“Free software is an irrepressible urge to transmit and share knowledge” – Interview with Jean-Paul Smets (Nexedi)

This article is an excerpt of the book Open models published in French in 2014 and translated in 2016.

Jean-Paul Smets, CEO of Nexedi, publisher of the free software ERP5, generates 90% of his business from export. Fifteen years ago, he co-authored a book whose title speaks for itself: « LOGICIELS LIBRES. Liberté, égalité, business (Edispher 1999[1]) ». Today, he still believes in free software and tells us why.

[1] FREE SOFTWARE. Freedom, equality, business

What are the business models of free software that work? What are the most and least favorable areas for free software?

More than ten years ago, I wrote a book with Benoit Faucon about the free software economy. My goal was to promote free software to confront Microsoft’s absolute domination. I was, and still am, a big fan of free competition. The existence of a business model is for some executives what makes something acceptable that would otherwise not be. My goal was however to promote free software so I wrote a book for those executives that have this perception of the business model being the foundation of all human exchanges. In fact, doing so is as grotesque as considering business models the reason for sensuality: one does not make love because there is a business model. For example, few years ago, Redhat’s founder described his business model being like the one of the water delivery business, bringing gallons of water to his final clients. Saying that, we have moved the focus to the business model of bottling water rather than to why water flows from springs. Ten years after the writing of the book, what’s interesting for me is to know why water flows from springs.

What is the purpose of a free software?

It is not to respond to a financial need. Mainly people that do not have income problems produce free software. In developing countries, when feeding your family is an issue, some free software production exists, but on a much smaller scale. Since Shanghai has become a rich city, we notice that a small production of free software has begun. In fact, the free software producers’ geography is pretty similar to the OECD’s. The free software is initially based on a need to contribute to knowledge. It is an irrepressible urge to transmit and share knowledge. Of course, this need also exists amongst the more modest, but it certainly has less space to express itself.

The production of free software lacks of incentive. To finance this production and for it to last, what business model is best suited? How do we fund the R&D of free software on the long term?

This desire to produce free software might be important enough to make someone decline a doubled paycheck offer by Google. At Google, a developer is very well treated, fed, and even bathed like a child. Materially speaking, it’s interesting for the developer, but they contribute less to program sharing than in a free software company.

So the discussion around free software business models must be balanced by the recognition that none of these models explains the irrepressible desire to produce free software. Without this willingness, there would be no free software. Free software development goes on and on like when a researcher wants to continue his research and to produce articles on his discovery. Very few researchers give up and lose their desire to contribute and share about it. As for R&D funding, I remember what Richard Stallman said 15 years ago: “it’s simple, do consulting in the morning to pay the bill, and program in the afternoon”. This may seem unsophisticated, but it’s ultimately what works best. This model is by far the best and the simplest. It is sustainable and stable whatever the size of the company.

Is it possible in France?

This model is often used individually in Germany, much more than in France, since the German tax system is more favorable for young freelancers. For a freelancer in Germany, one week of work funds a monthly wage in Germany, allowing for 75% of his time devoted to R&D. In France, a JEI (Jeune Entreprise Innovante/ Young Innovative Company) can easily spend 50% of turnover in R&D with 10 employees. That allows the easy creation of stable free software publishers. After 7 years the company is recognized on the market and it becomes possible to slightly raise prices thus stabilizing income and allowing the employer’s social security contribution costs to be paid while investing in R&D. Other models have arrived since, but for me, these models lead to the corruption of free software.
One day, some big players in need of comfort (eg. Bull, Cap Gemini…) wanted to compare the quality of existing open source software so they anchored quantified comparison criteria. Rather than comparing software with each other, like kitchens chefs would compare gastronomy (we know what good food is when we’re used to eating well), they compared numeric criteria. One example of those metric measurements that has been recognized by the market is the measurement of the size of the community.

How is that a problem?

The first drawback of metric measurements is that they are linked to the idea that “I am no longer part of the community, others are”. For example some of our clients ask us “could you make your community grow?” and the immediate answer we get when we say “yes, sure, come on, join us!” is “oh no! Not us!” To be more precise, in certain areas of free software the community is clearly “others” and in other areas, the community is “I”. Regarding infrastructure softwares, like a kernel network module, “the community is me”. And both giant web/telecom companies and small groups of individuals, like Nexedi, see themselves in this “me”. All these companies desire a reliable Linux kernel, one that doesn’t crash when deployed to give a customer a service. This way there will be no hesitation for companies to contribute and patch into the correction of critical defects. What lies behind this logic is the basis for the access to knowledge, contribution and sharing. Why do we use free software at Nexedi? Because if we encounter a bug we will be able to fix it, provided we put in the necessary effort. Free software is a technology where nobody will prevent me from fixing bugs. In other areas free software is also often misunderstood in thinking that “the community, is the other”, especially in business management applications like ERPs. There are just few pioneers that think that “the community, is me”. By the way, this backwardness was analyzed by Brian Prentice from Garnter in his article: “Open Source & Business Apps – Is There a Disconnect?” For business applications software, big companies have not yet realized that it is their role to create the community by sharing non-confidential management processes, in addition to the software.

Is it the only drawback?

No, a second one with this idea of community size is that capital can be used to acquire it. “Since we measure a software’s quality by the size of its community, let’s use capital to acquire it entirely!” It’s easy. We go to conferences, we sponsor events, we build up a good image by giving the opportunity for others to express themselves then we hire talented and well-known performers for some rewarding missions and in doing so everyone will know that this famous guy works for this company etc. Capital becomes the instrument to quickly build up a community. Developers who, ten years ago, were managing and financing their communities the hard way by selling tee-shirts for example, now receive sponsored goodies and cookies during developers’ conferences. So saying that the size of the community is a so-called “quality criterion” gives an advantage to capital based business models while small structures with no capital or other goodies become invisible on the market.

“Developers who, ten years ago, were managing and financing their communities the hard way by selling tee-shirts for example, now receive sponsored goodies and cookies during developers’ conferences.”

Once again, isn’t this a French trend?

In France it became rather difficult to sell free software because of our sensitivity to marketing and slick looks. On the contrary, in Germany and Japan much more importance is given to free software. For instance in France CloudWatt has never looked at SlapOS, which nevertheless has been functioning since 2010. A Thales executive once tried justifying this fact: “if your software was good, it would have been known in the US!” Let me highlight that market structure is very different between France and Germany. In France the free software market is dominated by few integrators: Smile, Alterway, Linagora, and OpenWide which could also be software publishers like Linagora. These actors mostly integrate US-based free software promoted by strong marketing because French buyers give attention to American software and their marketing. “If your software was good, it would have been known in the US!”… Everything is revealed in this sentence.

You also mentioned Germany?

In Germany, there are regional companies with regional customers, and they make well-targeted small free software, well maintained and of good quality, that one can use outside its original region. Over there the market is very fragmented while in France there is a strong concentration around few actors.
But the most annoying thing with free software publisher companies receiving capital to fund their marketing and accelerate the acquisition of their community is that they no longer produce free software. They have a showcased free software and they sell a proprietary software. Believing they are buying free software, customers will in reality buy proprietary software and that ruins all the perceived benefits of free software, like being able to fix bugs and improve the product we selected. “Open source” uses this ambiguity. Ultimately, thanks to clever marketing around open source, the client understanding of what free software actually is is far from the real one which provides the ability to fix bugs independently. The “Cigref”, for example, spoke of Gmail as an “open source” product!

Why do you think that community size being a measurement of quality is a myth?

What really is the community in free software? Of course there are big communities such as Debian, which have nothing to do with marketing-funded communities. Debian can rely on a large community because it ensures packaging of tens of thousands packages, bricks relatively independent from one another. But in most free software community is rarely composed than more of 5 people. GRUB, the most used free software in the world is the boot loader used to boot Linux, Windows or BSD. GRUB’s community is only of three to five people. Developers of the heart of a file system project are a single person. When the only developer of ReiserFS was sent to jail, his software died. The new Linux Btrfs file system took three years to be developed and only one person was able to finish it. In short, in most cases, the heart of the community is composed of maximum 5 people, a hundred who produced a few lines one day and thousands fussing over it.

How do you see the future of free software?


One thing that I am sure of is that all business models consuming capital threaten free software. To be more specific, a company whose field is not related to free software and whose development requires capital (but which uses free software to develop itself) often contributes positively to the free software ecosystem. But in a company whose business is to produce software, capital often kills free software. As soon as existing companies in the free software space started to get their funding from venture capitalists, the balance between Europe and the United States in the success of free software has been broken and the very idea that free software has advantages faded. Why? Because the true free software became inaudible against well-marketed software companies financed by venture capital. Cloud computing, as a capital-intensive business model, is a significant threat to both free software and proprietary software. Developers who get offers of doubling or quadrupling their salary find it hard to resist. This is what happened in Japan, where Google hired two-thirds of the free software community. This also happened in France, to a lesser extent, with clouds named “sovereign”. One can conclude that the existence of free software is based upon the fact that people accept to earn half their salary to work at least twice as much, simply to satisfy their urge to produce free software.
Interview by Karine Durand-Garçon

Translation by Dominique Pasquier with the help of George Husni


Leave a Reply

Your email address will not be published. Required fields are marked *