A piece about Wikipedia recently published by the Wall Street Journal raised several concerns about the future of project1. Since it started in 2001, critics have been skeptical about Wikipedia's ability to produce high quality articles. The fact that everyone can edit almost any article has always generated concerns because of both the apparent lack of credentials of the website's editors, and the possibility of vandalism. Despite these concerns the website has managed to achieve impressive results both quantitatively2, and qualitatively3. However, those problems never disappeared, much on the contrary, the website's popularity actually boosted them. As a consequence Wikipedia has generated an increasing amount of rules and is making plans to move from a decentralized decision making structure with regard to the publication of articles to a more centralized one. The purpose of this essay is to analyze some of the issues presented in the Wall Street Journal article, as well as some of the strategies Wikipedia is developing to address them.
When comparing market- or firm-based organization of production with peer production, Yochai Benkler4 demonstrates how the latter approach has relative advantages over the former ones in at least two dimensions. The first concerns information gains, or in other terms, how peer production systems do a better job at identifying the best suited person for a required task. The second dimension pertains to allocation efficiencies. The basic idea here is that if production is a function of how to organize agents, resources and tasks, and if firms and markets bind through contract and property a certain number of agents to a certain number of tasks and resources, peer production, by allowing a larger number of agents, resources and tasks to interact, means that the amount of possible combinations among them will be higher and consequently this system will be more productive in general5. This analysis suggests that Wikipedia would be far better off maintaining its current highly decentralized production model than switching to one of the other options. When comparing Wikipedia's proposal to centralize control over publication and maintain the decentralized production, we see that the change does not seem to affect any of the basic characteristics that make it more productive than the two others. Centralizing a certain degree of control in the hands of some of Wikipedia's top contributors means that the project might be turning into a hybrid model in terms of organizational form that combines characteristics of both the hierarchical centralized approach of firm-based production and the decentralization of peer production. The question would then change from whether one form should prevail over the other, to how much of each organizational arrangement should be used to achieve its purposes more efficiently.
In fact, when we look at the organizational structure of most open source software projects we see the exact kind of balance between centralization of control over the final result and decentralization of the tasks. A notorious example in the open source community is the Mozilla Foundation, the organization responsible for supervising the development of the Firefox web browser. Mozilla's browser now detains 25%6 of the world market for browsers and is developed through a process that combines both centralized control over the final product with decentralized participation of tens of thousands volunteers working throughout the world. Fixes made to the code of the browser are not released automatically to the final user before going through a well structured process of quality assurance and semi-centralized control. This is necessary to prevent security flaws, crashes, incompatibility and other malfunctioning in the software.
This hybrid system is not immune from criticism. One first problem relates to the impact that the change might have on the motivation of the agents involved in the process. In both market- and firms-based production systems, agents are motivated to engage in the productive process by economic incentives. However, motivations of the agents working in commons-based peer production systems are substantially different. Instead of being driven by extrinsic factors – such as economic incentives – motivations for agents in peer production systems are intrinsic and therefore subject more susceptible to social norms. Benkler points that one of the possible threats to motivation is “failure of integration”7. The concept represents the situation in which an individual participating in a peer production project perceives her contribution as being wasted – considering the person values this waste as impacting negatively her intrinsic motivation to participate8. The move to a centralized system to control quality and vandalism, would ultimately mean that many contributions would be refused, or wasted.
It is conceivable to think about another sort of threat to motivation operating in this case. Imagine the hypothesis that some long time Wikipedia editors participate in the project because they have developed affinity and friendship relationships with other editors as a consequence of the similarity of interests and of the amount of time they spent interacting with each other. Assuming that the social relations operate more or less the same way both the online and offline, the more one voluntarily interacts with someone else the stronger become the bounds that connect them. The project becomes, then the locus where they maintain part of their social life – the same reason why some people go to clubs, reading groups or bars. The sense of pertaining to a community is what is relevant here. Eventual changes in the organizational structure and/or technologic architecture of the website which affect their ability to participate in the project and continue to interact in the community, may undermine editors' motivations and cause them to stop contributing. In this sense, motivation and participation interact with each in a sort of ascending spiral, each contributing to increase the other.
If this is true, then given the particular nature of the motivations of the agents participating in peer production systems, Wikipedia's focus on increasing participation among existing users instead of directing efforts to increase the total number of contributors makes good sense. Nevertheless, there are at least two arguments that would justify a continuous effort to increase the total number of editors. First, because one of the organizational advantages that peer production systems have in comparison to either firms and markets can be attributed to the higher number of combinations possible between agents, resources and projects, an increase in any of these factors improves the productivity of the system in overall. Second, we cannot neglect the advantages that a large number of contributors have in identifying even the smallest of the problems and working on very complex problems, or as Eric Raymond put it: “given enough eyeballs, all bugs are shallow”9. On the other hand, as Jimmy Wales states in the article, it is hard to know how much volunteers would be necessary to keep Wikipedia running. It is also hard to know how the curve of actual contributions as a function of users would behave in a large scale.
A different analysis about the effects that a more centralized control over articles could have on Wikipedia can be extracted from an argument made by Jonathan Zittrain10. He cites an experiment made in the Dutch city of Drachten which concludes that – no matter how counterintuitive it might look – people tend to drive more cautiously in roads that have no signs than in the ones that do. This would suggest that enforcing a rule stating that all edits should be approved by top editors before going live would actually decrease rather than improve accuracy of articles. The rationale is that users would not be so careful about validating the information of their contribution because there would be someone else to check its appropriateness before authorizing it to be published. Top editors already oversee changes made in articles, however, the difference resides in the fact that when a change is immediately published in the website without anyone's prior supervision the user is compelled more effectively by social norms – like public shame and respectability – to behave according to a desired conduct. Moreover, the fact that social norms won't be effective for vandalism done anonymously would not be enough to justify disregarding the effect that they might have on all other legitimate activities.
Born as a highly generative unregulated platform Wikipedia is becoming more controlled over time. The increasing amount of rules in the website seem to have a direct impact in users participation in at least two ways. First, they work as a threat to motivation. As we discussed, if contributions that do not comply with the website's rules are discarded the user might feel demotivated to participate for seeing her work wasted. The case here is slightly different from the one discussed above because what generates this demotivation is the amount of rules itself and not the way of enforcing them. Second, these rules might be operating as an entrance barrier for new users. This argument could actually be made not only for Wikipedia's rules but also for the site's technology, the only difference being the layer upon which each one apply. The point here is quite obvious: the easier to operate a technology – be that code or a way to express language – the more people will be able to use it.
In sum, our analysis suggests that:
The higher editorial control over Wikipedia's articles might turn its organizational form into a hybrid one, similar to the structure of some open source software projects. This organizational form is not a priori inconsistent with the basic characteristics that make peer production systems more advantageous than production systems based on either markets or firms;
This hybrid organizational structure may, however, affect the motivation of some editors to collaborate;
If Wikipedia functions as a locus where people maintain a substantial part of their social relations, there is reason to believe that the more interaction its rules and technology allows for users the more motivated they will feel to contribute to the project;
Despite the advantages that increased participation might generate, two reasons would justify a concomitant effort to increase the total number of participants: (a) increased gains in allocation efficiencies, and (b) the maxim known as Linux's Law: “given enough eyeballs, all bugs are shallow”;
Some analysis suggest that implementing a centralized system of control that operates a priori might counterintuitively change users' behavior making them less cautious about their contributions;
The implementation of a large amount of excessively rigorous rules might demotivate new users to contribute to Wikipedia in at least two ways: (a) by reinforcing the “failure to integrate” effect, and (b) by raising entrance barriers.
1Julia Angwin and Geoffrey A. Fowler, Volunteers Log Off as Wikipedia Ages. Wall Street Journal http://online.wsj.com/article/SB125893981183759969.html
2As of December 2009, Wikipedia had more than 3 million articles written in English.
3J. Giles, “Special Report: Internet Encyclopedias Go Head to Head,” Nature, December 14, 2005, available at http://www.nature.com/nature/journal/v438/n7070/full/438900a.html
4Yochai Benkler, Coase's Penguim, or, Linux and the Nature of the Firm (2002).
5See Yochai Benkler (2002) for a more detailed description of the argument.
6According to data from November 2009 measured by Net Applications. http://marketshare.hitslink.com/browser-market-share.aspx?qprid=0&qpmr=100&qpdt=1&qpct=3&qptimeframe=W&qpsp=566&sample=9
7Yochai Benkler (2002).
8One of the comments made by an user about Wikipedia's response to the WSJ article, showed this argument could be quite true:
“Well, I have taken hours editing and polishing a biographical article about a scientist. There is nothing in the article now that is under dispute, yet it is probably going to be taken down and deleted as one editor is exercising his or hers petty power-plays. Take it out and I’m never contributing again (neither time or money).
It is unbearably sad that wikipedia seems to be becoming the encyclopedia of all things pokemon, and deletes (or will) pages of scientists who have influenced courses in major universities and who are doing serious change out there.
All things Pokemon smoothly flow, unchallenged, while serious researchers are deleted?
Is that the future of wikipedia?”
http://blog.wikimedia.org/2009/11/26/wikipedias-volunteer-story/#comment-1694
9Eric Raymond, The Cathedral and the Bazaar, O'Reilly (1999).
10Jonathan Zittrain, The Future of the Intenet – and how to stop it. Yale University Press (2008).
Nenhum comentário:
Postar um comentário