Feb

14

Peer Production, Bias, and Competition

Filed Under Peer Production | 

By way of InstaPundit, here’s an interesting description of the struggle behind the scenes of a controversial Wikipedia page. The summary is provided by James Purtilo, a CS Professor at UMD College Park. He notes, in particular, one of the dangers when such material is frozen and not flagged with any sort of disclaimer:

When journalists accept such material, whether innocently or by knowingly giving faint diligence to an obligation to get ‘outside’ authoritative sources, the quality of material presented on Wikipedia becomes inappropriately boosted in the eyes of the public.

This concern certainly seems rational, though I suspect over time it will take care of itself: Wikipedia will become less credible, and that lack of credibility will extend to sources that cite it. Perhaps more importantly, this sort of problem may spur the emergence of more competitors, along with aggregator and search technologies to help users sort out a broader set of sources.

WikiPedia is (for the most part correctly, I think) held up as an example of successful instance of “commons-based peer production”, a term coined by Yochai Benkler in his 2002 paper, “Coase’s Penguin” to describe various sort of open-source projects, online communities, and other activities that make use of small contributions from large numbers of distributed participants.

Benkler, noting the usual “tragedy of the commons” argument, addresses a number of potential issues of production in the absence of traditional property rights. One such argument is that it will be difficult to sustain contribution in the case where contributors cannot appropriate the results of their efforts. Clearly, however, in cases such as the one that Purtilo describes, lack of motivation is not an issue. Furthermore, one could argue that at the level of an individual entry, Wikipedia is no longer a commons. Freezing an entry is clearly a form of appropriation, as is obsessive editing and reverting. It may be a public park, but if I chase off anybody who tries to use the swingset, the swingset is effectively mine (until the cops arrive).

Ultimately, this weakness is probably inherent in any single centralized system. Nick Denton’s classic remark on flamers points towards a solution:

… this is the way to deal with flamers: let them post on their own damn sites. And then let everyone else ignore them. Weblogs are a gigantic interlinked discussion forum, in which it’s trivially easy to route around idiots.

Obviously resources like Wikipedia serve a purpose that would not be well served by the chronological structure of blogs, but the point translates fairly easily. Once they have control of an entry, there is no way to “route around the idiots” in the context of a single wiki. Appeals to administrators may or may not help. At a minimum, such appeals increase the overhead of integrating a peer-production effort.

The cost of bandwidth and software for setting up a competing Wikipedia-style site is minimal, and I wouldn’t be surprised if we see more competitors emerge. Perceptions of bias often drive exploration of new media, though the ability to focus on more specialized niches may prove to be a more significant motivation. For example, I would consider PlanetMath to be a better source than Wikipedia within a particular niche.

Competition will tend to punish the sort of behavior that Purtilo documented. However, it also makes the user’s job more difficult, requiring better aggregator and search technologies to manage more broadly distributed content. Aggregation and search algorithms can, and will, also be gamed, and the effective half-life of these technologies seems to be shrinking. The silver lining: No shortage of interesting work for software designers.

Comments

Comments are closed.