3.1- A tale about the bazaarThis paper pre … 3.1- A tale about the bazaarThis paper presents a cautionary tale about Raymond’s (1999) bazaar metaphorwhen applied to the analysis of F/LOS development, and also when implementedin it. We have focussed on two communities where the conjugation of opennessand growth has taken its toll, in terms of both output quality and delivery, andsocial stability. As we have shown, these problems have been tackled throughinstitutional strategies that re-dress the balance between openness and stability,and might lead to unintended consequences too, such as conflict, stagnation ordecreases in participation. This observation guides our caution about the bazaaras an analytical metaphor: norms, values, membership processes and structuresfor governance are always essential in order to understand the productive andsocial dynamics of a project, as even their absence (for example in the case ofWikipedia at its earlier stages) is a particular configuration of these elements.Having established the relevance of focussing on F/LOS institutions, we showthat the epistemic community and legitimate peripheral concepts, complementedwith a rationale for the emergence of hierarchies, based on the process ofauthority distribution sketched in the theoretical framework, constitute usefultools for this analysis. It is interesting to note that in the cases we have selectedfor this paper, this latter process, that of delegation of power based on trust, isconstrained by the decentralisation of the structures through which developmenttakes place. This explains the difficulties faced by Debian and Wikipedia’srespective project leaders when trying to exercise any sort of meaningful authorityover their projects, in contrast, for example, with the GNU/Linux kernel or theBSD Unix distribution, and illuminates how the nature of technical platforms fordevelopment (Debian’s package management system and Wikipedia’s wikisoftware tools) influence the social structure of the communities that adopt them.
3.2- The F/LOS museumWe would like to end our conclusions by arguing that instead of thinking ofF/LOS projects as oriental bazaars, an inherently flawed metaphor because of itsfocus on individual exchange instead of collective construction of a stablestructure, perhaps it would be more useful to conceive them as public museums:spaces anybody can enter (in that sense, open), in which the right to hang artworkfrom the walls is granted by a curator (i.e. the project leader or integrator)depending on more or less objective criteria of quality, the values the museumstands for, coherence with previous contributions, the experience or even status ofthe would-be exhibitor etc. We think that this metaphor accommodates theintegration elements of F/LOS development, and the possibility of participation asa form of self-expression (not precluding vandalism, graffiti and sabotage) muchbetter than that of the bazaar, while also conveying the public good aspects ofaccessibility by the general public. Thinking about this F/LOS museum inarchitectural terms also makes it possible conceptualise the difficulties that fastgrowth represents for co-ordination and quality assurance, inasmuch a visitormight decide to hang her own picture from the wall in a distant, rarely visitedroom, regardless of rules, regulations or the presence of “security guards”91.By establishing a continuum that goes from the open museum to the medievalcathedral with which Eric Raymond originally compared proprietary softwarecompanies, we propose a subtler, more flexible and tolerant conception ofF/LOSS methodologies as a large, diffuse constellation inside the universe ofpossible social configurations for software development, one that in occasions,depending on policy decisions (“more or less open”, “more or less managed”)made by project participants and leaders, will intersect with the traditional one.Since this metaphor makes it possible to understand the emerging diversity oforganisations engaged in F/LOS development in terms of institutionalarrangements established (or evolved) in order to find a balance between theachievement of different, sometimes conflicting goals (such as stability versusinnovation, or appropriability versus community good-will), we suggest that ourproposal does not constitute a mere re-definition or re-naming of a phenomenon,but a useful and necessary re-conceptualisation that should contribute to a betterunderstanding of its nature.
3.3- Implications: Our analysis has several implications relevant for F/LOS projects and companiesintent on fostering F/LOS communities for support and co-development of theirproducts:• Institutional structures are essential determinants of a project’soutcomes and its administrator should pay special attention to theirdesign and evolution. • Technical knowledge about a project area is the primary source oflegitimacy for exercises of authority inside it. Therefore, the nature ofthe platform adopted for development (which determines the form in which knowledge is accumulated or dispersed throughout thecommunity), constrains the kind of management actions that can beattempted successfully inside a project.• Decentralised platforms with components that interlink usingstandardised interfaces are easier to scale and facilitate paralleldevelopment, but also make it difficult and costly to oversee thequality of contributions, and to exercise authority pro-actively(something which might hinder, for example, architectural innovationor, as we have shown in the case of Debian, slow down developmentspeed).• On the other hand, “vertical structures” where contributions areprocessed upstream through “gatekeeper” positions (such is the case ofthe GNU/Linux project) might be strained by growth as coreparticipants struggle to keep up with the pace of contributions andbecome bottlenecks to development. • Administrators and companies interested in retaining a certain degreeof control over a project should attempt to establish the basic goals tobe accomplished, as well as elementary norms regulating participantbehaviour at its early stages in order to inform entrants and dissuade“problematic” participants, thus avoiding unproductive discussions orstruggles for control over the project’s direction92.• Further steps towards formalisation of institutional structures throughthe codification of project norms and values and membershipacquisition procedures can strengthen legitimate participationprocesses, increase participants’ feeling of “ownership” over theproject and contribute to the focus of discussions. However, they donot guarantee the elimination of conflict, might dampen participation(unless potential entrants expect important benefits from the eventualacquisition of insider status) and are costly in terms of resources.tus) and are costly in terms of resources.
We use the concepts of Epistemic Community … We use the concepts of Epistemic Community, Legitimate Peripheral Participation and Distributed Authority to elaborate a model for the analysis of social and organisational dynamics in Free/Libre/Open Source projects. We apply it to two large and successful F/LOS projects, Wikipedia and Debian, and show how openness, decentralisation and authority dispersion have facilitated growth in participant size, diversity and technical complexity, but has also resulted in conflict and concerns about the quality and delivery of these projects’ output. We analyse the processes of institutional redesign undertaken in order to address these problems and produce empirical implications relevant for F/LOS communities and companies. Finally, we argue for the need to complement predominant individualistic exchange-based models of F/LOS development with others that acknowledge its social aspects as a setting for the collective production of knowledge goods, and propose, as a step in that direction, the adoption of a “museum metaphor” to balance that of the ‘bazaar’ frequently used in the social science literature on F/LOS.in the social science literature on F/LOS.