Link]
This makes no sense. If i.e. MySQL is forked and multiple companies join in, this is absolutely no different than another non-Oracle related project.
The only worry here is patents, copyrights and who owns them does not matter at all, since distribution is already approved by the license.
Wel, you are right... kinda. Posted Aug 15, 2010 7:06 UTC (Sun) by khim (subscriber, #9252) [Link] This makes no sense. If i.e. MySQL is forked and multiple companies join in, this is absolutely no different than another non-Oracle related project. If it's forked and multiple companies join in then yes, it may be safe over time (you first need big enough userbase and Oracle can kill such coalition before it's powerful enough to resist it). What does not make sense is to work on Oracle-owned technology without forking it. This is what MariaDB does and this is why I'm not yet sure MariaDB is safe. Too few companies are involved is MariaDB - it still can be crushed easily. The only worry here is patents, copyrights and who owns them does not matter at all, since distribution is already approved by the license. Of course. This is what makes Oracle special. Years ago they were strongly against patents, yet today they are happy to use them to attack other companies directly. Similar to what Microsoft did. The remedy is the same: try to stay as far as possible from them. Just like with Microsoft is hard (they are big, own a lot of technologies which are a high demand today), but this is the only sensible approach. Forks make sense - but only as last resort, when the technology is valuable enough to risk it. All things being equal it's usually better to join some existing camp. I'm now investigating Vala: it looks similar enough to Java/C# yet totally different since it does not use virtual machine. It does not include GC, but it does include refcounting - and practically speaking it's enough: if you have such a complex structures as to make refcounting unusable usually you have too complex structures to write correct programs with them. I'm not yet convinced it's the way to go, so I'm looking on Go too. We'll see what will replace Java in the future, but it's clear that Java is thing of the past...
Wel, you are right... kinda. Posted Aug 15, 2010 9:34 UTC (Sun) by Los__D (guest, #15263) [Link]
Oh, agreed completely that it should be forked immediately. If it is big enough a target for Oracle to go after before enough has joined in, depends on a lot of factors, but I think that it usually wouldn't trigger Oracle until a lot of backers has already arrived.
While I'm not really all that interested myself, I think that MySQL, OO.org and OpenSolaris (well, at least Solaris) has proven big enough successes for great interest in a fork. Especially OO.org is also a key factor in the battles with MS.
If Java starts getting killed or closed by Oracle, I am sure that it will be forked, and that will be with a lot of companies behind it from the beginning. It is the core business for thousands of companies, and a big part of business for many more.
Java future is bleak... Posted Aug 15, 2010 10:12 UTC (Sun) by khim (subscriber, #9252) [Link] I don't really see what's the big deal about Solaris: yes, it does some things better then Linux but it looks like it'll be easier to duplicate the functionality and add it to Linux rather then keep the whole separate OS around. There are IllumOS but it remains to be seen how many backers and users it'll attract. MySQL is already forked and we'll soon see how it goes. As for OpenOffice.org... the problem here is the most successful fork (Go-OO) is well-known but is Novel-backed so it's not clear how it can help in "battles with MS". But Java... Java can be declared "dead for FOSS". Lots of Java users foolishly think they can avoid Google fate by using OpenJDK and even more of them just don't care (remember: Java was open-sourced just two years ago, before that it was proprietary... and it's backers were fine with it), but the problem here is the fact that Oracle is not Sun and it'll not spend billions on free product. Java will calcify more and more and over time will become the next COBOL. We can strike it out from the FOSS-relevant list of technologies. The only hope is Android: if Google will be able to defend Android and Dalvik then we'll have truly open J*va-based platfrom (like we had *nix-based platform for years in Linux) and then the FOSS development can be directed this way. But this is small hope: Google may say bold words today, but if Oracle will be able to stop Android sales it'll be different tune altogether. It's clear that people were abandoning the ship and Oracle was losing credibility in FOSS community for a long time, and now this process will speedup. But will it happen fast enough to save these projects from Oracle's iron patent-encrusted fist? This is open question...
Java future is bleak... Posted Aug 15, 2010 11:00 UTC (Sun) by Los__D (guest, #15263) [Link]
As I said, I don't really care about any of the three, but they all have a big following.
About Java, I think it is quite too early declaring it dead for FOSS (in fact I think it is way too early declaring anything about this situation), let's see how this plays out.
Please reread what I wrote... Posted Aug 19, 2010 15:24 UTC (Thu) by vonbrand (subscriber, #4458) [Link] Any addition to any project is potential patent grief. If you added some kind of database written in elisp to Emacs just the same as if you extend MySQL or Postgres, or ran afoul of some patent on code optimization in GCC, or doing JIT on Perl's intermediate code. The only difference is that Oracle is interested in milking Android for all it can get, and wouldn't expend the effort on going after Perl, Postgres or some MySQL fork. Simply because those projects haven't got the bread to line Oracle's spacious pockets.
There are bigger difference... Posted Aug 21, 2010 17:46 UTC (Sat) by khim (subscriber, #9252) [Link] The only difference is that Oracle is interested in milking Android for all it can get, and wouldn't expend the effort on going after Perl, Postgres or some MySQL fork. Simply because those projects haven't got the bread to line Oracle's spacious pockets. This is not the only difference and not even the most important. Sun traditionally demanded copyright assignment for contributions. It was always problematic practice but as long as these assignments were only used to sell proprietary offerings to help Sun to stay afloat it was justifiable to some degree. But now it's just stupid: Oracle will sue you for infringement while still retaining right for your work. Why will you want this? In a sense it means that btrfs is safe: this is rare Oracle-started project which is not also Oracle-owned. But most other projects are not safe and they must be forked or abandoned.
I've said this before and will say it again... Posted Aug 14, 2010 10:41 UTC (Sat) by AndreE (subscriber, #60148) [Link]
I've never thought of the Free Software movement as being about strict adherence to the current GPL license wording.
The principle of freedom of user and developer action that the GPL was intended to protect is clearly violated by this action.
The freedom of being able to create independently implementations competitor products is something the free software movement surely supports, even if it is not explictly protected by the GPL(v2 at least).
I've said this before and will say it again... Posted Aug 14, 2010 19:03 UTC (Sat) by allquixotic (subscriber, #61671) [Link]
"The freedom of being able to create independently implementations competitor products is something the free software movement surely supports, even if it is not explictly protected by the GPL(v2 at least)."
Correct, and this is exactly what Oracle management wants to squelch. If Oracle had their way, theoretically, then all of their currently open source projects would be re-licensed under a more restrictive license that allows you to read the code, make local modifications, but not distribute those modifications to any other legal entity. I'm not sure whether such a license would be legally consistent in very many jurisdictions, but management would create such a license if it thought that it could get away with it. I quote from the memo:
"We want the adoption of our technology and intellectual property to
I think the "before we do" is a conciliatory statement that was placed there just in case this memo gets leaked, and to reduce the chance that the reader will think they are intentionally trying to be anti-competitive. You might think, "Oh, they just want to be *first*, and *then* they'll let people compete with them". No. I think they want to stamp out competition *altogether*. That's the best way to improve their bottom line, right? -- they want to ensure that every Solaris user has to buy a license from Oracle, or else use a stripped-down, late, open source version that you can't properly fork into your own product without getting sued. Sounds almost as bad as MSFT. Almost.
I want to know what Red Hat is thinking and doing about this, in terms of moving their IP (sorry Stallman) chess pieces. I've always found Red Hat's decisions to be well-informed and sensible. If Red Hat starts moving away from projects stewarded by Oracle (OpenOffice and MySQL come to mind, as they are in the Fedora and Red Hat distributions), then we'd have a pretty good indication that it's high time to jump ship, and focus on replacements like KOffice and PostgreSQL.
Who's left? Posted Aug 15, 2010 3:42 UTC (Sun) by jone (guest, #62596) [Link]
Looks like Mike Shapiro got a bit more power now that Fishworks hasn't panned out as well as they had hoped (everybody makes an appliance) .. and now Oracle after the trial year (to see where to make deep cuts) probably isn't getting enough revenue from [Open]Solaris to sustain its engineering effort is probably looking to turn more maintenance and licensing revenue - so the logical step is to batten down the hatches a bit and try to retain more of the talent from walking out the door by taking out some of the legs they can stand on while promising to hire more talent and attempt to bring in some young blood .. so gauging from a few of the hordes who've left:
- Bryan Cantrill - Stephen Hahn - Sunay Tripathi - Bill Moore - Garret D'Amore (Looks like Adam and Brendan from team dTrace/Fishworks are still sticking it out with Mike) .. it sounds like they've got the option of either continuing to invest in Solaris, attempt a fork and potentially have their Licenses revoked and IP subsumed (not so nice looking at the CDDL from the other side), or dive into the linux kernel and attempt to start making things better there .. personally - i think it'd make sense to do the latter given the wide swath of things where there could really contribute now that they've taken off their jerseys and stepped into a larger co-contributing GPL biased world.
Who's left? Posted Aug 15, 2010 19:26 UTC (Sun) by dmarti (subscriber, #11625) [Link]
+1
(From a marketing POV, it looks like they're attempting to establish a Unique Selling Propsition for a product that's going into cash cow maintenance mode. And "The number one enterprise OS" is a better cash cow USP than "An open source POSIX OS, like Linux, but less so.")
Just wanted to say... Posted Aug 15, 2010 9:02 UTC (Sun) by khim (subscriber, #9252) [Link] The most fascinating thing here is the easiness with which Oracle ascended to the top. Just a week ago we had Microsoft as a "FOSS archenemy #1". The distant #2 Apple tried valiantly to win this title but was not vicious enough to even come close. And SCO was almost forgotten and it was not clear how long it'll be able to defend it's #3 position. But Oracle was able to claim the top spot for itself in just one day with it's one-two combination of Android lawsuit and Solaris closure! Well done, Oracle, well done indeed. I always knew Oracle is vicious but to go from "maybe friend of FOSS" to "FOSS archenemy #1" in a day? They are good. The sad thing is that Oracle is quite formidable archenemy: not only they are better organized then Microsoft (Microsoft was fearsome ten years ago, but today it's in disarray), but they literally own quite a few technologies important for today's FOSS ecosystem (lots of "normal user" can only use Linux because OpenOffice.org can read Microsoft formats, Berkeley DB is ubiquitous and so on) so they can not be just "cut off" from it! The fight will be long and bitter, I'm afraid... Oh, and they managed to clear Google's name in one fell swap - but this obvious: it's just like SCO did seven years ago. Lots of people were wary of IBM back then - but since SCO put IBM in David's spot the "big bad IBM" suddenly was "defender of the Holy Grail" and was loved again. Google is "defender of the Holy Grail" today - and gets the same love. The Holy Grail in question is not Linux, Dalvik or Android - these are important things but they are small potatoes in comparison to the Holy Grail of FOSS. The Holy Grail being: the ability to redo proprietary offerings. It's guarantee of eventual FOSS victory, but with this ability FOSS can hope to win the war against proprietary software. If masters of proprietary software can decide what can be reimplemented and what can not be ever rewritten - then all hope is lost. FOSS is defeated and we can only discuss terms of capitulation. GNU started as Unix reimplementation and Linux was reimplementation of critical part of it. Dalvik is reimplementation of "Java Machine". Parallels are obvious: Oracle attack is not directed to "sometimes ally of FOSS". it's directed to the heart of FOSS (or may be it's Achilles heel?), to the one thing which must be protected no matter the cost. The one thing which may make the best buddy of FOSS it's sworn enemy in a day (and Oracle was not best buddy of FOSS at the best of times). The Oracle declared war on FOSS. The only question: was it even aware they are declaring war on FOSS or were it's lawyers under illusion they are starting just another small shake-down operation? I doubt we'll ever know... but the closure of OpenSolaris hints that Oracle knew what they are doing: they decided they don't need the FOSS community anymore and can kick it to their liking. The war is scary, especially the war with powerful enemy and former ally like Oracle, but FOSS community have no choice... gauntlet is thrown and even if Google capitulates (like TomTom capitulated before Microsoft) it'll probably not stop the war. Only Oracle can stop the war - and it's just not in Oracle's DNA...
Just wanted to say... Posted Aug 15, 2010 10:24 UTC (Sun) by tialaramex (subscriber, #21167) [Link]
I don't know how Microsoft attained this #1 position in your chart.
They're all talk, talk, talk. An executive says "GPL will rot your teeth" and meanwhile they're shipping an OS with (optional) GPL'd components and providing source on their FTP server.
Obviously Microsoft would benefit from the non-existence of Linux, but they'd benefit from the non-existence of Sony, or MySQL, or a thousand other things, they have their fingers in every pie.
Microsoft was the only one... Posted Aug 15, 2010 10:54 UTC (Sun) by khim (subscriber, #9252) [Link] I don't know how Microsoft attained this #1 position in your chart. They limit freedom in many ways (especially on XBox), but they are hardly unique in this - Apple, Nintendo, SONY, TiVo and countless others easily surpasses them there. The unique Microsoft sin is related to patents and FOSS Holy Grail: it was the only serious company which tries to stop reimplementations with patent lawsuit (TomTom and others). And till recently they were prepared to do the same to Sambe - only fines measured in hundred millions stopped them. That's how it was able to take back #1 slot in the list. They're all talk, talk, talk. An executive says "GPL will rot your teeth" and meanwhile they're shipping an OS with (optional) GPL'd components and providing source on their FTP server. Well, talk matter too and like Apple they only allow GPLv2, but these are minor issues. And you are right: Microsoft was losing lead over Apple fast in recent years, but before they were able to lose their #1 position to Apple Oracle implemented it's one-two combination... and the rest is history... Now they fight for #2 position :)
FreeBSD and others? Posted Aug 15, 2010 11:14 UTC (Sun) by sgros (subscriber, #36440) [Link]
Why no one mentions what impact will this move, if it happens at all, have on all OSes that incorporated some of Solaris technologies?
DTrace is in MacOS X and FreeBSD, ZFS is in FreeBSD... now they won't have access to the latest developer release which means that when new Solaris version is released, along with the source, they'll need time to incorporate new features and bug fixes. In the mean time, they sit idle waiting for a new release... even worse yet, who'll fix the bugs in the mean time?
CDDL leaks Posted Aug 16, 2010 13:33 UTC (Mon) by marcH (guest, #57642) [Link]
Since most of the code will still be covered with the CDDL, could members of the Oracle Technology Network regularly, anonymously and legally leak the source?
CDDL leaks Posted Aug 16, 2010 16:38 UTC (Mon) by TRS-80 (subscriber, #1804) [Link] Snoracle still owns the copyright, so the OTN source could be stripped of the CDDL. Or OTN members could sign a contract saying they won't redistribute it. I've no idea how feasible either of these possibilities is.
CDDL leaks Posted Aug 20, 2010 20:42 UTC (Fri) by steffen780 (guest, #68142) [Link]
IANAL, but the copyright owner has the right to publish materials under any legal license. Whilst Oracle is (should) not be able to revoke existing licenses (ie. the CDDL of code already published) they're well within their rights to also publish that code, or improved versions of it, under any license they see fit. The same applies to the GPL etc., even the BSD license. The difference with the BSD license is that everyone has that right, "Apple's" MacOS X/iPhoneOS being the most high profile (ab-)user of this right.
CDDL leaks Posted Aug 21, 2010 2:35 UTC (Sat) by TRS-80 (subscriber, #1804) [Link] Yeah, I know Oracle can legally relicense to whatever they want. My comment about feasibility is to do with the effort of going through every file and stripping the CDDL header in the OTN version, then putting it back in the release version.
When Oracle just acts like Google Posted Aug 16, 2010 13:36 UTC (Mon) by alceste (guest, #39742) [Link]
As far as I can understand, the leaked memo shows that Oracle is refocusing its strategy around Solaris - but it does not (yet) mean that Solaris 11 will be turned into a completely proprietary operating system. I think that it is necessary to make a clear distinction between the development model of the OS, and its license.
Take, for example, the difference between Google's Android and Intel/Nokia's MeeGo.
Android is developed internally by Google, and the "community" receives occasional code dumps. Only a few Google commercial partners are allowed to follow its daily development and take technical decisions. Google made this choice it in order to keep its commercial advantage over its competitors: if they want an up-to-date Android system, they have to sign (and pay) a commercial partnership with Google. Otherwise, they'll have to either make their own fork of the project, or wait for the next public release --- that will be made available on the public repositories only when updated smartphones will have reached the stores.
On the other hand, MeeGo development happens on public repositories, using a public infrastructure, and several technical decisions are taken after public discussion. Intel/Nokia made this choice in order to speed up MeeGo adoption, even if competitors will be able to ship their products with the same software. It is the only possible way to gain some market share against Android.
Now, going back to Solaris: just take the previous two paragraphs, and replace "Android" and "Google" with "Solaris" and "Oracle". And replace "MeeGo" and "Intel/Nokia" with "Illumos" and "Nexenta/Belenix/etc...".
*If* the Oracle memo is not revised later, then Solaris 11 will still be (mostly) free software, as much as Android. And Illumos will be free software as much as MeeGo.
In other words, I would not use the Solaris issue as a proof that Oracle is "evil", unless Android is also considered a proof that Google is "evil" as well. (The Java issue, of course, is just another can of worms, and should not be confused with Solaris development decisions --- at least IMHO...).
When Oracle just acts like Google Posted Aug 17, 2010 1:15 UTC (Tue) by TRS-80 (subscriber, #1804) [Link] The difference is, Android has never been developed as a community project, while OpenSolaris has. And back in February Oracle said[12:09] <DanR> Oracle will continue to make OpenSolaris available as open source, and Oracle will continue to actively support and participate in the communityPerhaps not evidence of evil, but definitely evidence of going back on their word.
|
|||||||||||||