Communities? Wikipedia, YouTube, del.icio.us and Other Projects

Vancouver.
The next session at AoIR 2007 begins with a paper by Ralph Schroeder and Mattijs den Besten on the section in the Pynchon wiki which has sprung up to collaboratively annotate Thomas Pynchon's much-anticipated novel Against the Day. There are some interesting statistics on how user participation shifted from the Pynchon mailing-list before the release of the book to the wiki once it was released; today, the wiki works as a reference source, as a tool highlighting connections to other Pynchon novels, and interpreting the content of the book. By July 2007, it had 200 contributors, 5000 entries, and contained some 400,000 words.

Like other wikis, it is volunteer-based, but requires registration for participation. There is only light-touch coordination, and the wiki separates discussion and annotation pages, but there is some overlap between the two. Content is organised mostly page by page (as an annotation of specific pages in the book), but topical or alphabetically ordered content pages also exist- but interestingly, in the first two months of the wiki, the latter two forms of content dominated, and page-specific annotations emerged and took over in dominance only from the third month of the wiki onwards. What remains to be seen is whether the wiki will ultimately turn into a fan discussion space, or a more scholarly form of annotation.

Up next is Sheizaf Rafaeli with a co-authored paper that focusses on Wikipedia. There's by now been a great deal of scholarly interest in Wikipedia, which has taken a number of research approaches to understanding the project; this paper examines especially the community aspects of the Wikipedia - do Wikipedians consider themselves a community, or more precisely, what forms of engagement relate to what levels of (self-perceived) membership in the community? Also, who are the opinion leaders in Wikipedia communities? Is Wikipedia an invisible college, a virtual knowledge-building community, a community of practice?

The research method outlined here to answer some of these questions involves surveying Wikipedians themselves; mapping their interactions, and an analysing their discussions (across a number of lanuage-specific Wikipedia version). Of the 344 respondents, some 90% were male; their age ranged between 18 and 82 (with a mean of 30); their education was 60% high school, 25% university. Key reasons reported for participation were pleasure, learning new things, the intellectual challenge, and sharing thier knowledge - and long-term users of Wikipedia as well as those looking for discussion appeared most likely to contribute actively. 61% reported interactions with ither Wikipedians within the site, just under 50% also outside the site. On some Wikipedia versions, communal aspects were important drivers - but the (global) English version scored markedly different on this point, perhaps due to its community's global dispersion.

The study also identified the 500 most active contributors to discussion pages, and examined their contributions. Main attributes were that they focussed strongly on specific subjects and performed more interactions with group members than with others - two key attributes of opinion leaders in Katz & Lazarsfeld's traditional model of opinion leadership. Other expected attributes in that model would be that they form part of a peer group, are non-professionals but considered experts by their group, and serve as the groups' information gatekeepers - those attributes were less clearly present in this sample.

The next presenter is Alice Marwick, who shifts our focus to YouTube. She notes the 'Macaca' moment, in which Senatorial candidate George Allen was captured on tape using a racial slur against an Indian-born supporter of his opponent in the campaign. Is, as some argue, YouTube mainly used today to damage democracy - to destroy candidates rather than make constructive campaign contributions?

Alice has examined the representation of YouTube as a campaign tool in the mainstream media (focussing here on mainstream periodicals in the U.S.), and found a number of typical characterisations (YouTube as an interesting novelty, copyright issues, media and celebrity users of the site, marketing oportunities, the LonelyGirl15 controversy, and the political uses of YouTube). However, almost all of these articles also discussed YouTube as a demcratising site: democratising politics, entertainment, media production, and celebrity.

What, then, does it mean to be democratic - and perhaps, what is meant here is more the egalitarian nature of sharing content on YouTube, in which users decide what content is popular. At the same time, Google continues to have the final say over YouTube content, and copyrighted content is increasingly removed while more sponsored content is added; there are also suggestions that some content is being censored. The average age on YouTube is around 20, Alice suggests, and some 70% of users remain U.S.-based; of course, broadband access (and production technology access) is required for full participation - so, only some 10% of users contribute content at present.

YouTube continues to contain a fair share of misogynist, racist, and homophobic content, and stereotypical content tends to bubble up to the front page; the honour system is being undermined by fake profiles and page refresher bots driving up the access count. Additionally, projects such as LonelyGirl15 have been created to game the system, and have variously been described as misleading the YouTube community or greating a new form of alternate reality games and Web media content.

Part of the business model here is the deployment of viral commercial videos (some non-attributed) and sponsored channels, as well as the harnessing of fan video; on the other hand, increasingly elaborate filters for copyrighted content have been deployed to weed out what is seen as copyright infringement. Additionally, of course, the use of the site in current election campaigns has also generated a great deal of media interest.

Up next is Evangelia Berdou, who looks at non-programming contributors to open source projects and has conducted interviews with a number of programmers and non-programmers in the Gnome and KDE projects to examine their community of practice in pursuit of a shared enterprise. Such communities are organised around a centre/periphery division which is consistent with the meritocratic structure of open source culture.

Non-coding tasks here include translation, documentation, artwork, Website maintenance, and release management; the translation and internationalisation team, for example, translate the GUI as well as documentation for projects, do not require advanced technical skills, and work in numerous language teams involving large numbers of contributors. Development documentation teams require stronger technical skills and more advanced knowledge, and are typically smaller.

Such teams need to invite contributors, meaning that they need to lower to the barriers to participation whil ensuring high quality. Expertise remains important, however (just in an area other than programming), which means that insights gained at the periphery may not necessarily feed well into the ongoing development efforts. Communication between coders and non-coders appears to be modular and is usually moderated by individuals moving between the boundaries of both groups. Programmers have to conform to certain conventions in order to enable translation work (marking up strings for translation in certain ways), and translation work takes place mainly at periods of 'string-freeze' during which developers are no longer allowed to change strings so that translators and documenters can catch up with the work required - this turns translation into a seasonal form of work in these projects, and can create tensions between the two groups.

Translation and documentation activity constitute distinctive spheres of activity, then - they are characterised by autonomous peripherality rather than legitimate peripherality. Different groups within the project are hierarchically organised, with some activities considered more important than others. This has implications for other aspects of participation. The overall projects can therefore be described better as constellations of practice rather than as homogenous communities of practice - and the dynamics of autonomous peripherality depend on the projects's current domain and phase.

Finally now to a paper by Robert Mason, David Hendry, and Karine Barzilai-Nahon. The highlight the importance of user input into design processes, which has been well established by now. Traditional approaches would gather user requirements, set specifications, and then develop the code; more recently we've seen moves towards participatory design and rapid prototyping, and now towards a democratisation or disintermediation of the design process. David Hendry has begun to characterise some of the dimensions of such development - user access to extension mechanisms, user control over installation and configuration, source code availability, and user access to release schedules.

We may be in a transitional era, therefore; full democracy may be unlikely and may not be the best metaphor, and the emergent structure produces well-defined roles and differential powers for participants. From a study of the development processes around the social bookmarking site del.icio.us, such roles include chief architect, application developers, posters, reporters, blog commentators, and monitored users, for example - and it may be useful to structure future development projects in anticipation of such role distributions.

Technorati : Against the Day, AoIR 2007, Pynchon, Wikipedia, YouTube, aoir8, community, del.icio.us, democracy, development, open source, produsage, wiki
Del.icio.us : Against the Day, AoIR 2007, Pynchon, Wikipedia, YouTube, aoir8, community, del.icio.us, democracy, development, open source, produsage, wiki