Who Owns Commercial Open Source – and Can Forks Work?

Three years ago, Tom Foremski wrote an interesting piece called “Adapt or die--the choice facing the open source movement“, which began:

Can Larry Ellison be stopped? By which I mean could Oracle shut down the fledgling open-source software movement through a series of acquisitions??

Consider this: This week, (not) coincidentally with the open source conference at the Argent Hotel in San Francisco, Oracle announced the acquisition of Sleepycat, an up and coming open source database; MYSQL said Oracle tried to buy it; and industry insiders are saying the acquisition of JBoss by Oracle is imminent.

In one fell swoop, Oracle has drawn a square around the most active and interesting parts of the open source movement--the databases and tools. These are the platforms for applications. Applications are just skins on the database--if you own the database (Oracle) or access to the data (Net Apps) you are in the sweet spot.

It concluded:

I'll say it again: In one fell swoop Oracle drew a square around the open source movement and unless it can prove that it can remain independent--it is a dead movement. Unless the open source movement reorganizes to meet this challenge it will dwindle and become an interesting footnote in the history of the computing industry.

Needless to say, I wasn't too convinced by the argument. Here's a comment I wrote on the story:

I think it's important to distinguish between the open source community and open source companies: you can buy the latter, not the former. Mr Ellison could buy every open source company he likes, but he can't control the communities that support them: all they have to do is "fork" the code - start a new version - which will effectively render his investment worthless. This ability to fork code - a central freedom of open source software - is what keeps the communities vibrant, and the companies honest.

At that time, this was a purely theoretical discussion, but with the acquisition of Sun and hence MySQL by Oracle, those points suddenly gain a new pertinence.

Someone else has been thinking about them, too – and it's someone who knows much and cares deeply about the open source project in question: Michael Widenius, founder and original developer of MySQL. Here's what he's written:

I don't think that anyone can own an open source project; the projects are defined by the de-facto project leaders and the developers that are working on the project. If the company loses the trust of these people, they can go away and fork the project and turn it the way they want to.

Sun's acquisition of MySQL did not go smoothly; most of the MySQL leaders (both commercial and project) have left Sun and the people who are left are sitting with their CV and ready to press send.

Oracle, not having the best possible reputation in the Open Source space, will have a hard time keeping the remaining MySQL people in the company or even working on the MySQL project. Oracle will also have a hard time to ensure to the MySQL customers, community and users that it will keep MySQL "free and available for all".

Now, Widenius probably knows the developers of MySQL better than anyone, and if he says that most have already left, and that those who haven't are about to depart, that's probably correct. This has major implications for MySQL, since it would leave the corporate side of the project hollowed out. But as both Widenius and I emphasised, that does not mean the end of the project itself. Indeed, Widenius is willing to get actively involved to ensure the latter lives on:

Here I see where Monty Program Ab, can play a significant role. Since I left Sun, I have been working on making it to be for Sun what Fedora is for RedHat. With Oracle now owning MySQL, I think that the need for an independent true Open Source entity for MySQL is even bigger than ever before.

The biggest threat to MySQL future is not Oracle per se, but that the MySQL talent at Sun will spread like the wind and go to a lot of different companies which will set the MySQL development and support back years.

I would not like to see this happen and I am doing everything I can do to keep this talent pool together (after all, most of them are long time personal friends of mine). I am prepared to hire or find a good home (either at Monty Program Ab or close to it) for all core MySQL personnel.

Here's why that is noteworthy. “Pure” open source projects like Fedora have always been set up alongside the commercial variants like Red Hat Enterprise by the company owning the copyrights on some or all of the code – Red Hat in this case. That is, they are very much “official” complements to the commercial versions. Widenius' plan is daring, because it seems to hold out the prospect of a MySQL fork set up independently of the copyright owners, and irrespective of their wishes. Under the GNU GPL, that's certainly possible, but I don't think it's ever been tried before, certainly not on this scale and with such an important project.

It will be interesting to see whether Widenius is able to pull this off, and whether Oracle, or whoever ends up owning MySQL, decides to help or hinder the attempt. There's always been a tacit assumption that it's not really viable to take this route, because the open source company simply has too much of an advantage through its ownership of the copyright, and the fact that it can always incorporate any code produced by a fork into the commercial variants. If this attempt to create a self-standing but quite separate version of MySQL succeeds, it could have major ramifications for all open source companies that think they own the project simply because they own the copyright of the code.

Follow me on Twitter @glynmoody.

Comments

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

GPL = copyleft, noone or rather everyone has the copyright

Anonymous's picture

Here's why that is noteworthy. “Pure” open source projects like Fedora have always been set up alongside the commercial variants like Red Hat Enterprise by the company owning the copyrights on some or all of the code

There's always been a tacit assumption that it's not really viable to take this route, because the open source company simply has too much of an advantage through its ownership of the copyright

You seem to have a fundamental misunderstanding of the GPL license. The GPL license is the opposite of copyright, any code release under the GPL is code that has been given away, the company who released it dos not own it anymore from that moment onwards! (at least not anymore than anyone else)

There is no such thing as a copyright owner of GPL'ed code.

It is you Mr Anonymous who doesn't know the GPL

Mark Lane's picture

No where in the GPL does the copyright owner give up ownership of his code. It gives the right to obtain the source, redistribute, modify. It requires distributed modifications to be released as freely as source code. In fact if the rights holder were to give up his rights by releasing with the GPL that would in fact make the GPL uninforceable.

True

goblin's picture

True.
Mr Anonymous is obviously mixing up "copyright" and "license", a very common error these days.

to clarify my point further

Anonymous's picture

I'm primarily objecting to this sentence:

because the open source company simply has too much of an advantage through its ownership of the copyright

The ownership of the copyright is completely hollowed out by the GPL license. Legally you are right, the original author is still the copyright holder of the code he wrote (but that already gets messy when you have many authors like it's common with opensource projects), but by releasing his code under the GPL license, he has given up all privileges that the copyright law entitles him to.
Therefore the copyright owner has absolutely no advantage compared to anyone else interested in reusing and/or redistributing his code.

You should try it out!

goblin's picture

Therefore the copyright owner has absolutely no advantage compared to anyone else interested in reusing and/or redistributing his code.

That's simply not true, Anonymous.

The copyright owner is the only one who can distribute his own work under any other license, if he wishes, regardless of having distributed a version under the GPL.

But there's no point in discussing this - try it out yourself:
Oracle/Sun/MySQL AB can distribute MySQL under any license they want, in addition to distributing under the GPL. But you can't. So I dare you to download a copy of MySQL and redistribute it under a closed source license. But be warned: All hell will break loose upon you!

After you've tried this, you will ask: "Why can Oracle/Sun/MySQL AB distribute under many different licenses, while I can't?"
And until you understand the difference between copyright and license, you wont see the answer...

no mixup

Anonymous's picture

I'm not mixing up copyright and license.

The GPL license gives ANYONE the same rights to the code as the author, the only condition is that everyone has to also supply the source code.

So by using the GPL license for his code, the author effectively shares his copyright with everyone else, therefore there is no exclusive copyright on the code by the author anymore.

You are making a very common mistake of seeing the GPL as just another license, while in reality it's a clever way of public-domaining your code whilst making sure any derivative work stays in the public-domain, too.

It <b>is</b> just another license!

goblin's picture

You are making a very common mistake of seeing the GPL as just another license, while in reality it's a clever way of public-domaining your code whilst making sure any derivative work stays in the public-domain, too.

The GPL is just another license. GPL stands for General Public License, so I don't even want to discuss this with you. It's a fact that you need to understand before you can see the distinction between copyright and the GPL.

Besides, in my country the term "Public Domain" doesn't even make sense, legally speaking. But licenses (like the GPL) do. So once again your claims are wrong.

Clarification

goblin's picture

"The GPL license gives ANYONE the same rights to the code as the author, the only condition is that everyone has to also supply the source code."

The original author of a software product has copyright over his own work. He can use this copyright to distribute his work under one or more different licenses.

Contributers to a GPL licensed project, can only distribute the derived software product under the GPL - thus, later contributers to the GPL licensed line of software products, do not have the same rights with regards to the original code, as the original author.

The forkers don't necessarily have to do it without funding

geoff_f's picture

Don't just assume that a community fork has to do it all without funding. What's to stop IBM (or similar) from providing funding to a community effort?

The recession.

AaronD's picture

The recession.

Indeed

Glyn Moody's picture

that could make even more interesting

Oracle's gunsights

MG's picture

My guess is that Oracle's primary target wasn't MySQL in this acquisition. I'm thinking they were gunning for Java, Solaris and also hardware capability (what will happen to HP-based Exadata?)

I would also guess that they would want to keep open source MySQL alive as the M in LAMP and provide a migration path to Oracle for users that need to scale up. Whether or not they can do this is up for debate.

Based on the way they've operated in the past, they'll probably continue to support the installed base of MySQL Enterprise customers but sunset this technology with an upgrade path to Oracle.

mySQL appliance may have scared heck out of Oracle

Jose_X's picture

I agree about Java

Glyn Moody's picture

Here's a first analysis I wrote of the deal.

Give MariaDB Another Look

Vivek's picture

Before the Sun acquisition, I was looking at the two main forks of MySQL, MariaDB and Drizzle. I've gotta say now, I'm starting to give MariaDB a serious look.

Source code evolves using some type of crane as Richard Dawkins likes to call it. The longer some code (MySQL) spends within the vicinity of some other code (Oracle DB), the greater chance it has to share common traits. Now I feel there's a disadvantage to open source sometimes especially when source code can be integrated into commercial variants, which we may see Oracle doing in the near future. It's almost not fair because Oracle not gets the best of both worlds, while the open source community doesn't get to tap into Oracle DB.

I forsee a very slow death to MySQL unless Oracle does something out of the blue. We're in the Web 2.0 generation and now and if Widenius can get good at popularizing MariaDB to other open source projects like phpBB, Mediawiki, Drupal, etc, who knows, it may turn out to be something like Gandalf the Grey morphing into Gandalf the White!

Why kill MySQL?

Alan's picture

Leaving aside the fact that they can't effectively kill MySQL, I don't see why Oracle would. I am not savvy to the business side of these things, but I manage postgresql, mysql, and oracle 9i/10g databases at work. I can tell you right now if MySQL went away, my migration path would be postgresql, not oracle. Oracle is a ten ton nightmare to admin, it does not scale down at all; patching oracle requires reading pages of whitepapers and sometimes bringing in a consultant, whereas patching mysql/postgresql requires running apt-get update on my Debian boxes.

Some may snicker at my lameness as a DBA, but that's the point. I don't need a certification or a masters in computer science to manage MySQL or PostgreSQL. So if Oracle thinks killing MySQL will sell more Oracle licenses, they're nuts. The only reason we run Oracle is because vertical app vendors give us the option of Oracle or MS-SQL server. If they gave us the option of MySQL or PostgreSQL, we'd be running that. Maybe that's the turf Oracle is defending, I don't know. But their product is too expensive, resource-hungry, and complicated to replace MySQL in the majority of our deployments, and I'll wager that's the case for most MySQL users out there.

If they really bought Sun for Java, then their best bet is to sell off MySQL to someone else.

What he said

KenM's picture

I fully agree with Alan. mySQL is a dream to administer, patch, and DBA. It takes as much time to do it as type this sentence. I rolled out 3 Oracle instances, the Linux ones take only a few hours, but Solaris? Ugh. It is obvious Oracle is already favoring the Linux side of things.

On a different topic, I'm not fearing Oracle mishandling mySQL. If they do, it will be forked to yourSQL or something and be better for it. What I'm concerned about is them being just useful enough that they retain control of the initiative, while they slowly load it up with 10's of thousands of lines of really, really slow Java.

In other words, just like Oracle 10 and 11.

-Ken

It's not so much killing as dying

Glyn Moody's picture

I think the interesting question is more whether Oracle will be able to keep MySQL going as a vibrant project if it loses all the top coders.

PostgreSQL

JC's picture

People should be using PostgreSQL anyway. Just forget about MySQL, a third rate ``database''.

postgresql.org

I don't get it

Morgauxo's picture

Why do these proprietary corps buy MySQL? First Sun, now Oracle... Don't they realize that their only two alternatives are to either provide an environment where the developers can flourish and compete with their non-free products or they can watch them all leave, fork the project and continue to compete. I just don't see how they think they are changing anything for themselves other than spending a lot of money.

That in mind, have the developers stopped and asked what Oracle intends to do with them? Maybe Oracle would hire back the ones whom already left Sun and support MySQL in an open source friendly way. Ok, that seems unlikely as it would make no sense to be competing with their own expensive proprietary database but it makes as much sense as any other reason anyone would buy the MySQL copyright.

Unless Larry Ellison really is that dense that he thinks he can shut MySQL down just by buying Sun. It seems unlikely he could have gotten so far if so. Maybe they intend to try to challenge the GPL and the communities right to fork... Hasn't the GPL already been validated in court though? That one would sure keep groklaw busy for a very long time!

One power Oracle gains from this purchase which no one mentioned... release MySQL as public domain. I really don't see how that would benefit anyone including Oracle but if MySQL forks right out from under them and they wanted to prove that the copyright they spent so much one is good for something... That's the one thing they could do just to prove they can.

Uncharted territory

Glyn Moody's picture

I think the problem for Oracle - and all big proprietary companies - is that we're inventing this stuff as we go along. There are no precedents: it's all new.

What fun.

what about Mambo/Joomla

Ross Kendall's picture

Perhaps it is not the same scale, but it was interesting to watch the original company that developed Mambo CMS (Miro) try to retain control while the community and core developers sought a new direction with Joomla. I suspect Oracle has little control over the future of MySQL (except perhaps the name, which will probably be forgotten, just like Mambo)

Yes

Glyn Moody's picture

But as you say, the scale is quite different, and I think that has a big impact here: MySQL is really important for many sectors.

it's still a precedent

Ross Kendall's picture

At the time of the Joomla fork, Mambo was the most popular open source cms on the planet, so in terms of community size it was significant. Although the company Miro were very small.

Another thought is that MySQL having a larger community only increases it's strength.

In the end I think Monty is right in pointing out that where the core developers end up (and if they support a fork) is probably the key issue.

The community owns a proj, if there is one of sufficient size...

Jim Gettys's picture

The X Window System has been effectively forked several times during its existence to overcome issues:
1) XFree86 was able to force the rescission of The Open Group's license change by (threatening) a fork (along with pressure behind the scenes of the UNIX vendors).
2) The more recent XFree86->X.org fork also preserved the fundamental freedoms, but could not have succeeded without general community support.

Freedom zero is the ability to fork code. But without a community of people of interest, this isn't practical for large projects.

I submit MySQL has sufficient interest overall outside of Sun/Oracle this can/will succeed if necessary, and that such communities "own" the code in a more fundamental way that copyright, so long as the license starts sufficiently "free". The license itself is a relatively unimportant factor here, once a sufficiently free license is in place.

> Ok so it's got a shocking

Anonymous's picture

> Ok so it's got a shocking name but PostgreSQL is better
>than MySQL and no one entity owns it.

> http://www.postgresql.org/
Right

> MySQL is ripe for the last step of its exit strategy:
> 1) Build MySQL as a vendor-driven OSS project
> 2) Sell MySQL, the company, to Sun
> 3) Fork MySQL after Oracle bought Sun and make MySQL a
> community-driven project.
Right

No, the GPL wins!

AJ's picture

There's always been a tacit assumption that it's not really viable to take this route, because the open source company simply has too much of an advantage through its ownership of the copyright, and the fact that it can always incorporate any code produced by a fork into the commercial variants.

I don't think the phrase in bold is right, you have it backwards.

In most cases though the company is also selling licenses under non-GPL terms (that certainly has been true for MySQL in the past). Widenius' ability to fork the code relies on the GPL, so any changes he makes in his version can *only* be available under the GPL and cannot be incorporated into a non-GPL version (unless he's willing to sign over copyright to Oracle, which doesn't seem likely). Oracle will therefore have no legal ability to use the fork's modifications in its proprietary versions. The converse is not true however; as long as future versions of MySQL are made available by Oracle under the GPL, Widenius will always be able to incorporate their changes into his branch of the fork.

Sorry, badly phrased

Glyn Moody's picture

You're right, it reads badly: I meant the commercial *open source* versions - those released under the GNU GPL, not the commercial ones, which means that Oracle could keep incorporating innovations from any fork into the main codeline. As you rightly say, *proprietary* versions could not, but I don't think MySQL has a proprietary version currently - just proprietary addons (anyone know otherwise?). Thanks for pointing this out.

MySQL Commercial licensing

goblin's picture

but I don't think MySQL has a proprietary version currently - just proprietary addons (anyone know otherwise?).

MySQL, and later Sun, offered to provide MySQL under other licenses than the GPL:
OEMs, ISVs, VARs and other distributors that combine and distribute commercially licensed software with MySQL software and do not wish to distribute the source code for the commercially licensed software under version 2 of the GNU General Public License (the "GPL") must enter into a commercial license agreement with Sun.

Thus there is an option to license MySQL database server and client libraries under terms different form the GPL. And I think this option is being used.

thanks

Glyn Moody's picture

So that's to OEMs etc., rather than directly to customers, yes?

Alternative licensing only makes sense when redistributing

goblin's picture

Correct. Paying for alternatively (non-GPL) licensed MySQL only makes sense when redistributing MySQL as part of a non-GPL product. Ie OEMs etc.

If the project was say GPL

Anonymous's picture

If the project was say GPL V2 or above, then yes they could have the changes back for their GPL 2 variant but they couldn't then sell a proprietary licensed version of this code though.

In fact as the fork owns the copyright for their code, they are in as strong a position if not stronger than the originator, if they have community backing.

For example a fork could go GPL 3 on a GPL 2 or above project, forcing the originating project to GPL 3 if they want the code back.

Other notable forks I can think of are ejcs and Cedega. The Cedega example is a good one to show that though a commercial company can blind-side a community and even encourage the community to think that they don't need to work on aspects of a project (like Direct-X), the version with community backing will, in the long term, succeed.

This also shows why Novell shouldn't think about forking Open Office fully. They complain about Sun's leadership of the project in the past but they don't have an abundance of community good will behind them.

Whoops. EGCS.

Anonymous's picture

Whoops. EGCS.

Mambo / Joomla

vicm3's picture

Yes I know is not the same scale, but has happened before, the best case I remember it's Mambo with the fork Joomla.

Excellent point

Glyn Moody's picture

As far as I can tell, they're both flourishing - anyone have more info?

Mambo who?

Ross Kendall's picture

I don't think Mambo could be described as flourishing but Joomla sure is

PostgreSQL!

pd's picture

Ok so it's got a shocking name but PostgreSQL is better than MySQL and no one entity owns it.

http://www.postgresql.org/

The optimal Open Source exit strategy

Sandro Groganz's picture

MySQL is ripe for the last step of its exit strategy: 1) Build MySQL as a vendor-driven OSS project, 2) Sell MySQL, the company, to Sun, 3) Fork MySQL after Oracle bought Sun and make MySQL a community-driven project.

In the end, all successful vendor-driven OSS products might end up as community-driven projects. Which is great for founders such as Monty: They will have the last laugh plus some cash and can continue what they always wanted to do: great software.

InnoDB?

goblin's picture

Thanks for sharing your thoughts, Glyn.
I really think you should comment on Oracles (indirect) ownership of InnoDB, the storage engine in MySQL.
How can this affect the scenarios concerning a fork of MySQL? Could Oracle put that initiative to sleep through their InnoDB ownership?

I've emailed Michael Widenius about this...

Glyn Moody's picture

...since he probably knows the details better than I do. I'll keep you informed.

Webinar
One Click, Universal Protection: Implementing Centralized Security Policies on Linux Systems

As Linux continues to play an ever increasing role in corporate data centers and institutions, ensuring the integrity and protection of these systems must be a priority. With 60% of the world's websites and an increasing share of organization's mission-critical workloads running on Linux, failing to stop malware and other advanced threats on Linux can increasingly impact an organization's reputation and bottom line.

Learn More

Sponsored by Bit9

Webinar
Linux Backup and Recovery Webinar

Most companies incorporate backup procedures for critical data, which can be restored quickly if a loss occurs. However, fewer companies are prepared for catastrophic system failures, in which they lose all data, the entire operating system, applications, settings, patches and more, reducing their system(s) to “bare metal.” After all, before data can be restored to a system, there must be a system to restore it to.

In this one hour webinar, learn how to enhance your existing backup strategies for better disaster recovery preparedness using Storix System Backup Administrator (SBAdmin), a highly flexible bare-metal recovery solution for UNIX and Linux systems.

Learn More

Sponsored by Storix