MARSOUIN.ORG

Firms' contribution to open source software and the dominant user skill.

Jean-Benoît Zimmermann, Nicolas Jullien — 2009-03-16T16:33:40Z

Keywords: "Free"/"libre" or "open source" software, Industrial economics, dominant user's skill

JEL Classifications: L11, L15, L22, L86

Free, libre or open source software (FLOSS) is nowadays produced not only by individual benevolent developers but, in a growing proportion, by firms that hire programmers for their own objectives of development in open source or for contributing to open source projects in the context of dedicated communities. A recent literature has focused on the question of the business models explaining how and why firms may draw benefits from such involvement and their connected activities. They can be considered as the building blocks of a new modus operandi of an industry, built on an alternative approach to intellectual property management. Its prospects will depend on both the firms' willingness to rally and its ability to compete with the traditional “proprietary” approach. As a matter of fact firms' involvement in FLOSS, while growing, remains very contrasted, depending on the nature of the products and the characteristics of the markets. The paper asks why do for-profit firms contribute to FLOSS development and why some firms contribute more than the others. The common explanation is that FLOSS is often a complement to proprietary software (or hardware or services) that the for-profit firm sells at a positive price. We present an alternative explanation based the users' skill level. When users are skilled, opening the software is likely to result in a better product because the user base will contribute improvements (find bugs, write fixes and produce new features). We introduce the concept of the dominant user's skill and we set up a theoretical model to better understand how it may condition the nature and outcome of the competition between a FLOSS firm and a proprietary firm. We discuss these results in the light of empirical stylized facts drawn from the recent trends in the software industry.

[Cahier de recherche] Free/Libre/Open Source Software (FLOSS) : lessons for intellectual property rights management in a knowledge-based economy.

Jean-Benoît Zimmermann, Nicolas Jullien — 2006-08-30T14:09:07Z

Résumé.

Dans cet article nous nous intéressons au fait qu'aujourd'hui, le logiciel libre n'est plus produit simplement par des individus, mais par des entreprises, qui ont leurs propores stratégies/objectifs de développement et qui embauchent des développeurs pour les atteindre. Il est alors important de comprendre les bénéfices qu'elles tirent d'un tel engagement. Autrement dit, nous nous intéressons aux différents modèles économiques de ces entreprises et leur évolution possible. Nous montrons d'abord que logiciel libre est une nouvelle manière de penser et de gérer sa propriété intellectuelle, qui amène à repenser les bénéfices qu'on peut retirer des taches de production de connaissances. Nous proposons ensuite une typologie des différentes stratégies industrielles vis-à-vis du logiciel libre, et comment ces stratégies correspondent aux caractéristiques techniques, économique de chaque sous-secteur. Dans une partie conclusive, nous essayons de tirer les principales leçons de l'expérience industrielle du logiciel libre et de son extension à d'autres activités commerciales basées sur la connaissance.

ABSTRACT.

The aim of this paper is to focus on the emerging situation in which open source software is nowadays produced not only by individual developers but in a growing proportion by firms that hire programmers for their own objectives of development in open source or for contributing to open source projects in the context of dedicated communities. As commercial firms it is important to analyze how and why they are capable to draw benefits from such involvement and their connected activities. In other way we want to stress the different types of business model these firms are leaning on and the possible evolution they are likely to know in a near future. We shown how Open Source principles provide an alternative way of thinking and managing intellectual property that do not come up against the same problems but needs a radical change in the way of drawing commercial benefits from knowledge development tasks. Then we analyze the growing involvement of commercial actors by setting up of a typology of the different business model that can be observed in the OS landscape, how they correspond to different strategies of industrial firms according to the main characteristics of their technical skills and market position. Finally, in a conclusive section we will intent to draw the main lessons of the FLOSS experience for a possible enlargement of those principles of IPR management and business to other knowledge based commercial activities.

Keywords : FLOSS, industrial organization in ICT.

Sommaire

Introduction.

Introduction.

A “free”/“libre” or “open source” software (FLOSS) is a software whose source-code, that is the explicit expression of the programming work, remains openly accessible. It appears as an alternative solution to the question of intellectual property in the computer software field, in which neither copyright nor patents can bring an acceptable balance between innovation incentives and knowledge diffusion. Copyright protecting not the ideas but a given expression of the ideas, software editors do not generally reveal the explicit expression of the programs (the source code) and suit those who try to disclose it. This behavior impedes knowledge diffusion in contradiction with the principles of IPR protection and handicaps the accumulative characteristic of innovation and the software products interoperability. On the other side patenting software could lead to a progressive partitioning of the field into proprietary owned procedures or algorithm and contradicts the recent evolution of programming techniques that are based on a closer relation to scientific knowledge and a combinatorial assembly of reusable components.

On the contrary, the alternative model of Open Source Software -OSS-, is based on a very innovative juridical concept called GPL "General Public License" and its diverse variations, and consists in imposing the producers to disclose both the source-code of the concerned programs and any further improvement if they are re-distributed/re-sold. It corresponds to a totally different approach of intellectual property rights, based on a weaker protection and the ability for all the actors to benefit from the whole set of innovations and progresses from a shared knowledge base. Of course, as in any public good question, this raises immediately problems of possible free-riding and of the incentive to disclose such knowledge in so far as accessing to knowledge doesn't depend on having contributed or not. That's the reason why the viability of this new way of IPR management will depend on the sustainability of associated business models and institutional supports.

Until recently, FLOSS was considered as only concerning programmers motivated by the building and the sharing of a base of programs developed for their own needs. Today, the open source model involves commercial enterprises and also an enlarged market of simple users. This brings to a paradoxical situation in which the development of business relies on the existence and durability of an activity of non-market nature. In former works, we have shown that solving such a paradox requires the setting up of new modes of incentives involving a pecuniary dimension additionally to the motivations of programmers originated in the initial movement. Such a turn still appears to be part of the actual way of working of FLOSS in so far as a growing amount of the code is produced by salaries that are paid for doing so. Such “hybridization”, mixing market and non-market rationales, nowadays appears as an inescapable evolution that also challenges policy makers for integrating support to FLOSS in the instruments of technological policy. It is then of growing importance to better understand under which conditions such a model of IPR management could extend to a growing number of knowledge intensive economic activities.

The aim of this paper is to focus on the emerging situation in which FLOSS is nowadays produced not only by individual developers but in a growing proportion by firms that hire programmers for their own objectives of development in open source or for contributing to open source projects in the context of dedicated communities. As commercial firms it is important to analyze how and why they are capable to draw benefits from such involvement and their connected activities. In other way we want to stress the different types of business model these firms are leaning on and the possible evolution they are likely to know in a near future.

Section 2 is devoted to the analysis of IPR traditional protecting ways in the software industry and their failure. We will explain how Open Source principles provide an alternative way of thinking and managing intellectual property that do not come up against the same problems but needs a radical change in the way of drawing commercial benefits from knowledge development tasks. In the section 3 we will describe how FLOSS has progressively switched from a contribution model based on individuals' benevolent efforts to an actual industrial one. Then section 4 will aim the setting up of a typology of the different business model that can be observed in the OS landscape, how they correspond to different strategies of industrial firms according to the main characteristics of their technical skills and market position. Finally, in a conclusive section we will intent to draw the main lessons of the FLOSS experience for a possible enlargement of those principles of IPR management and business to other knowledge based commercial activities.

[Cahier de recherche] Peut-on envisager une écologie du logiciel libre favorable aux nuls ?

Jean-Benoît Zimmermann, Nicolas Jullien — 2005-11-29T09:38:00Z

Nous abordons dans cet article la question de la diffusion des logiciels libres vers les simples utilisateurs. Si l'appropriation par ces utilisateurs des logiciels libres est souvent possible, si elle est souhaitable pour la pérennité du modèle, nous montrons que cette diffusion risque de faire évoluer sensiblement l'équilibre de l'organisation libre, aujourd'hui fondée sur une coopération technique.

Cela fait maintenant plusieurs années que le phénomène libre a émergé dans la sphère économique et de nombreux travaux scientifiques se sont intéressés au phénomène, parmi lesquels ceux publiés dans Terminal en 1999 [1] ont été précurseurs. Six ans plus tard, s'il reste beaucoup à faire pour comprendre le fonctionnement des communautés de développement, certaines questions posées alors ont trouvé des éléments de réponse, notamment en ce qui concerne les motivation des développeurs individuels (von Hippel (2002), Pénard et Dang Nguyen (2001), Jullien (2001), Foray et Zimmermann (2001), Demazière, Horn, Jullien (2004)...) ou les modèles économiques (succès de RedHat, Jullien (2001), Coris (2004).

Elles reposent toutes sur la relative homogénéité des utilisateurs et des développeurs de logiciel libre ; pour résumer, et jusqu'à récemment, le logiciel libre était affaire d'informaticiens, qui co-développaient leurs outils de travail et avaient intérêt à travailler ensemble car leurs compétences étaient complémentaires, les 'clients' des entreprises du libre étant, à l'époque, plutôt les services informatiques des grands groupes (Jullien 2003).

Ce constat est aujourd'hui remis en cause, avec la diffusion de logiciels destinés aux postes de travail, fonctionnant sous Windows, comme Open Office ou Mozilla et leur installation sur des ordinateurs de bureau [2]. La plupart des annonces relatives aux logiciels libres par des grands comptes porte d'ailleurs sur l'adoption de ces logiciels (villes de Munich et de Paris, gendarmerie française...) et leur diffusion dans la population (exemple des lycées avec le Conseil Régional PACA). Les entreprises proposant des solutions « libres » se tournent aussi de plus en plus vers les PME, comme en témoigne le dossier du Journal du Net de septembre 2005 [3].

Bref, le niveau de compétence des utilisateurs des logiciels libres baisse dans la mesure où, après une première vague de ralliement d'utilisateurs éclairés, ce sont désormais les simples utilisateurs rejoignent le monde du Libre [4].

Cette diffusion, voulue dès le départ par les 'inventeurs' du concept de logiciel libre [5], nous semble remettre en cause l'écologie de l'organisation libre, basée sur la logique du "don-contre don" [6] (que ce soit en termes de temps passé ou de contribution financière). En moindre mal elle ne profitera pas aux simples utilisateurs, au pire des cas, elle peut dissuader les développeurs de continuer à collaborer à ce genre de logiciels. Nous proposons dans cet article, de discuter des risques de cette diffusion.

Dans les deux premières parties, nous montrerons pourquoi un nombre croissant de simples utilisateurs se rallient au logiciel libre et pourquoi la diffusion vers ces utilisateurs est effectivement une nécessité pour assurer le succès à long terme de l'organisation libre. Mais la présence de ces simples utilisateurs ne rentre pas de manière directe dans les modalités de l'organisation actuelle du libre fondée sur la logique du don-contre don (partie trois). Nous proposerons alors dans la partie 4 des pistes sur ce que pourrait être le nouvel équilibre de l'organisation libre.

Dans cette réflexion, nous nous appuierons sur l'étude de deux exemples de logiciels libres pour les non-informaticiens, celui d'Open Office et d'un ERP [7] libre, Dolibarr. Open office comme emblème d'un logiciel que tout le monde a ou peut avoir, avec une adoption plutôt répandue dans les grandes organisations et le soutien de grands groupes informatique, et L'ERP comme exemple de logiciel 'métier' à destination des PME. Nous étaierons aussi notre analyse avec sur les entretiens auprès de développeurs du libre que Nicolas Jullien a réalisés conjointement avec Didier Demazière et François Horn.

[1] "Logiciels Libres : de l'utopie au marché", Terminal, numéro

spécial N°s 80-81, Automne-Hiver 1999.

[2] HP vend aujourd'hui des portables sous Linux

(http://svmblogs.vnunet.fr/svm/2005/09/pc_portables_so.h

tml) disponible sur le site de Carrefour (http://www.carrefourmultimedia.

com/) quand, aux États-Unis, Wal-Mart

vend des station de travail (http://solutions.journaldunet.

com/0404/040406_walmart.shtml).

[3] http://solutions.journaldunet.com/dossiers/libre/sommaire.

shtml

[4] Gérard-Varet et Zimmermann (1985) proposent une typologie

des utilisateurs à laquelle nous faisons référence

ici dans un langage plus courant. Ils dénomment utilisateurs

"naïfs" ceux que nous désignons comme simples utilisateurs

et qui sont essentiellement demandeurs de "caractéristiques

d'utilisation", c'est-à-dire de modalités et de

performances d'utilisation, utilisateurs "sophistiqués" ceux

qui sont capable d'intervenir dans la composition des

éléments matériels et logiciels du système informatique

auquel ils ont recours et enfin utilisateurs "designers" qui

ne sont autres que ceux que nous désignons en ce qui nous

concerne ici comme les utilisateurs-développeurs qui sont

à la base du modèle de production du logiciel libre.

[5] http://www.gnu.org/gnu/manifesto.html

[6] Mauss M. (1923-1924)

[7] Enterprise Resource Planning.

[Cahier de recherche] New approaches to intellectual property : from open software to knowledge based industrial activities.

Jean-Benoît Zimmermann, Nicolas Jullien — 2005-10-18T08:55:19Z

Nous abordons dans cet article la propriété intellectuelle dans le secteur du logiciel. Nous montrons que ni le droit d'auteur, ni le brevet ne sont parfaitement adaptés aux besoins et spécificités de cette industrie. Le modèle « alternatif » du logiciel libre (open source), basé sur un concept juridique innovant, la licence GPL, prend une importance croissante. Nous en présentons les principales caractéristiques, qui imposent au producteur de rendre public le code source du programme protégé par une licence de ce type, ainsi que celui de toutes les modifications postérieures en cas de redistribution. Cela entraine une modification profonde des stratégies industrielles de valorisation de la propriété intellectuelle, allant vers un plus faible niveau de protection. Nous analysons les conséquences d'un tel mouvement en terme institutionnel et de politique publique. Enfin, nous notons que cette approche apparaît exemplaire pour un nombre croissant d'industries dans le contexte d'une économie basée sur la connaissance.

Though software intellectual property could not satisfactorily fall into any existing legal framework, all countries have taken the decision to range it under the category of copyright. Then the double objective of intellectual property protection is not satisfied, which consists, on the one hand, to grant to the inventor a provisional monopoly for exploiting his invention and, on the other hand, to oblige him to disclose the principles of his invention. To resort to the patents system as it became more and more usual in the US and which is in debate in Europe, raise other kind of problems. Now the alternative model of Open Source Software, based on a very peculiar juridical tool called GPL "General Public Licence", tends to take a growing importance. Its main principle is to impose to its adopters to disclose the source-code of the concerned programs and of any further improvement if they circulate them, as well as the free circulation of the code under the sole condition to maintain its "open" character.

This does not exclude a possible commercialization of these programs and do not limit “open source software” to a non-marketable sphere. Understanding that, firms have more recently joined the world of cooperative development and of free access to sources-codes. This enlargement is concerned with two types of strategies. On the one hand firms distribute open software products to enlarge base of users with the services that can help to use them : training, adaptation to specific case or context, hotline, maintenance, updating, ... On the other hand, a growing number of enterprises began to "free" part of their software products aiming to draw benefits from the potential of development of the free software community or to favour a large diffusion of a key-product imposing it as a de facto standard and looking for gains from the commercialisation of proprietary complementary products.

By doing that they introduce a totally different approach of intellectual property within their industrial strategies. But these strategies didn't accommodate very well with the GPL terms as they were and this actors enlargement has led to a juridical enlargement beyond the strict framework of the GPL. Number of “hybrid” licenses have been designed in order to control the extent of their openness.

So open software approach doesn't represent a denial of intellectual property but a new way to manage intellectual property. Through the GPL, intellectual property is not rejected, authors do not renounce to their rights but to the sole monopoly rent, such rights would authorize in a copyright regime. The main legal aspect is that, when a program is declared under GPL license, any code derived from it or integrating GPL code lines must also be available under GPL License. Hence GPL status is “contagious” in the sense that this status attached to any number of lines is automatically transmitted to the whole program into which they are incorporated. The authors do authorize anyone who wants to make use of their work (modifications, improvements, additional features ...) under the sole condition that the new product could also circulate freely.

Such an approach appears exemplary in the context of a knowledge based economy, for a growing number of industrial activities, for which the scope of knowledge that has to be mastered appears to be too large for a single even powerful agent. Consider knowledge as a mutual resource implies a reshaping of the value chain concept, cash flow being drawn from the usage of the knowledge base (services, complementary products), not from the knowledge itself. We shall illustrate this idea on a short list of knowledge based activities, like biotechnologies and health in order to bring to openness approach a larger extent.

In this paper, we propose to explain the reasons why the copyright framework has been chosen for computer intellectual property protection, why it does not work very well, but also why a patent system may not be better suited (part 1). We will then consider the alternative model of Open Source Software, based on a new concept of authorship rights as expressed by the GPL "General Public License". We shall replace the appearance of the free-software movement in its peculiar context and explain the main reasons of its present industrial success (part 2 and 3). We will defend the idea that this new way of considering intellectual property management echoes debated in other industries, such as biotechnologies (part 4). This has different implications in institutional and public policy we discuss in the last part (part 5).