Mark Shuttleworth, the founder of Ubuntu Linux company Canonical, today published a blog posting on the lingering controversy over whether his company's code contributions are reasonably proportionate to the scale of its commercial activities. The South African one-time astronaut argues that "the body of free software needs many organs, many cells, each of which has their own priorities and interests. The body can only exist thanks to all of them."
He doesn't want his company to be measured just by what it contributes to the open source code base (which is little in some people's opinion and he doesn't seem to deny that). He realizes that commercializing open source "without contributing features might just feel like leeching" and basically wants a broader definition of what is considered a meaningful contribution, preferably including the kind of design, packaging and marketing job Canonical does.
This debate started a few months back and doesn't seem to go away soon.
More than a decade ago, when Red Hat went public and at some point attained a market capitalization of around $20 billion (about three times of what Red Hat is worth now), there was also some concern in the community about whether this was fair.
Even though the quantity of Red Hat's contributions to the Linux code base in recent years may by now appear reasonable compared to its share of the Linux market, the fact of the matter is that Red Hat didn't create Linux and made rather replaceable contributions to its adoption in the early years.
I'm not aware of any other company having derived so much commercial value from a product it neither developed nor acquired. Only in open source is this possible, it seems. Is that a good thing? I'm skeptical and I'll explain my view further below.
Red Hat gave stock worth millions of dollars to Linus Torvalds, and hired other key developers, such as Alan Cox, who became a Red Hat Fellow (now he works for Intel). While Red Hat's gifts to Linux developers were pocket change compared to its own value, there was no legal obligation and it was certainly a significant token of gratitude. But a Red Hat investor I talked to a few years ago nevertheless called the company's business model "parasitic" (and not "symbiotic", which is how Red Hat apologists would rather put it).
The second Linux distributor, SUSE, was acquired in early 2004 by Novell for about $210 million. That amount is dwarfed by the gains Red Hat realized, but still significant for a company that didn't create (or acquire) Linux either, apart from a pretty popular setup tool (YaST).
Now there are constant rumors that Red Hat, too, may be acquired soon. I don't know how much stock to put into that assumption, but any publicly traded company can be acquired if someone is ready, willing and able to lay down the right price. Red Hat's stock has done well in recent weeks, but is still significantly below its IPO price. So Red Hat has been a better deal for its pre-IPO shareholders and for short-term speculators than for long-term investors who came in early and held.
Another major open source project was built on the basis of a business model that is diametrically opposed to the Linux approach.
In 2001, I became involved with MySQL AB, the startup that made the namesake database and was later acquired by Sun (which in turn was bought by Oracle). That was the year the company was founded. When I first met the two most active founders (there was a third one who financed them early on), they had not yet moved into their office.
We had our conversation at the kitchen table in the appartment of one of them (David Axmark) in Uppsala near Stockholm. We had already started the meeting when there was some loud noise from the living room: Monty, the original author of MySQL, was just waking up. Shortly after the dotcom boom that was characterized by lavish spending and (figuratively speaking) marble floors, this was quite some contrast.
But the really impressive part was their IP-centric business model: they always ensured that if someone made an essential contribution to the code base, they would get the relevant intellectual property rights assigned. In some cases, this included that they hired capable contributors as part of the deal.
At the time we also discussed how copyleft (the obligation to publish derived versions of a GPL'd program under the GPL as well, unless they are used only internally) enabled the dual licensing business model. David said that -- even though it appears counterintuitive -- the stricter the copyleft rules are, the better it actually is for dual licensing because it creates demand for a non-copyleft license to the same program code. But the ability to make a GPL'd program available on non-GPL terms requires copyright ownership.
The European Commission (as well as antitrust regulators in Russia and China) looked into Oracle's acquisition of Sun of several more months after the US government had already approved the deal because it understood that MySQL wasn't a Linux-like project belonging to everybody and nobody at the same time. MySQL was, even though available under the GPL, essentially a company product and very much an IP-based business.
As a former strategy adviser to MySQL's CEO and shareholder of the company from the early stages until its acquisition by Sun, I knew about that. It just wasn't easy to explain this to other people because Oracle and Sun argued aggressively that an open source project doesn't need any particular company behind it. Interestingly, Richard Stallman agreed with Monty and me rather than with Oracle, Sun and the likes of Eben Moglen. He also explained the legitimacy of MySQL's business model (from a free software point of view) on the GNU.org website.
Given Richard's support, it's unbelievable that a pathological liar told the community that my work related to the merger control process was directed against the GPL. I debunked that smear in this recent blog posting. I argued for -- not against -- the GPL in connection with MySQL's business model.
The key thing in the current context is that MySQL never had the kind of debate over commensurate contributions that now surrounds Linux.
In the controversy over Oracle's acquisition of MySQL, some argued that MySQL would have been even more successful with a Linux-like model. Those who said so either just wanted to help Oracle and Sun push the deal through or some of them might have vastly (!) overrated their understanding of business issues (or both in some cases).
The people who took that position, including but not limited to Eben Moglen, certainly never built any (or any significant) business. MySQL AB was sold to Sun for $1 billion less than seven years after being founded, which should have upped the ante for those who thought they knew better.
Moreover, MySQL's lead venture capital investor, Benchmark Capital, was the original financier of Red Hat and certainly didn't lack an understanding of the Linux business model. Benchmark also financed eBay and other major successes. If a Linux-like approach had indeed worked better for MySQL, those experts would have been in a perfect position to identify and seize that opportunity, whose understanding of business models in general and FOSS business models in particular is hugely greater than that of Eben Moglen (who effectively joined Oracle's legal team for the merger case). The claim that MySQL picked the wrong business model is an insult to human intelligence, but such absurdities aren't against the law, so we have to live with them in all sorts of contexts.
The one lesson that I believe many more people in the FOSS community should learn from the Ubuntu debate is that intellectual property is a perfectly valid concept. Intellectual property rights (IPRs) were created by governments in order to ensure that innovators and other creative people get to benefit from their work. The IPR system may have its counterproductive protuberances, but the basic underlying idea is a good one.
I firmly believe that one can be pro-FOSS and pro-IP at the same time. That's been my approach ever since my involvement with MySQL started in 2001, and my appreciation of open source goes back to the time when my online gaming startup in the late 1990's used Linux, PostgreSQL and other free software on the server side.
I often disagree with the anti-IP rhetoric of many other FOSS advocates, to the extent that some radicals misperceive my support and respect for intellectual property as a hidden non-FOSS (or even anti-FOSS) agenda. I don't claim that FOSS and IP mix easily. It's non-trivial to strike the balance and reconcile the two value systems. But I won't waver in my tireless efforts and I won't be satisfied with anything less than the best of both worlds.
The fact of the matter is that whether or not one wants to categorize a Red Hat or Ubuntu business model as "parasitic", it certainly can't be a model for the economy at large. It can work for a few, but nature teaches us the limit: a biological system needs mechanisms that create real substance, not just parasites or little symbiotic creatures who can't exist without a host.
Even though I know that some are ideological about this and won't agree with me, I think FOSS will only do better if the community shows increasing respect for intellectual property and recognizes that in some areas, for certain kinds of innovation and creative production, IPRs are needed. One can deny that fact and claim that "sharing" and "the Commons" can take care of everything. One can even go as far as Eben Moglen recently did in India (a country that is way behind not only the West but also China in terms of respecting IPRs) and claim that "property" is a bad thing in general. In that speech, which was ideologically much closer to Fidel Castro than Hugo Chávez, he even claimed that mankind can only survive on the basis of "the Commons". I oppose such fundamentalism, and if necessary I will also call out people on hypocrisy if they preach water while actually drinking wine that is financed by patent aggressors like IBM.
I want to be thought-provoking, if necessary even irreverent, and identify and address problems. I don't want to preach to the converted and spread anti-IP propaganda because that only makes problems worse instead of contributing to solutions. Even when I fought against the EU software patent directive, I highlighted my pro-IP values.
It's not always easy to strike the balance. For instance, Richard Stallman argues that the term IP is a "seductive mirage" because of the differences between various IPRs. This isn't as bad as denouncing the concept of property. But RMS overstates the differences between non-material property rights. By rejecting the term IP without proposing a reasonably acceptable alternative, he raises concerns among decision-makers that the free software movement is anti-IP. His claim that ownership of software is immoral adds massively to that impression, even though the position he took on MySQL's dual licensing model shows a kind of pragmatism that a lot of people wouldn't consider him capable of having.
The whole debate over whether Ubuntu contributes enough code to open source is in my opinion the ultimate empirical evidence of our innate respect for rights that ensure a fair reward for true creators rather than free riders. Deep in our hearts, we all understand the legitimacy of intellectual property. We should not let an ill-conceived ideology turn our value system upside down.
Notwithstanding the foregoing, where IPRs are overreaching or where they are used in exclusionary and destructive ways, we should stand up and fight. But you won't seee me throw out the baby with the bathwater only to pander to an ideology.
If you'd like to be updated on patent issues affecting free software and open source, please subscribe to my RSS feed (in the right-hand column) and/or follow me on Twitter @FOSSpatents.